{"schema":2,"meta":{"source_fingerprint":"92f36c5d4fa83aef:1048577","last_updated":"2026-11-02","note":"Built by scripts/build-research-graph.mjs from the unified literature vault; re-running on an unchanged vault reproduces this payload byte-for-byte (the manifest carries the build timestamp).","counts":{"papers_total":2669,"papers_verified":2659,"edges_curated":2391,"nodes_drawn":746,"edges_drawn":1299,"semantic_edges":638,"cites_edges":661,"dangling_dropped":81,"authored_total":23,"authored_drawn":10,"authored_without_edges":13,"retracted_shown":1,"relation_counts":{"CITES":661,"EXTENDS":210,"COMPLEMENTS":262,"USES_METHOD":36,"OPERATIONALIZES":61,"CONTRASTS_WITH":64,"SHARES_BENCHMARK":2,"REFUTES":3}}},"story_sentence":"Across areas, 570 drawn relations touch the foundations layer; the strongest program-to-program link is Applied AI in Operations ↔ Distributed Systems (15 relations).","areas":[{"id":"ai-operations","label":"Applied AI in Operations","short":"Applied AI in Ops","desc":"Agents, retrieval, and verification gates on live production workloads. One program, deliberately kept whole: the study series publishes here, so the count is high and the area card says exactly how much of it is drawn. 1130 papers in the vault, 165 drawn here.","prio":"primary","n":165,"e":468,"authored":4,"top":[{"k":"glerum2009wer","t":"Debugging in the (very) large","c":1000},{"k":"he2017drain","t":"Drain: An Online Log Parsing Approach with Fixed Depth Tree","c":998},{"k":"bettenburg2008goodbug","t":"What Makes a Good Bug Report?","c":948}]},{"id":"calibration-conformal","label":"Calibration & Conformal Prediction","short":"Calibration","desc":"When a model's confidence is allowed to mean something: prior shift, selective prediction, risk-controlled decisions. 60 papers in the vault, 10 drawn here.","prio":"primary","n":10,"e":25,"authored":0,"top":[{"k":"zheng2023judging","t":"Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena","c":624},{"k":"guo2017calibration","t":"On Calibration of Modern Neural Networks","c":521},{"k":"angelopoulos2022ltt","t":"Learn then Test: Calibrating Predictive Algorithms to Achieve Risk Control","c":487}]},{"id":"security-governance","label":"AI Governance & Security","short":"Governance","desc":"ISO/IEC 42001 and 5259 controls, EU AI Act obligations, privacy firewalls, sandboxed agent tools. 85 papers in the vault, 32 drawn here.","prio":"primary","n":32,"e":64,"authored":2,"top":[{"k":"mitchell2019modelcards","t":"Model Cards for Model Reporting","c":690},{"k":"sweeney2002kanonymity","t":"k-ANONYMITY: A MODEL FOR PROTECTING PRIVACY","c":663},{"k":"iso-42001","t":"ISO/IEC 42001:2023 Information technology - Artificial intelligence - Management system","c":650}]},{"id":"distributed-systems","label":"Distributed Systems","short":"Distributed Systems","desc":"The database floor: replication, LSM engines, grey failures, fleet upgrades under measurement. 76 papers in the vault, 23 drawn here.","prio":"primary","n":23,"e":78,"authored":4,"top":[{"k":"o’neil-1996","t":"The log-structured merge-tree (LSM-tree)","c":994},{"k":"corbett2012spanner","t":"Spanner: Google's Globally-Distributed Database","c":641},{"k":"berenson-1995","t":"A Critique of ANSI SQL Isolation Levels.","c":487}]},{"id":"econometrics","label":"Econometrics & Causal Inference","short":"Causal Inference","desc":"Deconfounding and causal discovery — separating what correlates from what an intervention changes. 26 papers in the vault, 13 drawn here.","prio":"primary","n":13,"e":43,"authored":0,"top":[{"k":"callaway2021did","t":"Difference-in-Differences with Multiple Time Periods","c":553},{"k":"heckman1979selection","t":"Sample Selection Bias as a Specification Error","c":522},{"k":"sun2021event","t":"Estimating Dynamic Treatment Effects in Event Studies with Heterogeneous Treatment Effects","c":506}]},{"id":"sports-analytics","label":"Sports Analytics","short":"Sports Analytics","desc":"Public-data football studies: the proving ground where the site's method was formed. 15 papers in the vault, 9 drawn here.","prio":"primary","n":9,"e":6,"authored":0,"top":[{"k":"boyko2007referee","t":"Referee Bias Contributes to Home Advantage in English Premiership Football","c":430},{"k":"dawson2007referees","t":"Are Football Referees Really Biased and Inconsistent? Evidence on the Incidence of Disciplinary Sanction in the English Premier League","c":420},{"k":"wright2003professional","t":"The Professional Foul in Football: Tactics and Deterrents","c":413}]},{"id":"maritime","label":"Maritime Emissions","short":"Maritime","desc":"AIS telemetry and EU maritime regulation (ETS, FuelEU Maritime) behind the live dashboard.","prio":"primary","n":4,"e":5,"authored":0,"top":[{"k":"jalkanen-2009","t":"A modelling system for the exhaust emissions of marine traffic and its application in the Baltic Sea area","c":425},{"k":"johansson-2017","t":"Global assessment of shipping emissions in 2015 on a high spatial and temporal resolution","c":372},{"k":"jalkanen-2012","t":"Extension of an assessment model of ship traffic exhaust emissions for particulate matter and carbon monoxide","c":372}]},{"id":"energy-grids","label":"Energy & Grids","short":"Energy","desc":"Grid economics and the energy cost of computation — the counterpart ledger.","prio":"primary","n":3,"e":2,"authored":0,"top":[{"k":"pereira2017energy","t":"Energy efficiency across programming languages: how do energy, time, and memory relate?","c":374},{"k":"pereira2021ranking","t":"Ranking programming languages by energy efficiency","c":358},{"k":"vankempen2024green","t":"It's Not Easy Being Green: On the Energy Efficiency of Programming Languages","c":349}]},{"id":"foundations-viz","label":"Graphics & Perception","short":"Graphics & Perception","desc":"Graphical perception, visual encoding, and honest data communication: the evidence this page's own design choices follow. 90 papers in the vault, 67 drawn here.","prio":"secondary","n":67,"e":189,"authored":0,"top":[{"k":"satyanarayan-2017","t":"Vega-Lite: A Grammar of Interactive Graphics","c":984},{"k":"cleveland1984graphical","t":"Graphical Perception: Theory, Experimentation, and Application to the Development of Graphical Methods","c":855},{"k":"heer2010crowdsourcing","t":"Crowdsourcing graphical perception","c":764}]},{"id":"foundations-stats","label":"Statistics & Experimentation","short":"Statistics","desc":"Inference, calibration, survival, experimentation: the statistical standing orders behind every study. 199 papers in the vault, 101 drawn here.","prio":"secondary","n":101,"e":235,"authored":0,"top":[{"k":"cox-1972","t":"Regression Models and Life-Tables","c":859},{"k":"kaplan-1958","t":"Nonparametric Estimation from Incomplete Observations","c":787},{"k":"georges2007statistically","t":"Statistically rigorous java performance evaluation","c":779}]},{"id":"foundations-telemetry","label":"Log & Failure Analysis Classics","short":"Log Classics","desc":"Console-log mining, anomaly detection, debugging at scale — the direct ancestry of the AI-operations work. 162 papers in the vault, 80 drawn here.","prio":"secondary","n":80,"e":291,"authored":0,"top":[{"k":"fu2009execution","t":"Execution Anomaly Detection in Distributed Systems through Unstructured Log Analysis","c":741},{"k":"hooimeijer-2007","t":"Modeling bug report quality","c":694},{"k":"kim2006bugfixtime","t":"How long did it take to fix bugs?","c":679}]},{"id":"foundations-systems","label":"Systems & Infrastructure","short":"Systems","desc":"Storage engines, networking, cloud operations: where distributed-systems research begins. 161 papers in the vault, 86 drawn here.","prio":"secondary","n":86,"e":210,"authored":0,"top":[{"k":"qiu2013schemacoevolution","t":"An empirical analysis of the co-evolution of schema and code in database applications","c":686},{"k":"mytkowicz2009wrong","t":"Producing wrong data without doing anything obviously wrong!","c":615},{"k":"schroeder-2006","t":"A large-scale study of failures in high-performance computing systems","c":597}]},{"id":"foundations-se","label":"Software Engineering & Human Factors","short":"Software Eng.","desc":"Modularity, bug reports, performance evaluation, reproducibility, usability: the engineering discipline layer. 142 papers in the vault, 81 drawn here.","prio":"secondary","n":81,"e":163,"authored":0,"top":[{"k":"parnas1972criteria","t":"On the criteria to be used in decomposing systems into modules","c":684},{"k":"brodie2005symptom","t":"Quickly Finding Known Software Problems via Automated Symptom Matching","c":631},{"k":"lee-2004","t":"Trust in Automation: Designing for Appropriate Reliance","c":583}]},{"id":"foundations-other","label":"Foundations — Other Fields","short":"Other Classics","desc":"Classics from adjacent fields the keyword rules did not place — medicine, forecasting, domain science. Kept honest by the count. 297 papers in the vault, 72 drawn here.","prio":"secondary","n":72,"e":139,"authored":0,"top":[{"k":"wang1996dataquality","t":"Beyond Accuracy: What Data Quality Means to Data Consumers","c":644},{"k":"khatri2010designing","t":"Designing data governance","c":519},{"k":"breck2017mltestscore","t":"The ML test score: A rubric for ML production readiness and technical debt reduction","c":508}]}],"map_viewbox":{"w":1484,"h":620},"map_nodes":[{"id":"ai-operations","x":623.3,"y":387.2,"n":165},{"id":"calibration-conformal","x":528.2,"y":524.6,"n":10},{"id":"security-governance","x":653.3,"y":234.4,"n":32},{"id":"distributed-systems","x":896.1,"y":520,"n":23},{"id":"econometrics","x":413.2,"y":369.6,"n":13},{"id":"sports-analytics","x":1409.9,"y":326.8,"n":9},{"id":"maritime","x":736.3,"y":74,"n":4},{"id":"energy-grids","x":74,"y":252.3,"n":3},{"id":"foundations-viz","x":930.4,"y":339.1,"n":67},{"id":"foundations-stats","x":474.4,"y":217.8,"n":101},{"id":"foundations-telemetry","x":838.1,"y":193.2,"n":80},{"id":"foundations-systems","x":671.7,"y":546.4,"n":86},{"id":"foundations-se","x":578.1,"y":77.4,"n":81},{"id":"foundations-other","x":778.2,"y":432.4,"n":72}],"map_edges":[{"s":"ai-operations","t":"calibration-conformal","r":"COMPLEMENTS","w":6},{"s":"ai-operations","t":"calibration-conformal","r":"CONTRASTS_WITH","w":2},{"s":"ai-operations","t":"calibration-conformal","r":"EXTENDS","w":1},{"s":"ai-operations","t":"calibration-conformal","r":"OPERATIONALIZES","w":1},{"s":"ai-operations","t":"calibration-conformal","r":"REFUTES","w":1},{"s":"ai-operations","t":"calibration-conformal","r":"USES_METHOD","w":1},{"s":"ai-operations","t":"distributed-systems","r":"COMPLEMENTS","w":2},{"s":"ai-operations","t":"distributed-systems","r":"CONTRASTS_WITH","w":3},{"s":"ai-operations","t":"distributed-systems","r":"EXTENDS","w":2},{"s":"ai-operations","t":"distributed-systems","r":"OPERATIONALIZES","w":1},{"s":"ai-operations","t":"distributed-systems","r":"USES_METHOD","w":2},{"s":"ai-operations","t":"econometrics","r":"EXTENDS","w":1},{"s":"ai-operations","t":"foundations-other","r":"COMPLEMENTS","w":7},{"s":"ai-operations","t":"foundations-other","r":"CONTRASTS_WITH","w":3},{"s":"ai-operations","t":"foundations-other","r":"EXTENDS","w":5},{"s":"ai-operations","t":"foundations-other","r":"OPERATIONALIZES","w":1},{"s":"ai-operations","t":"foundations-other","r":"USES_METHOD","w":2},{"s":"ai-operations","t":"foundations-se","r":"CONTRASTS_WITH","w":1},{"s":"ai-operations","t":"foundations-se","r":"EXTENDS","w":3},{"s":"ai-operations","t":"foundations-se","r":"OPERATIONALIZES","w":1},{"s":"ai-operations","t":"foundations-stats","r":"COMPLEMENTS","w":12},{"s":"ai-operations","t":"foundations-stats","r":"CONTRASTS_WITH","w":3},{"s":"ai-operations","t":"foundations-stats","r":"EXTENDS","w":7},{"s":"ai-operations","t":"foundations-stats","r":"OPERATIONALIZES","w":2},{"s":"ai-operations","t":"foundations-stats","r":"USES_METHOD","w":2},{"s":"ai-operations","t":"foundations-systems","r":"COMPLEMENTS","w":16},{"s":"ai-operations","t":"foundations-systems","r":"CONTRASTS_WITH","w":1},{"s":"ai-operations","t":"foundations-systems","r":"EXTENDS","w":7},{"s":"ai-operations","t":"foundations-systems","r":"OPERATIONALIZES","w":1},{"s":"ai-operations","t":"foundations-systems","r":"USES_METHOD","w":5},{"s":"ai-operations","t":"foundations-telemetry","r":"COMPLEMENTS","w":17},{"s":"ai-operations","t":"foundations-telemetry","r":"CONTRASTS_WITH","w":8},{"s":"ai-operations","t":"foundations-telemetry","r":"EXTENDS","w":15},{"s":"ai-operations","t":"foundations-telemetry","r":"OPERATIONALIZES","w":2},{"s":"ai-operations","t":"foundations-telemetry","r":"REFUTES","w":1},{"s":"ai-operations","t":"foundations-telemetry","r":"USES_METHOD","w":1},{"s":"ai-operations","t":"foundations-viz","r":"COMPLEMENTS","w":2},{"s":"ai-operations","t":"foundations-viz","r":"OPERATIONALIZES","w":1},{"s":"ai-operations","t":"security-governance","r":"COMPLEMENTS","w":4},{"s":"ai-operations","t":"security-governance","r":"CONTRASTS_WITH","w":1},{"s":"ai-operations","t":"security-governance","r":"EXTENDS","w":2},{"s":"ai-operations","t":"security-governance","r":"OPERATIONALIZES","w":3},{"s":"calibration-conformal","t":"foundations-other","r":"COMPLEMENTS","w":2},{"s":"calibration-conformal","t":"foundations-se","r":"EXTENDS","w":1},{"s":"calibration-conformal","t":"foundations-systems","r":"OPERATIONALIZES","w":1},{"s":"distributed-systems","t":"foundations-other","r":"COMPLEMENTS","w":1},{"s":"distributed-systems","t":"foundations-other","r":"EXTENDS","w":1},{"s":"distributed-systems","t":"foundations-se","r":"CONTRASTS_WITH","w":1},{"s":"distributed-systems","t":"foundations-se","r":"EXTENDS","w":1},{"s":"distributed-systems","t":"foundations-se","r":"OPERATIONALIZES","w":5},{"s":"distributed-systems","t":"foundations-systems","r":"COMPLEMENTS","w":2},{"s":"distributed-systems","t":"foundations-systems","r":"CONTRASTS_WITH","w":2},{"s":"distributed-systems","t":"foundations-systems","r":"EXTENDS","w":2},{"s":"distributed-systems","t":"foundations-systems","r":"OPERATIONALIZES","w":2},{"s":"distributed-systems","t":"foundations-systems","r":"USES_METHOD","w":1},{"s":"distributed-systems","t":"foundations-telemetry","r":"COMPLEMENTS","w":1},{"s":"distributed-systems","t":"foundations-telemetry","r":"OPERATIONALIZES","w":1},{"s":"econometrics","t":"foundations-stats","r":"COMPLEMENTS","w":6},{"s":"econometrics","t":"foundations-stats","r":"EXTENDS","w":3},{"s":"econometrics","t":"foundations-stats","r":"USES_METHOD","w":1},{"s":"foundations-other","t":"foundations-se","r":"COMPLEMENTS","w":7},{"s":"foundations-other","t":"foundations-se","r":"EXTENDS","w":5},{"s":"foundations-other","t":"foundations-se","r":"OPERATIONALIZES","w":2},{"s":"foundations-other","t":"foundations-se","r":"USES_METHOD","w":1},{"s":"foundations-other","t":"foundations-stats","r":"COMPLEMENTS","w":7},{"s":"foundations-other","t":"foundations-stats","r":"CONTRASTS_WITH","w":1},{"s":"foundations-other","t":"foundations-stats","r":"EXTENDS","w":3},{"s":"foundations-other","t":"foundations-stats","r":"OPERATIONALIZES","w":1},{"s":"foundations-other","t":"foundations-systems","r":"COMPLEMENTS","w":3},{"s":"foundations-other","t":"foundations-systems","r":"CONTRASTS_WITH","w":1},{"s":"foundations-other","t":"foundations-systems","r":"EXTENDS","w":2},{"s":"foundations-other","t":"foundations-systems","r":"OPERATIONALIZES","w":2},{"s":"foundations-other","t":"foundations-telemetry","r":"COMPLEMENTS","w":1},{"s":"foundations-other","t":"foundations-telemetry","r":"CONTRASTS_WITH","w":1},{"s":"foundations-other","t":"foundations-telemetry","r":"EXTENDS","w":2},{"s":"foundations-other","t":"foundations-viz","r":"COMPLEMENTS","w":1},{"s":"foundations-other","t":"security-governance","r":"COMPLEMENTS","w":2},{"s":"foundations-other","t":"security-governance","r":"OPERATIONALIZES","w":1},{"s":"foundations-se","t":"foundations-stats","r":"COMPLEMENTS","w":4},{"s":"foundations-se","t":"foundations-stats","r":"CONTRASTS_WITH","w":1},{"s":"foundations-se","t":"foundations-stats","r":"EXTENDS","w":2},{"s":"foundations-se","t":"foundations-stats","r":"OPERATIONALIZES","w":1},{"s":"foundations-se","t":"foundations-stats","r":"USES_METHOD","w":1},{"s":"foundations-se","t":"foundations-systems","r":"COMPLEMENTS","w":3},{"s":"foundations-se","t":"foundations-systems","r":"EXTENDS","w":2},{"s":"foundations-se","t":"foundations-systems","r":"OPERATIONALIZES","w":1},{"s":"foundations-se","t":"foundations-systems","r":"USES_METHOD","w":2},{"s":"foundations-se","t":"foundations-telemetry","r":"COMPLEMENTS","w":1},{"s":"foundations-se","t":"foundations-telemetry","r":"CONTRASTS_WITH","w":3},{"s":"foundations-se","t":"foundations-telemetry","r":"EXTENDS","w":2},{"s":"foundations-se","t":"foundations-telemetry","r":"OPERATIONALIZES","w":1},{"s":"foundations-se","t":"foundations-viz","r":"COMPLEMENTS","w":1},{"s":"foundations-se","t":"foundations-viz","r":"EXTENDS","w":2},{"s":"foundations-se","t":"foundations-viz","r":"OPERATIONALIZES","w":3},{"s":"foundations-se","t":"security-governance","r":"COMPLEMENTS","w":2},{"s":"foundations-stats","t":"foundations-systems","r":"COMPLEMENTS","w":1},{"s":"foundations-stats","t":"foundations-systems","r":"CONTRASTS_WITH","w":1},{"s":"foundations-stats","t":"foundations-systems","r":"EXTENDS","w":1},{"s":"foundations-stats","t":"foundations-systems","r":"USES_METHOD","w":1},{"s":"foundations-stats","t":"foundations-telemetry","r":"COMPLEMENTS","w":3},{"s":"foundations-stats","t":"foundations-telemetry","r":"CONTRASTS_WITH","w":1},{"s":"foundations-stats","t":"foundations-telemetry","r":"EXTENDS","w":5},{"s":"foundations-stats","t":"foundations-telemetry","r":"OPERATIONALIZES","w":1},{"s":"foundations-stats","t":"foundations-telemetry","r":"USES_METHOD","w":1},{"s":"foundations-stats","t":"foundations-viz","r":"COMPLEMENTS","w":1},{"s":"foundations-stats","t":"foundations-viz","r":"EXTENDS","w":3},{"s":"foundations-stats","t":"security-governance","r":"EXTENDS","w":1},{"s":"foundations-stats","t":"security-governance","r":"OPERATIONALIZES","w":1},{"s":"foundations-systems","t":"foundations-telemetry","r":"COMPLEMENTS","w":6},{"s":"foundations-systems","t":"foundations-telemetry","r":"CONTRASTS_WITH","w":2},{"s":"foundations-systems","t":"foundations-telemetry","r":"EXTENDS","w":1},{"s":"foundations-systems","t":"foundations-telemetry","r":"OPERATIONALIZES","w":1},{"s":"foundations-systems","t":"foundations-telemetry","r":"USES_METHOD","w":2},{"s":"foundations-telemetry","t":"security-governance","r":"COMPLEMENTS","w":4},{"s":"foundations-viz","t":"security-governance","r":"EXTENDS","w":1},{"s":"foundations-viz","t":"security-governance","r":"OPERATIONALIZES","w":1},{"s":"maritime","t":"security-governance","r":"COMPLEMENTS","w":1}],"nodes":[{"k":"abadie-2021","a":"foundations-stats","t":"Using Synthetic Controls: Feasibility, Data Requirements, and Methodological Aspects","y":2021,"s":"peer_reviewed","e":3,"d":"10.1257/jel.20191450","v":"Journal of Economic Literature","a_":["Abadie"],"bl":"https://doi.org/10.1257/jel.20191450","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Practical guide to synthetic control: feasibility conditions (a good pre-period fit, no anticipation, no interference, a donor pool that is not itself treated), placebo-in-space and placebo-in-time tests, leave-one-out donors, and the ratio of post- to pre-period RMSPE for inference."}]},{"k":"abadie2010sc","a":"ai-operations","t":"Synthetic Control Methods for Comparative Case Studies: Estimating the Effect of California's Tobacco Control Program","y":2010,"s":"peer_reviewed","e":5,"d":"10.1198/jasa.2009.ap08746","v":"Journal of the American Statistical Association","a_":["Abadie et al."],"bl":"https://doi.org/10.1198/jasa.2009.ap08746","pc":1,"sm":[{"m":"agent draft (claude)","l":"support human ai","tx":"Synthetic control constructs a weighted combination of untreated units that matches the treated unit's pre-period path, with placebo-in-space inference. Placebo tests on other queues give an honest sense of whether the post-period TTR shift is unusual."}]},{"k":"abadie2023clustering","a":"foundations-systems","t":"When Should You Adjust Standard Errors for Clustering?","y":2022,"s":"peer_reviewed","e":1,"d":"10.1093/qje/qjac038","v":"The Quarterly Journal of Economics","a_":["Abadie et al."],"bl":"https://doi.org/10.1093/qje/qjac038"},{"k":"abeba-2024-arxiv","a":"security-governance","t":"AI auditing: The Broken Bus on the Road to AI Accountability","y":2024,"s":"preprint","e":1,"x":"2401.14462","v":"arXiv:2401.14462","a_":["Birhane et al."],"bl":"https://arxiv.org/abs/2401.14462","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Most AI audit work stops at evaluation and never reaches accountability outcomes. In ISO 42001 terms, a red tile that never opens a corrective action (9.1 without 10.2) is the broken bus."}]},{"k":"abedi2017rmit","a":"foundations-stats","t":"Conducting Repeatable Experiments in Highly Variable Cloud Computing Environments","y":2017,"s":"peer_reviewed","e":2,"d":"10.1145/3030207.3030229","v":"Proceedings of the 8th ACM/SPEC on International Conference on Performance Engineering","a_":["Abedi & Brecht"],"bl":"https://doi.org/10.1145/3030207.3030229","pc":1,"sm":[{"m":"agent draft (gemini)","l":"methodology","tx":"Establishes that unpaired or sequential cloud benchmarks fail under multi-tenant drift and introduces interleaved single-instance trial schedules, providing the methodological foundation for our single-VM palindromic ABBA and ABCDDCBA schedules in what_the_log."}]},{"k":"abitbol2026revolut","a":"foundations-other","t":"Shipping Like a Startup, Regulated Like a Bank: How Revolut Cut AI Coding Costs 90% With Self-Hosted Open Models","y":2026,"s":"external","e":2,"v":"Proceedings of the AI Engineer Summit New York 2026","a_":["Abitbol"],"bl":"https://ai.engineer/nyc/2026/talks/revolut","pc":1,"sm":[{"m":"agent draft (gemini)","l":"local inference economics","tx":"Documents how Revolut reduced AI coding and verification inference costs by 90% in a regulated European banking environment by migrating high-volume loops from per-token frontier APIs to self-hosted open-weight models."}]},{"k":"abowd2018census","a":"security-governance","t":"The U.S. Census Bureau Adopts Differential Privacy","y":2018,"s":"peer_reviewed","e":2,"d":"10.1145/3219819.3226070","v":"Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery &amp; Data Mining","a_":["Abowd"],"bl":"https://doi.org/10.1145/3219819.3226070","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Census 2020 adopted DP after reconstruction experiments showed exact published tables re-identify respondents; it trades accuracy at small geographies for a formal budget. Lesson for fleet advisories: many exact cross-tabulations from one population compose into a reconstruction risk even when each table looks harmless."}]},{"k":"abraham2019datagovernance","a":"foundations-other","t":"Data governance: A conceptual framework, structured review, and research agenda","y":2019,"s":"peer_reviewed","e":1,"d":"10.1016/j.ijinfomgt.2019.07.008","v":"International Journal of Information Management","a_":["Abraham et al."],"bl":"https://doi.org/10.1016/j.ijinfomgt.2019.07.008","pc":1,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Synthesizes 145 data governance publications (2001-2019) into a 6-dimensional conceptual framework: Governance Mechanisms (structural roles/accountabilities, procedural policies/standards/compliance monitoring, and relational training/shared language), Organizational Scope (intra- vs. inter-organizational), Data Scope (traditional structured vs."}]},{"k":"adams2020monarch","a":"foundations-systems","t":"A system design for elastically scaling transaction processing engines in virtualized servers","y":2020,"s":"peer_reviewed","e":4,"d":"10.14778/3415478.3415536","v":"Proceedings of the VLDB Endowment","a_":["Anadiotis et al."],"bl":"https://doi.org/10.14778/3415478.3415536","pc":1,"sm":[{"m":"agent draft (gemini)","l":"systems performance","tx":"Formalizes planet-scale time-series data modeling with explicit cumulative counter reset semantics and series-scoped delta aggregation prior to cross-series spatial rollups."}]},{"k":"afroozeh-2023","a":"ai-operations","t":"ALP: Adaptive Lossless floating-Point Compression","y":2023,"s":"peer_reviewed","e":3,"d":"10.1145/3626717","v":"Proceedings of the ACM on Management of Data","a_":["Afroozeh et al."],"bl":"https://doi.org/10.1145/3626717","pc":1,"sm":[{"m":"agent draft (qwen)","l":"operational translation","tx":"DocDB analogue: block/key layout and encoding changes, not more prefetches."}]},{"k":"afroozeh-2025","a":"ai-operations","t":"The FastLanes File Format","y":2025,"s":"peer_reviewed","e":5,"d":"10.14778/3749646.3749718","v":"Proceedings of the VLDB Endowment","a_":["Afroozeh & Boncz"],"bl":"https://doi.org/10.14778/3749646.3749718","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"FastLanes file format (PVLDB 18): lightweight data-parallel (SIMD/GPU) cascaded encodings instead of general-purpose block compression; reports better ratio than Parquet with much faster decoding. ALP from the same group became a Parquet encoding (Preview, parquet-format 2.14)."}]},{"k":"aggarwal2024automix","a":"ai-operations","t":"AutoMix: Automatically Mixing Language Models","y":2024,"s":"external","e":2,"v":"Advances in Neural Information Processing Systems (NeurIPS 2024)","a_":["Aggarwal et al."],"pc":1,"sm":[{"m":"agent draft (qwen)","l":"operational translation","tx":"AutoMix: small model answers, a zero-shot self-consistency-based surrogate verifier decides when to escalate to a stronger model, giving better cost-quality tradeoffs than always-strong."}]},{"k":"aghili2025logs","a":"security-governance","t":"Protecting Privacy in Software Logs: What Should Be Anonymized?","y":2025,"s":"peer_reviewed","e":3,"d":"10.1145/3715779","v":"Proceedings of the ACM on Software Engineering","a_":["Aghili et al."],"bl":"https://doi.org/10.1145/3715779","pc":7,"sm":[{"m":"agent draft (gemini)","l":"format driven log redaction","tx":"Empirical study of sensitive attributes in software logs (FSE 2025) showing that sensitive tokens (credentials, IPs, emails, session IDs, user paths) leak across diverse log severities and unexpected error/exception templates rather than only dedicated auth/audit lines."}]},{"k":"agresti1998approximate","a":"foundations-stats","t":"Approximate is Better than “Exact” for Interval Estimation of Binomial Proportions","y":1998,"s":"peer_reviewed","e":1,"d":"10.1080/00031305.1998.10480550","v":"The American Statistician","a_":["Agresti & Coull"],"bl":"https://doi.org/10.1080/00031305.1998.10480550","pc":1,"sm":[{"m":"agent draft (claude)","l":"dashboard fix validation","tx":"Shows the Wald interval performs badly and the Clopper-Pearson 'exact' interval is overly conservative; the score (Wilson) interval or the 'add two successes and two failures' adjusted Wald interval gives near-nominal coverage even for small n and p near 0 or 1. Directly applicable to SLA attainment tiles near 100% with small n (e.g."}]},{"k":"ahmed2023recommending","a":"ai-operations","t":"Recommending Root-Cause and Mitigation Steps for Cloud Incidents using Large Language Models","y":2023,"s":"peer_reviewed","e":8,"d":"10.1109/ICSE48619.2023.00149","v":"2023 IEEE/ACM 45th International Conference on Software Engineering (ICSE)","a_":["Ahmed et al."],"bl":"https://doi.org/10.1109/ICSE48619.2023.00149","pc":4,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Evaluates LLMs fine-tuned on 40k+ historical Microsoft incidents to recommend root cause and mitigation; useful as a draft judged by on-call engineers, weaker on novel incidents. For the gallery: AI-note hypotheses must be compared to the human CE resolution class, not to LLM self-scores."}]},{"k":"aho1975efficient","a":"ai-operations","t":"Efficient string matching","y":1975,"s":"peer_reviewed","e":3,"d":"10.1145/360825.360855","v":"Communications of the ACM","a_":["Aho & Corasick"],"bl":"https://doi.org/10.1145/360825.360855","pc":4,"sm":[{"m":"agent draft (gemini)","l":"log template mining","tx":"Aho and Corasick (CACM 1975) construct a finite-state string-matching automaton from a finite set of keyword needles by combining a prefix trie (goto function), suffix failure links (failure function), and output sets, finding all overlapping occurrences of all needles in a text of length n in a single O(n + m) pass independent of dictionary size."},{"m":"agent draft (gemini)","l":"methodology","tx":"Aho and Corasick's 1975 CACM paper introducing the finite-state string matching machine with failure transitions that locates all occurrences of any finite set of keywords in a text string in a single pass in O(n) time independent of the number of keywords."}]},{"k":"ajani-2022","a":"foundations-viz","t":"Declutter and Focus: Empirically Evaluating Design Guidelines for Effective Data Communication","y":2022,"s":"peer_reviewed","e":2,"d":"10.1109/TVCG.2021.3068337","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Ajani et al."],"bl":"https://doi.org/10.1109/TVCG.2021.3068337"},{"k":"akbarinasaji-2017","a":"foundations-telemetry","t":"Predicting bug-fixing time: A replication study using an open source software project","y":2018,"s":"peer_reviewed","e":5,"d":"10.1016/j.jss.2017.02.021","v":"Journal of Systems and Software","a_":["Akbarinasaji et al."],"bl":"https://doi.org/10.1016/j.jss.2017.02.021","pc":1,"sm":[{"m":"agent draft (claude-sonnet-5-5)","l":"complexity measurement without effort","tx":"Replication of bug-fixing-time prediction on an open-source project (Bugzilla-style data). Replication reports weaker/different predictive structure than the original commercial-project study, illustrating that elapsed time is a noisy, context-dependent target."}]},{"k":"akidau2013millwheel","a":"foundations-systems","t":"MillWheel","y":2013,"s":"peer_reviewed","e":2,"d":"10.14778/2536222.2536229","v":"Proceedings of the VLDB Endowment","a_":["Akidau et al."],"bl":"https://doi.org/10.14778/2536222.2536229","sm":[{"m":"vault note","l":"summary","tx":"Akidau et al. (PVLDB 2013) MillWheel: Fault-Tolerant Stream Processing at Internet Scale"}]},{"k":"akidau2015dataflow","a":"foundations-systems","t":"The dataflow model","y":2015,"s":"peer_reviewed","e":4,"d":"10.14778/2824032.2824076","v":"Proceedings of the VLDB Endowment","a_":["Akidau et al."],"bl":"https://doi.org/10.14778/2824032.2824076","pc":3,"sm":[{"m":"agent draft (claude)","l":"sre platform","tx":"Separates event time from processing time and defines the watermark: the point in event time up to which the system believes it has seen all input. For platform health, watermark lag (now minus watermark) is the right per-source freshness signal for incremental syncs."},{"m":"agent draft (gemini)","l":"streaming rle watermarks","tx":"Formalizes unbounded stream processing along What (transformations), Where (event-time session/tumbling windows), When (watermarks and early/on-time/late triggers), and How (discarding vs accumulating refinements)."},{"m":"agent draft (gemini)","l":"streaming windowing rle","tx":"Establishes the What/Where/When/How decomposition for unbounded out-of-order streams using event-time watermarks, session/tumbling windows, triggers, and accumulation modes (Discarding, Accumulating, Accumulating & Retracting)."}]},{"k":"aksin2007moderncallcenter","a":"foundations-telemetry","t":"The Modern Call Center: A Multi‐Disciplinary Perspective on Operations Management Research","y":2007,"s":"peer_reviewed","e":12,"d":"10.1111/j.1937-5956.2007.tb00288.x","v":"Production and Operations Management","a_":["Aksin et al."],"bl":"https://doi.org/10.1111/j.1937-5956.2007.tb00288.x","pc":2,"sm":[{"m":"agent draft (claude)","l":"support ops review","tx":"Multi-disciplinary review of call-centre OM: capacity planning, routing, skill-based staffing, and the human-resource side (agent burnout, turnover, learning)."},{"m":"agent draft (claude)","l":"vp support dashboard","tx":"POMS review of call-center OM research covering demand forecasting, capacity sizing, staffing/scheduling, routing and skills, and the interfaces with HR (agent attrition, learning) and marketing (service quality, revenue)."}]},{"k":"alagappan2018par","a":"ai-operations","t":"Protocol-Aware Recovery for Consensus-Based Distributed Storage","y":2018,"s":"peer_reviewed","e":3,"d":"10.1145/3241062","v":"ACM Transactions on Storage","a_":["Alagappan et al."],"bl":"https://doi.org/10.1145/3241062","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Protocol-aware recovery: storage faults in one replica's persisted Raft state (log, term, vote) must be repaired using the protocol, not by crashing or trusting local state. Council-5 DocDB lens: the 'Unexpected new leader in same term' FATAL family (GitHub #21808, volume remount under a running tserver;"}]},{"k":"alcott2005jevons","a":"foundations-other","t":"Jevons' paradox","y":2005,"s":"peer_reviewed","e":2,"d":"10.1016/j.ecolecon.2005.03.020","v":"Ecological Economics","a_":["Alcott"],"bl":"https://doi.org/10.1016/j.ecolecon.2005.03.020","pc":2,"sm":[{"m":"agent draft (gemini)","l":"local inference economics","tx":"Formalizes Jevons' paradox: technological increases in resource efficiency lower the effective price of a work unit, causing elastic demand to expand total consumption rather than reducing it."}]},{"k":"alexander-2023-arxiv","a":"foundations-telemetry","t":"SoK: Log Based Transparency Enhancing Technologies","y":2023,"s":"preprint","e":2,"x":"2305.01378","v":"arXiv:2305.01378","a_":["Hicks"],"bl":"https://arxiv.org/abs/2305.01378","pc":1,"sm":[{"m":"agent draft (qwen)","l":"operational translation","tx":"For a support tool consumed by vendors, the SoK's 'constrained-audience transparency' designs (hashed commitments + out-of-band checkpointing) are the realistic target, not public logs, because bundles carry customer PII."}]},{"k":"an2015crashinducing","a":"foundations-telemetry","t":"An Empirical Study of Crash-inducing Commits in Mozilla Firefox","y":2015,"s":"peer_reviewed","e":2,"d":"10.1145/2810146.2810152","v":"Proceedings of the 11th International Conference on Predictive Models and Data Analytics in Software Engineering","a_":["An & Khomh"],"bl":"https://doi.org/10.1145/2810146.2810152","sm":[{"m":"vault note","l":"summary","tx":"Crash-inducing commits in Firefox: links crash buckets to inducing changes/releases; basis for per-release signature emergence"}]},{"k":"andersen-1982","a":"foundations-stats","t":"Cox's Regression Model for Counting Processes: A Large Sample Study","y":1982,"s":"peer_reviewed","e":3,"d":"10.1214/aos/1176345976","v":"The Annals of Statistics","a_":["Andersen & Gill"],"bl":"https://doi.org/10.1214/aos/1176345976","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Counting-process extension of the Cox model for recurrent events with time-varying covariates and delayed entry. The right model for tickets per universe as a recurrent event with release (and release age) as time-varying covariates and universes entering exposure late (left truncation)."}]},{"k":"anderson2004satisfaction","a":"foundations-se","t":"Customer Satisfaction and Shareholder Value","y":2004,"s":"peer_reviewed","e":2,"d":"10.1509/jmkg.68.4.172.42723","v":"Journal of Marketing","a_":["Anderson et al."],"bl":"https://doi.org/10.1509/jmkg.68.4.172.42723","pc":1,"sm":[{"m":"agent draft (claude)","l":"vp support dashboard","tx":"Finds a positive association between ACSI customer satisfaction and shareholder value (Tobin's q). Evidence that satisfaction is a legitimate business-outcome dimension, which a support dashboard can link to account retention/expansion once CRM revenue data is joined."}]},{"k":"angelopoulos2022ltt","a":"calibration-conformal","t":"Learn then Test: Calibrating Predictive Algorithms to Achieve Risk Control","y":2022,"s":"peer_reviewed","e":5,"d":"10.1214/24-AOAS1998","v":"Proceedings of the Tenth International Conference on Learning Representations (ICLR 2022)","a_":["Angelopoulos et al."],"bl":"https://doi.org/10.1214/24-AOAS1998","pc":4,"sm":[{"m":"agent draft (gemini)","l":"systems performance","tx":"Formulates Learn Then Test (LTT) calibration for finite-sample risk-controlling prediction sets and abstention thresholds."}]},{"k":"angelopoulos2023ppi","a":"ai-operations","t":"Prediction-Powered Inference","y":2023,"s":"peer_reviewed","e":4,"d":"10.1126/science.adi6000","v":"Science","a_":["Angelopoulos et al."],"bl":"https://doi.org/10.1126/science.adi6000","pc":6,"sm":[{"m":"agent draft (claude)","l":"ai platform review","tx":"Until then the dashboard should show no quality level, only uncalibrated check trends, and every rate (e.g. the ready-to-close reopen proxy, 17.3% vs 14.9%, n=810/478, difference not significant) should carry an interval."},{"m":"agent draft (claude)","l":"llmops","tx":"Prediction-powered inference combines many model predictions with a small gold-labelled set, using the labelled set to estimate and correct the predictor's bias, yielding valid confidence intervals narrower than labels alone."},{"m":"agent draft (claude)","l":"metric statistics review","tx":"Prediction-powered inference combines many model predictions with a small gold-labelled sample, using the labelled residual (rectifier) to give valid confidence intervals for population quantities."}]},{"k":"aoyang2026openrca2","a":"ai-operations","t":"OpenRCA 2.0: From Outcome Labels to Causal Process Supervision","y":2026,"s":"preprint","e":1,"x":"2606.27154","a_":["Fang et al."],"bl":"https://arxiv.org/abs/2606.27154","pc":2,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"Introduces PAVE and OpenRCA 2.0 (500 instances across 3 microservice systems and 27 fault types). Across 11 frontier LLMs, exact root-cause recovery averages 20.7%, and although agents name at least one correct service in 76.0% of cases (AnySvc), only 61.5% satisfy Path Reachability (PR), exposing a 14.5 pp ungrounded diagnosis gap."},{"m":"agent draft (claude)","l":"empirical findings","tx":"OpenRCA 2.0 adds step-wise causal annotations to 500 RCA instances; across 11 frontier models exact root-cause-set recovery is about 20.7% while about 76% name at least one correct service, and only 61.5% tie it to a verified propagation path ('ungrounded diagnosis')."}]},{"k":"argote-2011","a":"foundations-se","t":"Organizational Learning: From Experience to Knowledge","y":2011,"s":"peer_reviewed","e":2,"d":"10.1287/orsc.1100.0621","v":"Organization Science","a_":["Argote & Miron-Spektor"],"bl":"https://doi.org/10.1287/orsc.1100.0621","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Framework: organisational learning converts task experience into knowledge embedded in members, tools and routines; knowledge embedded in tools transfers and persists better than knowledge held by individuals."}]},{"k":"armbrust-2020","a":"ai-operations","t":"Delta lake","y":2020,"s":"peer_reviewed","e":4,"d":"10.14778/3415478.3415560","v":"Proceedings of the VLDB Endowment","a_":["Armbrust et al."],"bl":"https://doi.org/10.14778/3415478.3415560","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Delta transaction log with per-file stats enables data skipping before Parquet row-group skipping; verify files scanned vs total."}]},{"k":"armbrust2018structured","a":"foundations-systems","t":"Structured Streaming","y":2018,"s":"peer_reviewed","e":2,"d":"10.1145/3183713.3190664","v":"Proceedings of the 2018 International Conference on Management of Data","a_":["Armbrust et al."],"bl":"https://doi.org/10.1145/3183713.3190664","sm":[{"m":"vault note","l":"summary","tx":"Armbrust et al. (SIGMOD 2018) Structured Streaming: A Declarative API for Real-Time Applications in Apache Spark"}]},{"k":"armbrust2021lakehouse","a":"foundations-systems","t":"Lakehouse: A New Generation of Open Platforms that Unify Data Warehousing and Advanced Analytics","y":2021,"s":"peer_reviewed","e":1,"v":"11th Conference on Innovative Data Systems Research (CIDR 2021)","a_":["Armbrust et al."],"bl":"https://www.cidrdb.org/cidr2021/papers/cidr2021_paper17.pdf","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Argues for one open-format lake (Parquet plus a transactional metadata layer) serving BI and ML, with ACID commits, versioning and quality enforcement in the table layer. For the bundle lake: silver and _derived are plain Hive-partitioned Parquet with no table log, so a glob reader cannot tell a current from a stale snapshot;"}]},{"k":"arndt2017tethered","a":"foundations-telemetry","t":"Tethered to the EHR: Primary Care Physician Workload Assessment Using EHR Event Log Data and Time-Motion Observations","y":2017,"s":"peer_reviewed","e":2,"d":"10.1370/afm.2121","v":"The Annals of Family Medicine","a_":["Arndt et al."],"bl":"https://doi.org/10.1370/afm.2121","pc":1,"sm":[{"m":"agent draft (claude)","l":"measurement validity","tx":"Template for the design we lack: estimate workload from system event logs (Epic, 142 physicians, 3 years) and calibrate against a small direct time-motion observation subset (14 physicians). Found 5.9 h/day in the EHR of an 11.4 h day, 1.4 h of it after hours."}]},{"k":"artstein-2008","a":"foundations-stats","t":"Inter-Coder Agreement for Computational Linguistics","y":2008,"s":"peer_reviewed","e":4,"d":"10.1162/coli.07-034-R2","v":"Computational Linguistics","a_":["Artstein & Poesio"],"bl":"https://doi.org/10.1162/coli.07-034-R2","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Survey of agreement coefficients for coding: favours single-distribution coefficients (alpha, pi) for reliability, weighted coefficients for ordinal scales, and rejects Landis-Koch adjectives as thresholds."}]},{"k":"ascarza-2017","a":"foundations-other","t":"In Pursuit of Enhanced Customer Retention Management: Review, Key Issues, and Future Directions","y":2017,"s":"peer_reviewed","e":2,"d":"10.1007/s40547-017-0080-0","v":"Customer Needs and Solutions","a_":["Ascarza et al."],"bl":"https://doi.org/10.1007/s40547-017-0080-0","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Review of retention management: prediction, targeting, and measuring the incremental effect of retention actions."}]},{"k":"ascarza2018futility","a":"foundations-other","t":"Retention Futility: Targeting High-Risk Customers Might be Ineffective","y":2018,"s":"peer_reviewed","e":2,"d":"10.1509/jmr.16.0163","v":"Journal of Marketing Research","a_":["Ascarza"],"bl":"https://doi.org/10.1509/jmr.16.0163","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Targeting the highest churn-risk customers can be futile; target those most responsive to intervention (uplift). A support-derived risk list (stalls, repeat P1s) is decision-useful only with an intervention whose effect is measured, e.g. proactive outreach to accounts with recurrent disk stalls."},{"m":"agent draft (claude)","l":"vp support dashboard","tx":"Field experiments show customers with the highest predicted churn risk are not necessarily those most responsive to retention interventions; targeting by estimated uplift beats targeting by risk. Implication for an account-health tab: a 'risk' list is descriptive; intervention priority needs evidence of what support actions change outcomes."}]},{"k":"ascarza2018silent","a":"ai-operations","t":"Some Customers Would Rather Leave Without Saying Goodbye","y":2018,"s":"peer_reviewed","e":2,"d":"10.1287/mksc.2017.1057","v":"Marketing Science","a_":["Ascarza et al."],"bl":"https://doi.org/10.1287/mksc.2017.1057"},{"k":"athanassoulis-2016","a":"foundations-other","t":"Designing Access Methods: The RUM Conjecture","y":2016,"s":"peer_reviewed","e":2,"d":"10.5441/002/edbt.2016.42","v":"OpenProceedings.org","a_":["Athanassoulis et al."],"bl":"https://doi.org/10.5441/002/edbt.2016.42","pc":1,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Articulates the RUM Conjecture: every storage access method faces a fundamental 3-way trade-off among Read Amplification (RA), Update/Write Amplification (WA), and Memory/Space Amplification (SA), such that optimizing any two necessarily degrades the third."}]},{"k":"athey-2017","a":"econometrics","t":"The State of Applied Econometrics: Causality and Policy Evaluation","y":2017,"s":"peer_reviewed","e":5,"d":"10.1257/jep.31.2.3","v":"Journal of Economic Perspectives","a_":["Athey & Imbens"],"bl":"https://doi.org/10.1257/jep.31.2.3","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Survey of applied causal econometrics; argues for supplementary analyses (placebo tests on pseudo-outcomes and pseudo-treatments, sensitivity to unconfoundedness, robustness) beside point estimates. Applied as placebo cluster labels (recurrence), org-level shuffled release labels (relrate) and placebo intervention dates (ITS)."}]},{"k":"austin-2011","a":"foundations-stats","t":"An Introduction to Propensity Score Methods for Reducing the Effects of Confounding in Observational Studies","y":2011,"s":"peer_reviewed","e":1,"d":"10.1080/00273171.2011.568786","v":"Multivariate Behavioral Research","a_":["Austin"],"bl":"https://doi.org/10.1080/00273171.2011.568786"},{"k":"austin2016competingrisks","a":"foundations-stats","t":"Introduction to the Analysis of Survival Data in the Presence of Competing Risks","y":2016,"s":"peer_reviewed","e":1,"d":"10.1161/CIRCULATIONAHA.115.017719","v":"Circulation","a_":["Austin et al."],"bl":"https://doi.org/10.1161/CIRCULATIONAHA.115.017719","sm":[{"m":"vault note","l":"summary","tx":"Competing risks: 1-KM overstates cumulative incidence when a competing event (e.g. ticket closed without bundle) precludes the event; use Aalen-Johansen CIF"}]},{"k":"author-2008","a":"foundations-stats","t":"Handbook on Constructing Composite Indicators: Methodology and User Guide","y":2008,"s":"peer_reviewed","e":1,"d":"10.1787/9789264043466-en","v":"OECD","a_":["OECD and Joint Research Centre"],"bl":"https://doi.org/10.1787/9789264043466-en","pc":1,"sm":[{"m":"agent draft (claude)","l":"composite indicator methodology applied to support workload scores","tx":"OECD/JRC Handbook on Constructing Composite Indicators (Nardo, Saisana, Saltelli, Tarantola et al., 2008). A ten-step checklist: theoretical framework, data selection, imputation, multivariate analysis, normalisation, weighting and aggregation, uncertainty and sensitivity analysis, back to the data, links to other indicators, and visualisation."}]},{"k":"azmat2024xb","a":"sports-analytics","t":"Investigating Fouling Efficiency in Football Using Expected Booking (xB) Model","y":2024,"s":"preprint","e":1,"d":"10.48550/arXiv.2401.08718","x":"2401.08718","v":"arXiv preprint","a_":["Azmat & Yi"],"bl":"https://doi.org/10.48550/arXiv.2401.08718","pc":3,"sm":[{"m":"vault note","l":"summary","tx":"Models P(card | foul context). The correctly specified alternative to a raw fouls-per-card ratio. NOT peer-reviewed -- cite as preprint."}]},{"k":"azure2026dasv5","a":"foundations-other","t":"Dasv5 size series","y":2026,"s":"external","e":1,"a_":["Microsoft"],"bl":"https://learn.microsoft.com/en-us/azure/virtual-machines/sizes/general-purpose/dasv5-series","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"D16as_v5 uncached cap is 384 MBps vs 768 for D32as_v5. With an 8 TiB-class Premium v1 disk (P70, 750 MB/s), NewBlack's D16 VM cap binds before the disk; on D32 the disk binds. NewBlack had S4 stalls on both shapes (42.5 s on D16, 8.9 s on D32), so the VM cap alone does not explain them."}]},{"k":"azure2026diskperf","a":"foundations-systems","t":"Virtual machine and disk performance","y":2026,"s":"external","e":2,"a_":["Microsoft"],"bl":"https://learn.microsoft.com/en-us/azure/virtual-machines/disks-performance","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Azure caps uncached I/O at both the disk tier and the VM uncached limit; whichever is lower throttles and raises latency. Metrics to request: VM Uncached IOPS/Bandwidth Consumed Percentage and Data Disk IOPS/Bandwidth Consumed Percentage. Host caching ReadWrite acknowledges writes at the host cache; fsync/FUA forces flush."}]},{"k":"azure2026disktypes","a":"foundations-systems","t":"Select a disk type for Azure IaaS VMs: managed disks","y":2026,"s":"external","e":1,"a_":["Microsoft"],"bl":"https://learn.microsoft.com/en-us/azure/virtual-machines/disks-types","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Premium SSD v1 performance is tied to size tier (P10 500 IOPS/100 MB/s; P30 5,000/200; P70 18,000/750) with single-digit ms latency; Premium SSD v2 gives 3,000/125 free and sub-ms; Ultra sub-ms. In our fleet YBA Azure universes are Premium v1 (10/11 with device info), consistent with Azure's higher mean log_sync p50 (6.7 ms)."}]},{"k":"bacon2017spannersql","a":"distributed-systems","t":"Spanner","y":2017,"s":"peer_reviewed","e":2,"d":"10.1145/3035918.3056103","v":"Proceedings of the 2017 ACM International Conference on Management of Data","a_":["Bacon et al."],"bl":"https://doi.org/10.1145/3035918.3056103"},{"k":"bairavasundaram2007latent","a":"ai-operations","t":"An analysis of latent sector errors in disk drives","y":2007,"s":"peer_reviewed","e":3,"d":"10.1145/1254882.1254917","v":"Proceedings of the 2007 ACM SIGMETRICS international conference on Measurement and modeling of computer systems","a_":["Bairavasundaram et al."],"bl":"https://doi.org/10.1145/1254882.1254917","pc":1},{"k":"balkesen2013joins","a":"foundations-systems","t":"Multi-core, main-memory joins","y":2013,"s":"peer_reviewed","e":2,"d":"10.14778/2732219.2732227","v":"Proceedings of the VLDB Endowment","a_":["Balkesen et al."],"bl":"https://doi.org/10.14778/2732219.2732227","pc":1,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Balkesen et al. (VLDB '13) demonstrate that multi-core sort-merge pipelines only scale with core count when both the partitioning/run-generation phase and the final multi-way merge phase are range-partitioned across threads with NUMA/cache-aware buffers."}]},{"k":"balmau-2018","a":"distributed-systems","t":"SILK+ Preventing Latency Spikes in Log-Structured Merge Key-Value Stores Running Heterogeneous Workloads","y":2018,"s":"peer_reviewed","e":8,"d":"10.1145/3380905","v":"ACM Transactions on Computer Systems","a_":["Balmau et al."],"bl":"https://doi.org/10.1145/3380905","pc":1,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"SILK / SILK+ demonstrates that p99 tail latency in RocksDB spikes by 2-3 orders of magnitude (from sub-ms to ~1s) under sustained writes due to I/O bandwidth competition between foreground flushes/L0-L1 compactions and deep L2+ compactions (which trigger memtable/L0 write stalls)."}]},{"k":"bamberg2012oncall","a":"foundations-other","t":"Effects of On‐Call Work on Well‐Being: Results of a Daily Survey 1","y":2012,"s":"peer_reviewed","e":1,"d":"10.1111/j.1758-0854.2012.01075.x","v":"Applied Psychology: Health and Well-Being","a_":["Bamberg et al."],"bl":"https://doi.org/10.1111/j.1758-0854.2012.01075.x","pc":1,"sm":[{"m":"agent draft (claude)","l":"dashboard fix validation","tx":"Daily-diary study of on-call workers: being on call (even without being called out) is associated with worse well-being and impaired recovery the same day compared with off-call days."}]},{"k":"bansal-2020","a":"foundations-se","t":"DeCaf","y":2020,"s":"peer_reviewed","e":2,"d":"10.1145/3377813.3381353","v":"Proceedings of the ACM/IEEE 42nd International Conference on Software Engineering: Software Engineering in Practice","a_":["Bansal et al."],"bl":"https://doi.org/10.1145/3377813.3381353","pc":1,"sm":[{"m":"agent draft (claude-sonnet-5-5)","l":"complexity measurement without effort","tx":"DeCaf: diagnosing and triaging performance issues in large-scale cloud services using decision-tree/feature-based classification of telemetry. Shows that diagnostic class can be inferred from telemetry/symptom features rather than from handling history."}]},{"k":"bansal-2021","a":"foundations-se","t":"Does the Whole Exceed its Parts? The Effect of AI Explanations on Complementary Team Performance","y":2021,"s":"peer_reviewed","e":5,"d":"10.1145/3411764.3445717","v":"Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems","a_":["Bansal et al."],"bl":"https://doi.org/10.1145/3411764.3445717","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Explanations raised acceptance of AI advice whether the AI was right or wrong, so they did not produce complementary human-AI performance beyond the better of the two alone."},{"m":"agent draft (claude)","l":"support human ai","tx":"CHI 2021. Across tasks where humans and AI had comparable accuracy, AI explanations increased the rate at which people accepted AI recommendations whether or not they were correct. Explanations did not produce complementary team performance (team better than both)."}]},{"k":"barik2017readerrors","a":"foundations-se","t":"Do Developers Read Compiler Error Messages?","y":2017,"s":"peer_reviewed","e":2,"d":"10.1109/ICSE.2017.59","v":"2017 IEEE/ACM 39th International Conference on Software Engineering (ICSE)","a_":["Barik et al."],"bl":"https://doi.org/10.1109/ICSE.2017.59"},{"k":"barr2015oracle","a":"foundations-se","t":"The Oracle Problem in Software Testing: A Survey","y":2015,"s":"peer_reviewed","e":1,"d":"10.1109/TSE.2014.2372785","v":"IEEE Transactions on Software Engineering","a_":["Barr et al."],"bl":"https://doi.org/10.1109/TSE.2014.2372785","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Oracle-problem survey: implicit, derived and specified oracles."}]},{"k":"basak2023comparative","a":"ai-operations","t":"A Comparative Study of Software Secrets Reporting by Secret Detection Tools","y":2023,"s":"peer_reviewed","e":1,"d":"10.1109/esem56168.2023.10304853","v":"2023 ACM/IEEE International Symposium on Empirical Software Engineering and Measurement (ESEM)","a_":["Basak et al."],"bl":"https://doi.org/10.1109/esem56168.2023.10304853","pc":3},{"k":"basiri2016chaos","a":"ai-operations","t":"Chaos Engineering","y":2016,"s":"peer_reviewed","e":1,"d":"10.1109/MS.2016.60","v":"IEEE Software","a_":["Basiri et al."],"bl":"https://doi.org/10.1109/MS.2016.60","pc":6,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Chaos engineering principles: define steady state, hypothesise it holds, inject real-world faults, try to disprove. Adapted to a data lake: steady state is a card headline and its CI; faults are duplicates, dropped partitions, null keys, schema drift and truncation on copies of pinned inputs."}]},{"k":"bastani-2025","a":"foundations-other","t":"Generative AI without guardrails can harm learning: Evidence from high school mathematics","y":2025,"s":"peer_reviewed","e":2,"d":"10.1073/pnas.2422633122","v":"Proceedings of the National Academy of Sciences","a_":["Bastani et al."],"bl":"https://doi.org/10.1073/pnas.2422633122","pc":1,"sm":[{"m":"agent draft (claude)","l":"contrasting view","tx":"Field experiment with ~1,000 high-school students: unguarded GPT-4 access raised practice performance (+48%) but lowered later unassisted exam scores (-17%); a tutor version with guardrails avoided the harm. For support onboarding: AI notes may raise junior engineers' immediate output while slowing skill acquisition."}]},{"k":"batini2009methodologies","a":"foundations-other","t":"Methodologies for data quality assessment and improvement","y":2009,"s":"peer_reviewed","e":3,"d":"10.1145/1541880.1541883","v":"ACM Computing Surveys","a_":["Batini et al."],"bl":"https://doi.org/10.1145/1541880.1541883","pc":1,"sm":[{"m":"agent draft (claude)","l":"data quality","tx":"Comparative survey of 13 DQ assessment/improvement methodologies; consolidates dimensions (accuracy, completeness, consistency, and time-related currency/timeliness/volatility) with their metric definitions. Useful as the canonical reference for distinguishing currency (age vs real world) from timeliness (fit for the task) and volatility."}]},{"k":"bayerl2011intercoder","a":"foundations-stats","t":"What Determines Inter-Coder Agreement in Manual Annotations? A Meta-Analytic Investigation","y":2011,"s":"peer_reviewed","e":4,"d":"10.1162/COLI_a_00074","v":"Computational Linguistics","a_":["Bayerl & Paul"],"bl":"https://doi.org/10.1162/COLI_a_00074","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Meta-analysis of inter-coder agreement: agreement rises with annotator domain expertise, intensive training, fewer categories, and scheme maturity. Supports SE (expert) labellers, a calibration round, and collapsing fine scales when alpha is low."}]},{"k":"becker2019compilererrors","a":"foundations-se","t":"Compiler Error Messages Considered Unhelpful","y":2019,"s":"peer_reviewed","e":3,"d":"10.1145/3344429.3372508","v":"Proceedings of the Working Group Reports on Innovation and Technology in Computer Science Education","a_":["Becker et al."],"bl":"https://doi.org/10.1145/3344429.3372508","pc":1,"sm":[{"m":"agent draft (claude)","l":"bibliographic context","tx":"ITiCSE 2019 working-group landscape review of text-based programming error-message research. Consolidates design guidelines (readability, specificity, actionable fix hints, context, no blame) and shows evidence for enhanced messages is mixed and measurement-dependent (time-to-fix, repeated-error rates, population)."}]},{"k":"begg1983verification","a":"foundations-stats","t":"Assessment of Diagnostic Tests When Disease Verification is Subject to Selection Bias","y":1983,"s":"peer_reviewed","e":2,"d":"10.2307/2530820","v":"Biometrics","a_":["Begg & Greenes"],"bl":"https://doi.org/10.2307/2530820","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Begg & Greenes show that when the reference standard is verified only for a subset selected on the test result, naive sensitivity/specificity are biased, and that the bias is correctable when verification depends only on observed covariates (MAR) by inverse-verification-probability reweighting."}]},{"k":"beller2015idetesting","a":"foundations-se","t":"When, how, and why developers (do not) test in their IDEs","y":2015,"s":"peer_reviewed","e":1,"d":"10.1145/2786805.2786843","v":"Proceedings of the 2015 10th Joint Meeting on Foundations of Software Engineering","a_":["Beller et al."],"bl":"https://doi.org/10.1145/2786805.2786843","pc":1,"sm":[{"m":"agent draft (claude)","l":"measurement validity","tx":"WatchDog IDE telemetry on 416 developers (2,443 in the TSE extension) versus their own estimates: developers believed they spent about half their time testing but telemetry showed about a quarter, a twofold self-report overestimate."}]},{"k":"bendeck-2025","a":"ai-operations","t":"An Empirical Evaluation of the GPT-4 Multimodal Language Model on Visualization Literacy Tasks","y":2025,"s":"peer_reviewed","e":3,"d":"10.1109/TVCG.2024.3456155","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Bendeck & Stasko"],"bl":"https://doi.org/10.1109/TVCG.2024.3456155"},{"k":"benjamini1995fdr","a":"ai-operations","t":"Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing","y":1995,"s":"peer_reviewed","e":2,"d":"10.1111/j.2517-6161.1995.tb02031.x","v":"Journal of the Royal Statistical Society. Series B (Methodological)","a_":["Benjamini & Hochberg"],"bl":"https://doi.org/10.1111/j.2517-6161.1995.tb02031.x","pc":3},{"k":"benneyan-2003","a":"foundations-stats","t":"Statistical process control as a tool for research and healthcare improvement","y":2003,"s":"peer_reviewed","e":1,"d":"10.1136/qhc.12.6.458","v":"Quality and Safety in Health Care","a_":["Benneyan"],"bl":"https://doi.org/10.1136/qhc.12.6.458","pc":1,"sm":[{"m":"agent draft (claude)","l":"metric statistics review","tx":"Practical guide to SPC in healthcare: choose chart by data type (c/u for counts, p for proportions, XmR for continuous or when distributional assumptions fail), use about 20-25 subgroups to set limits, and read special-cause rules rather than reacting to single points."}]},{"k":"berenson-1995","a":"distributed-systems","t":"A Critique of ANSI SQL Isolation Levels.","y":1995,"s":"peer_reviewed","e":2,"d":"10.1145/223784.223785","v":"Proceedings of the 1995 ACM SIGMOD International Conference on Management of Data","a_":["Berenson & O'Neil"],"bl":"https://doi.org/10.1145/223784.223785","pc":2,"sm":[{"m":"agent draft (gemini)","l":"systems architecture","tx":"Demonstrates that the ANSI SQL-92 phenomenon definitions (dirty read, non-repeatable read, phantom) are ambiguous, formalizes Snapshot Isolation (SI), and characterizes write-skew anomalies."}]},{"k":"bertini-2020","a":"foundations-viz","t":"Why Shouldn’t All Charts Be Scatter Plots? Beyond Precision-Driven Visualizations","y":2020,"s":"peer_reviewed","e":3,"d":"10.1109/VIS47514.2020.00048","v":"2020 IEEE Visualization Conference (VIS)","a_":["Bertini et al."],"bl":"https://doi.org/10.1109/VIS47514.2020.00048"},{"k":"berzak2016anchoring","a":"foundations-stats","t":"Anchoring and Agreement in Syntactic Annotations","y":2016,"s":"preprint","e":1,"x":"1605.04481","v":"arXiv:1605.04481","a_":["Berzak et al."],"bl":"https://arxiv.org/abs/1605.04481","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Shows that annotators who edit parser pre-annotations anchor on them: agreement with the pre-annotation is inflated and human-human agreement is overestimated versus annotation from scratch. Direct evidence that showing AI pre-labels to human raters contaminates human-AI agreement estimates."}]},{"k":"beschastnikh2011synoptic","a":"ai-operations","t":"Leveraging existing instrumentation to automatically infer invariant-constrained models","y":2011,"s":"peer_reviewed","e":4,"d":"10.1145/2025113.2025151","v":"Proceedings of the 19th ACM SIGSOFT symposium and the 13th European conference on Foundations of software engineering","a_":["Beschastnikh et al."],"bl":"https://doi.org/10.1145/2025113.2025151","pc":2,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Beschastnikh et al. (ESEC/FSE 2011) infer temporal Ordering invariants (`a always precedes b`, `a never followed by b`) from system logs in Synoptic."}]},{"k":"bettenburg2008goodbug","a":"ai-operations","t":"What Makes a Good Bug Report?","y":2008,"s":"peer_reviewed","e":11,"d":"10.1145/1453101.1453146","v":"Proceedings of the 16th ACM SIGSOFT International Symposium on Foundations of Software Engineering (FSE '08)","a_":["Bettenburg et al."],"bl":"https://doi.org/10.1145/1453101.1453146","pc":5,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Survey of 466 Apache/Eclipse/Mozilla developers and reporters: developers rank steps to reproduce, stack traces and test cases as most useful, while these are the items reporters find hardest to provide; the mismatch (not reporter effort) is the core quality gap. CUEZILLA scored report quality from text features."},{"m":"agent draft (gemini)","l":"T0 support ticket completeness & reporter developer mismatch","tx":"Empirical survey of Apache, Eclipse, and Mozilla developers and users quantifying the severe mismatch between what reporters provide at T0 and what engineers need to resolve issues rapidly (steps to reproduce, stack traces, observed vs. expected behavior, and exact environment/version metadata)."},{"m":"agent draft (claude)","l":"empirical findings","tx":"Journal extension of Bettenburg et al. 2008: developer and reporter surveys; wrong or incomplete information (missing steps, versions, stack traces) is the most cited problem; a supervised model on report features predicts developer-perceived quality. Supports treating intake completeness as a measurable, predictive ticket attribute;"}]},{"k":"bevan2006targets","a":"foundations-se","t":"WHAT’S MEASURED IS WHAT MATTERS: TARGETS AND GAMING IN THE ENGLISH PUBLIC HEALTH CARE SYSTEM","y":2006,"s":"peer_reviewed","e":4,"d":"10.1111/j.1467-9299.2006.00600.x","v":"Public Administration","a_":["BEVAN & HOOD"],"bl":"https://doi.org/10.1111/j.1467-9299.2006.00600.x","pc":3,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Hard-threshold targets (NHS 4-hour, ambulance 8-minute) induced gaming: reclassification, clock-boundary manipulation, hitting the target and missing the point. A support SLA defined as a one-tap PagerDuty ack is exposed to the same mechanism; counter-measure is a paired substantive-engagement metric and audits of the ack-to-reply gap."},{"m":"agent draft (claude)","l":"dashboard fix validation","tx":"Documents gaming under English NHS performance targets (e.g. 4-hour A&E, ambulance response): reclassification, threshold effects, manipulating when the clock starts, hitting the target while missing the point. Typology: ratchet, threshold and output-distortion effects."},{"m":"agent draft (claude)","l":"support human ai","tx":"Documents gaming of NHS performance targets: ratchet effects, threshold effects and output distortion ('hitting the target and missing the point'), worsened when targets carry sanctions. Support implication: if TTR or AI-close rates become targets, expect premature closes, ticket splitting and threshold clustering."}]},{"k":"beyer2016sre","a":"ai-operations","t":"Site Reliability Engineering: How Google Runs Production Systems","y":2016,"s":"external","e":3,"v":"O'Reilly Media","a_":["Beyer et al."],"bl":"https://sre.google/sre-book/monitoring-distributed-systems/","pc":7,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Ch.12 Effective Troubleshooting: triage, examine, diagnose, test/treat; ask what changed (recent changes are the most common cause), what, where, why; correlation is not causation; negative results are valuable. For support tooling: the case-open panel should answer what/where/when and what changed, and every empty answer must say whether data was absent."},{"m":"agent draft (claude)","l":"support ops dashboard","tx":"SRE book chapters on SLOs (ch.4: use percentiles, not means; SLO as attainment over a window), Being On-Call (ch.11: response times of 5 min for user-facing vs 30 min for less critical; on-call load limits, max ~2 incidents per 12h shift; dual-region follow-the-sun), and Monitoring (ch.6)."}]},{"k":"beyer2018workbook","a":"ai-operations","t":"The Site Reliability Workbook: Practical Ways to Implement SRE","y":2018,"s":"external","e":5,"v":"O'Reilly Media","a_":["Beyer et al."],"bl":"https://sre.google/workbook/alerting-on-slos/","pc":8,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"PM seat (council 5): Error budgets need a policy and multiwindow burn-rate alerts; GF1 has a per-event bound and VOLUME.md proposes a 30-day budget, but the fs_asks card still adopts GF1/GF2 only as acceptance criteria with no budget policy."},{"m":"agent draft (claude)","l":"ai platform review","tx":"Ch.2/5: an SLO needs a persisted SLI time series, a target, a window and an error budget; page on multi-window multi-burn-rate (14.4x 1h/5m, 6x 6h/30m), ticket on slow burn."},{"m":"agent draft (claude)","l":"support ops review","tx":"SRE workbook: SLO-based alerting with error budgets and burn rates; on-call chapter recommends bounded load (e.g. few incidents per shift), handoff hygiene, and dry runs. Applied to support: express P1 response as an SLO with error budget, alert on burn during events like BFCM, and test escalation paths (Premium15 test page, dry run) before the window."}]},{"k":"bhuiyan2025sokredos","a":"foundations-se","t":"SoK: A Literature and Engineering Review of Regular Expression Denial of Service (ReDoS)","y":2025,"s":"peer_reviewed","e":3,"d":"10.1145/3708821.3733912","v":"Proceedings of the 20th ACM Asia Conference on Computer and Communications Security","a_":["Bhuiyan et al."],"bl":"https://doi.org/10.1145/3708821.3733912","pc":1,"sm":[{"m":"agent draft (gemini)","l":"security governance","tx":"Systematizes 93 papers on Regular Expression Denial of Service (ReDoS), contrasting Spencer-style backtracking NFA engines (Python `re`, PCRE, Java, JavaScript) that exhibit worst-case O(2^N) or polynomial O(N^k) super-linear complexity with Thompson NFA / lazy-DFA engines (Rust `regex`, Google RE2) that guarantee O(N) time complexity by tracking…"}]},{"k":"bird2009fair","a":"ai-operations","t":"Fair and balanced?","y":2009,"s":"peer_reviewed","e":2,"d":"10.1145/1595696.1595716","v":"Proceedings of the 7th joint meeting of the European software engineering conference and the ACM SIGSOFT symposium on The foundations of software engineering","a_":["Bird et al."],"bl":"https://doi.org/10.1145/1595696.1595716","pc":4,"sm":[{"m":"agent draft (claude)","l":"limitations caveats","tx":"Linked bug-fix samples are biased (by severity, experience); fixtest finds only 37% (DB 25%) of field-linked issues in commit messages, so its shares describe the linked subset."}]},{"k":"blackburn-2006","a":"ai-operations","t":"The DaCapo benchmarks","y":2006,"s":"peer_reviewed","e":5,"d":"10.1145/1167473.1167488","v":"Proceedings of the 21st annual ACM SIGPLAN conference on Object-oriented programming systems, languages, and applications","a_":["Blackburn et al."],"bl":"https://doi.org/10.1145/1167473.1167488"},{"k":"blackburn-2025","a":"ai-operations","t":"Rethinking Java Performance Analysis","y":2025,"s":"peer_reviewed","e":3,"d":"10.1145/3669940.3707217","v":"Proceedings of the 30th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 1","a_":["Blackburn et al."],"bl":"https://doi.org/10.1145/3669940.3707217","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Introduces DaCapo Chopin and argues Java performance analysis must report heap size (as a multiple of minimum heap), user-experienced latency (simple and metered), and total GC cost rather than pause time; demonstrates misleading conclusions from fixed-heap, single-metric evaluations."}]},{"k":"blelloch1989scans","a":"foundations-systems","t":"Scans as primitive parallel operations","y":1989,"s":"peer_reviewed","e":1,"d":"10.1109/12.42122","v":"IEEE Transactions on Computers","a_":["Blelloch"],"bl":"https://doi.org/10.1109/12.42122","pc":1,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Blelloch (1989/1990) establishes that any sequential recurrence driven by an associative binary operator (including segmented accumulations, stream compaction, and first-order recurrences) reduces to a work-efficient parallel prefix scan in O(N) work and O(log N) depth."}]},{"k":"bloom1970space","a":"foundations-systems","t":"Space/time trade-offs in hash coding with allowable errors","y":1970,"s":"peer_reviewed","e":7,"d":"10.1145/362686.362692","v":"Communications of the ACM","a_":["Bloom"],"bl":"https://doi.org/10.1145/362686.362692"},{"k":"borkin2013memorable","a":"foundations-viz","t":"What Makes a Visualization Memorable?","y":2013,"s":"peer_reviewed","e":3,"d":"10.1109/TVCG.2013.234","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Borkin et al."],"bl":"https://doi.org/10.1109/TVCG.2013.234","pc":1,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"Evaluates 2,070 single-panel visualizations across 261 participants and establishes that intrinsic visualization memorability is consistent across observers (Spearman rank correlation r=0.75)."}]},{"k":"borkin2016beyond","a":"foundations-viz","t":"Beyond Memorability: Visualization Recognition and Recall","y":2016,"s":"peer_reviewed","e":4,"d":"10.1109/TVCG.2015.2467732","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Borkin et al."],"bl":"https://doi.org/10.1109/TVCG.2015.2467732","pc":1,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Analyzes eye-tracking fixations and thousands of free-text recall descriptions across 393 visualizations. Proves that readers fixate first on the title and direct data labels, and that message redundancy (reinforcing the core quantitative contrast in the title, direct bar callouts, and takeaway banner) directly increases accurate recall without…"}]},{"k":"borland-2007","a":"foundations-viz","t":"Rainbow Color Map (Still) Considered Harmful","y":2007,"s":"peer_reviewed","e":2,"d":"10.1109/MCG.2007.323435","v":"IEEE Computer Graphics and Applications","a_":["Borland & Ii"],"bl":"https://doi.org/10.1109/MCG.2007.323435"},{"k":"bornholt2021s3","a":"ai-operations","t":"Using Lightweight Formal Methods to Validate a Key-Value Storage Node in Amazon S3","y":2021,"s":"peer_reviewed","e":2,"d":"10.1145/3477132.3483540","v":"Proceedings of the ACM SIGOPS 28th Symposium on Operating Systems Principles","a_":["Bornholt et al."],"bl":"https://doi.org/10.1145/3477132.3483540","pc":1,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Demonstrates lightweight formal specifications in Rust paired with property-based conformance checking to validate Amazon S3's ShardStore key-value storage node (SOSP 2021). Directly parallels Report 08's architecture of checking production storage/log-engine implementations against executable reference models."}]},{"k":"borovkov2026selective","a":"ai-operations","t":"Learning Selective LLM Autonomy from Copilot Feedback in Enterprise Customer Support Workflows","y":2026,"s":"preprint","e":1,"x":"2604.23855","v":"arXiv:2604.23855","a_":["Borovkov et al."],"bl":"https://arxiv.org/abs/2604.23855","pc":2,"sm":[{"m":"agent draft (claude)","l":"support human ai","tx":"Deployed enterprise customer-support BPM system that uses copilot accept/correct feedback and UI traces to train a policy plus an abstention critic. It auto-executes only high-confidence steps and hands uncertain ones to operators. Reported 45% of sessions automated and 39% lower average handling time without degraded quality (authors' claim)."},{"m":"agent draft (claude)","l":"eval kpis","tx":"Deployed enterprise customer-support system that learns a next-action policy and an abstention critic from operator accept/correct copilot feedback; in live A/B tests it automated ~45% of sessions and cut average handling time ~39% without degrading quality."}]},{"k":"borst-2004","a":"foundations-telemetry","t":"Dimensioning Large Call Centers","y":2004,"s":"peer_reviewed","e":4,"d":"10.1287/opre.1030.0081","v":"Operations Research","a_":["Borst et al."],"bl":"https://doi.org/10.1287/opre.1030.0081","pc":1,"sm":[{"m":"agent draft (claude)","l":"support ops dashboard","tx":"Derives asymptotically optimal staffing (square-root safety staffing) trading off staffing cost against delay cost, with ED/QD/QED regimes. Relevance: formalizes why very tight response targets (e.g., 5-min P1) imply over-provisioning relative to load in small systems; supports framing 5 vs 15 min P1 targets as cost-of-capacity decisions."}]},{"k":"bortnikov-2018","a":"foundations-systems","t":"Accordion","y":2018,"s":"peer_reviewed","e":4,"d":"10.14778/3229863.3229873","v":"Proceedings of the VLDB Endowment","a_":["Bortnikov et al."],"bl":"https://doi.org/10.14778/3229863.3229873","pc":1,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"Accordion applies LSM principles to the in-memory MemTable layer via proactive in-memory flushes and compactions into flat arrays, shrinking memory footprint, allocator fragmentation, and flush frequency (reducing Disk Write Amplification by 20-50%)."}]},{"k":"borusyak2024imputation","a":"ai-operations","t":"Revisiting Event-Study Designs: Robust and Efficient Estimation","y":2024,"s":"peer_reviewed","e":2,"d":"10.1093/restud/rdae007","v":"The Review of Economic Studies","a_":["Borusyak et al."],"bl":"https://doi.org/10.1093/restud/rdae007","pc":1},{"k":"bouzeghoub2004freshness","a":"foundations-other","t":"A framework for analysis of data freshness","y":2004,"s":"peer_reviewed","e":1,"d":"10.1145/1012453.1012464","v":"Proceedings of the 2004 international workshop on Information quality in information systems","a_":["Bouzeghoub"],"bl":"https://doi.org/10.1145/1012453.1012464","pc":1,"sm":[{"m":"agent draft (claude)","l":"data quality","tx":"Taxonomy of freshness: currency (time since source change not reflected) vs timeliness (age of data relative to update frequency), and propagation of freshness through integration pipelines."}]},{"k":"boyer1977fast","a":"foundations-systems","t":"A fast string searching algorithm","y":1977,"s":"peer_reviewed","e":3,"d":"10.1145/359842.359859","v":"Communications of the ACM","a_":["Boyer & Moore"],"bl":"https://doi.org/10.1145/359842.359859","pc":1,"sm":[{"m":"agent draft (gemini)","l":"systems performance","tx":"Foundational sub-linear string searching algorithm proving that right-to-left pattern comparison with bad-character shifts skips large fractions of text bytes."}]},{"k":"boyko2007referee","a":"sports-analytics","t":"Referee Bias Contributes to Home Advantage in English Premiership Football","y":2007,"s":"peer_reviewed","e":2,"d":"10.1080/02640410601038576","v":"Journal of Sports Sciences","a_":["Boyko et al."],"bl":"https://doi.org/10.1080/02640410601038576","pc":1,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"Separates referee-specific disciplinary bias from crowd-induced home advantage across 5,244 English Premier League matches using fixed-effects regression on yellow cards, red cards, and penalties."}]},{"k":"breck2017mltestscore","a":"foundations-other","t":"The ML test score: A rubric for ML production readiness and technical debt reduction","y":2017,"s":"peer_reviewed","e":4,"d":"10.1109/BigData.2017.8258038","v":"2017 IEEE International Conference on Big Data (Big Data)","a_":["Breck et al."],"bl":"https://doi.org/10.1109/BigData.2017.8258038","pc":3,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Readiness rubric over data, model, infra, monitoring."},{"m":"agent draft (claude)","l":"ai platform review","tx":"Score roughly 2 of 7 monitoring items fully met."},{"m":"agent draft (claude)","l":"llmops","tx":"28-item rubric across data, model, infrastructure and monitoring tests; the monitoring section asks for dependency-change alerts, training/serving skew, data invariants, staleness ('model is not too stale'), NaN/infra checks and regressions in prediction quality and latency."}]},{"k":"breck2019datavalidation","a":"foundations-other","t":"Data Validation for Machine Learning","y":2019,"s":"peer_reviewed","e":3,"v":"Proceedings of Machine Learning and Systems (MLSys / SysML 2019)","a_":["Breck et al."],"bl":"https://mlsys.org/Conferences/2019/doc/2019/167.pdf","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Google data validation: schema-based checks per batch, alerting on drift, data treated like code. Motivates per-table contracts for silver/_derived (key uniqueness, timestamps inside the bundle window, schema version) run server-side before the gallery export."},{"m":"agent draft (gemini)","l":"operational translation","tx":"Describes Google's production ML data validation system (TensorFlow Data Validation / TFDV), treating data schemas as code-checked contracts that enforce single-batch invariants, cross-batch distribution drift bounds ($L_\\infty$ distance), and training-serving skew detection across billions of examples per day."}]},{"k":"brettel-1997","a":"foundations-viz","t":"Computerized simulation of color appearance for dichromats","y":1997,"s":"peer_reviewed","e":2,"d":"10.1364/JOSAA.14.002647","v":"Journal of the Optical Society of America A","a_":["Brettel et al."],"bl":"https://doi.org/10.1364/JOSAA.14.002647"},{"k":"breu2010infoneeds","a":"ai-operations","t":"Information needs in bug reports","y":2010,"s":"peer_reviewed","e":3,"d":"10.1145/1718918.1718973","v":"Proceedings of the 2010 ACM conference on Computer supported cooperative work","a_":["Breu et al."],"bl":"https://doi.org/10.1145/1718918.1718973","pc":2,"sm":[{"m":"agent draft (gemini)","l":"Cost of follow up questions in support/issue threads","tx":"Analyzes 600 bug reports from Mozilla and Eclipse to categorize developer follow-up questions across 8 information need categories, proving that when developers must ask follow-up questions for missing T0 context, issue resolution time roughly doubles and many questions go unanswered or delayed."}]},{"k":"brewer-2000","a":"distributed-systems","t":"Towards Robust Distributed Systems.","y":2000,"s":"peer_reviewed","e":2,"d":"10.1145/564585.564601","v":"Keynote at the 19th Annual ACM Symposium on Principles of Distributed Computing (PODC)","a_":["Brewer"],"bl":"https://doi.org/10.1145/564585.564601","pc":1,"sm":[{"m":"vault note","l":"summary","tx":"Brewer, E. A. (2000). \"`Towards Robust Distributed Systems.`\" _Keynote at the 19th Annual ACM Symposium on Principles of Distributed Computing (PODC)_; formalized by Gilbert, S. and Lynch, N. (2002), \"`Brewer's Conjecture and the Feasibility of Consistent, Available, Partition-Tolerant Web Services,`\" _ACM SIGACT News_, 33(2), 51–59."}]},{"k":"breznau2022manyanalysts","a":"foundations-stats","t":"Observing many researchers using the same data and hypothesis reveals a hidden universe of uncertainty","y":2022,"s":"peer_reviewed","e":2,"d":"10.1073/pnas.2203150119","v":"Proceedings of the National Academy of Sciences","a_":["Breznau et al."],"bl":"https://doi.org/10.1073/pnas.2203150119","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"73 teams on one hypothesis: results spanned significant negative to significant positive, and measured analyst choices explained little of the variance (hidden universe of uncertainty). Supports treating single-spec fleet claims (e.g. 'imbalance does not predict CE') as uninformative unless the spec curve is tight."}]},{"k":"broadwell2003scrash","a":"ai-operations","t":"Scrash: A System for Generating Secure Crash Information","y":2003,"s":"peer_reviewed","e":1,"v":"12th USENIX Security Symposium","a_":["Broadwell et al."],"bl":"https://www.usenix.org/conference/12th-usenix-security-symposium/scrash-system-generating-secure-crash-information","pc":6,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Broadwell, Harren, and Sastry (USENIX Security 2003) demonstrate in Scrash that raw core dumps indiscriminately mix sensitive application heap data (passwords, keys, customer rows) with non-sensitive execution state (call stacks, symbol tables, control registers), and show how to strip sensitive memory regions prior to crash transmission."},{"m":"agent draft (gemini)","l":"security and privacy","tx":"Demonstrates that raw core dumps expose arbitrary process memory secrets (keys, passwords, PII); motivates yb-collector's default exclusion of raw core files in favor of pure-Rust/Go ELF64 NT_FILE inspection + on-node read-only lldb/gdb thread backtrace extraction (~30 KB)."}]},{"k":"brodie2005symptom","a":"foundations-se","t":"Quickly Finding Known Software Problems via Automated Symptom Matching","y":2005,"s":"peer_reviewed","e":7,"d":"10.1109/icac.2005.49","v":"Second International Conference on Autonomic Computing (ICAC'05)","a_":["Brodie et al."],"bl":"https://doi.org/10.1109/icac.2005.49","pc":4,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Brodie et al. (IBM, ICAC'05) match symptoms from logs/stack traces against a database of known problems to find known issues quickly; precision depends on curated, versioned symptom definitions."},{"m":"agent draft (claude)","l":"operational translation","tx":"IBM's symptom-database approach: match call stacks and log symptoms from a new problem against a curated database of known problems, so that recurrences are recognised without re-diagnosis."},{"m":"agent draft (gemini)","l":"operational translation","tx":"Brodie et al. (ICAC 2005) formalize automated symptom matching for known software defects by normalizing raw call stacks (stripping recursive loops, signal-handler boilerplate, and memory offsets) and applying weighted sequence alignment that prioritizes frames closest to the faulting instruction pointer."}]},{"k":"brooks1987nosilverbullet","a":"foundations-se","t":"No Silver Bullet Essence and Accidents of Software Engineering","y":1987,"s":"peer_reviewed","e":2,"d":"10.1109/MC.1987.1663532","v":"Computer","a_":["Brooks"],"bl":"https://doi.org/10.1109/MC.1987.1663532","pc":2,"sm":[{"m":"agent draft (gemini)","l":"systems architecture","tx":"Brooks (IEEE Computer 1987) distinguishes essential complexity (inherent domain logic: YugabyteDB Raft consensus, DocDB compaction, YSQL query planning, Bayesian System-1 calibration, and LLM orchestration) from accidental complexity (version skew across 4 worktrees, 4 forked Finding schemas, 3 regex redactors, and OOM-prone multi-pass tarball unpacking)."},{"m":"agent draft (gemini)","l":"essential vs accidental complexity","tx":"Brooks (1987) separates essential complexity (the domain rules, diagnostic physics, Bayesian priors, and YugabyteDB distributed consensus invariants) from accidental complexity (Python GIL contention, O(N^2) string scans, regex backtracking, and packaging drift)."}]},{"k":"brown-2005","a":"foundations-stats","t":"Statistical Analysis of a Telephone Call Center","y":2005,"s":"peer_reviewed","e":9,"d":"10.1198/016214504000001808","v":"Journal of the American Statistical Association","a_":["Brown et al."],"bl":"https://doi.org/10.1198/016214504000001808","pc":1,"sm":[{"m":"agent draft (claude)","l":"support ops dashboard","tx":"Empirical study of an Israeli bank call center: arrivals are time-inhomogeneous Poisson with overdispersed day-level rates; service times are approximately lognormal, not exponential; patience/abandonment estimated with survival methods for censored data."}]},{"k":"brown2001binomial","a":"ai-operations","t":"Interval Estimation for a Binomial Proportion","y":2001,"s":"peer_reviewed","e":1,"d":"10.1214/ss/1009213286","v":"Statistical Science","a_":["Brown et al."],"bl":"https://doi.org/10.1214/ss/1009213286","pc":4,"sm":[{"m":"agent draft (claude)","l":"metric statistics review","tx":"Shows the Wald interval for a binomial proportion has erratic, often badly low coverage even at moderate n and especially near 0 or 1, and recommends Wilson or Jeffreys for n<40 and Agresti-Coull for larger n. Applied to support SLA tiles: per-customer P1 attainment (n=1-10 pages) and BFCM event windows need Wilson/Jeffreys bounds;"},{"m":"agent draft (claude)","l":"support ops dashboard","tx":"Shows Wald intervals are erratic for binomial proportions and recommends Wilson/Agresti-Coull/Jeffreys. For SLA attainment and CSAT with small monthly n (e.g., 27/28 P1 acks within target gives Wilson 95% CI ~82-99%), dashboards should show Wilson intervals, not bare percentages."}]},{"k":"brundage2020trustworthy","a":"security-governance","t":"Toward Trustworthy AI Development: Mechanisms for Supporting Verifiable Claims","y":2020,"s":"preprint","e":3,"x":"2004.07213","v":"arXiv:2004.07213","a_":["Brundage et al."],"bl":"https://arxiv.org/abs/2004.07213","pc":1,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Introduces the taxonomy of institutional mechanisms (red-teaming, third-party auditing, incident sharing, bounties), software mechanisms (audit trails, interpretability, privacy-preserving ML, reproducible pipelines), and hardware mechanisms (compute accounting, secure enclaves) needed to turn high-level AI safety/governance assertions into…"}]},{"k":"brynjolfsson-2025","a":"ai-operations","t":"Generative AI at Work.","y":2025,"s":"external","e":10,"v":"The Quarterly Journal of Economics","a_":["Brynjolfsson & Raymond"],"pc":8,"sm":[{"m":"agent draft (claude)","l":"Applied AI leader","tx":"Staggered rollout of a GPT-based assistant to 5,179 support agents. Resolutions per hour rose about 15% on average and about 30% for novices; top performers saw no speed gain and a small quality decline. Customer sentiment improved and attrition fell."},{"m":"agent draft (claude)","l":"empirical findings","tx":"Staggered rollout of an LLM conversational assistant to about 5,000 customer-support agents. Productivity (resolutions per hour) rose about 14% on average and about 34% for novice and low-skilled workers, with little effect on the most experienced. Customer sentiment improved and attrition fell."},{"m":"agent draft (claude)","l":"limitations caveats","tx":"Field evidence (5,179 chat agents, staggered rollout): an LLM assistant raised resolutions per hour ~14% on average and ~34% for novices, little for top performers."}]},{"k":"buell-2017","a":"foundations-other","t":"Creating Reciprocal Value Through Operational Transparency","y":2017,"s":"peer_reviewed","e":1,"d":"10.1287/mnsc.2015.2411","v":"Management Science","a_":["Buell et al."],"bl":"https://doi.org/10.1287/mnsc.2015.2411","pc":1,"sm":[{"m":"agent draft (claude)","l":"support ops review","tx":"Field and lab experiments (cafeteria cooks/diners) show transparency works in both directions: customers who see the work value it more, and workers who see the customer exert more effort and produce higher quality. For a support dashboard, an automated 'acknowledged by engineer X' relay is one-way transparency;"}]},{"k":"buneman-2001","a":"foundations-systems","t":"Why and Where: A Characterization of Data Provenance","y":2001,"s":"peer_reviewed","e":2,"d":"10.1007/3-540-44503-X_20","v":"Lecture Notes in Computer Science","a_":["Buneman et al."],"bl":"https://doi.org/10.1007/3-540-44503-X_20","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Distinguishes why-provenance (which inputs justify an output) from where-provenance (where a value was copied from). The _derived marker objects list is where-provenance at file granularity; gallery numbers lack why-provenance (which bundle set and file generations produced a card)."}]},{"k":"burchell-2023","a":"ai-operations","t":"Don’t Trust Your Profiler: An Empirical Study on the Precision and Accuracy of Java Profilers","y":2023,"s":"peer_reviewed","e":3,"d":"10.1145/3617651.3622985","v":"Proceedings of the 20th ACM SIGPLAN International Conference on Managed Programming Languages and Runtimes","a_":["Burchell et al."],"bl":"https://doi.org/10.1145/3617651.3622985","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Revisits Mytkowicz PLDI'10 with modern samplers (async-profiler, Honest Profiler, JFR, YourKit, JProfiler, VisualVM) on DaCapo/Renaissance. Profilers are fairly self-consistent across ~30 runs and usually agree on the top method, but disagree beyond it and on per-method shares."}]},{"k":"burchell-2025","a":"ai-operations","t":"Divining Profiler Accuracy: An Approach to Approximate Profiler Accuracy through Machine Code-Level Slowdown","y":2025,"s":"peer_reviewed","e":4,"d":"10.1145/3763180","v":"Proceedings of the ACM on Programming Languages","a_":["Burchell & Marr"],"bl":"https://doi.org/10.1145/3763180"},{"k":"burns2023weakstrong","a":"ai-operations","t":"Weak-to-Strong Generalization: Eliciting Strong Capabilities With Weak Supervision","y":2023,"s":"preprint","e":1,"x":"2312.09390","v":"arXiv:2312.09390","a_":["Burns et al."],"bl":"https://arxiv.org/abs/2312.09390","pc":1,"sm":[{"m":"agent draft (qwen)","l":"operational translation","tx":"Weak supervisor (GPT-2-class) finetuning a much stronger model only partially elicits its capability (logistic-regression-style gap): the weak model's labels define a ceiling on the strong model's behavior if it is trained/selected on them."}]},{"k":"buçinca-2021","a":"foundations-se","t":"To Trust or to Think","y":2021,"s":"peer_reviewed","e":4,"d":"10.1145/3449287","v":"Proceedings of the ACM on Human-Computer Interaction","a_":["Buçinca et al."],"bl":"https://doi.org/10.1145/3449287","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Cognitive forcing functions (asking people to decide before seeing the AI, or adding a wait) reduce over-reliance on wrong AI advice more than explanations do, but people like them least; benefits are largest for people high in need for cognition."},{"m":"agent draft (claude)","l":"support human ai","tx":"CSCW 2021. Cognitive forcing designs (decide before seeing the AI, wait, AI on demand) reduced over-reliance on incorrect AI advice compared with simple explainable-AI displays, but users rated them less preferred. Benefits skewed toward people high in need for cognition."}]},{"k":"bystrom-1995","a":"foundations-other","t":"Task complexity affects information seeking and use","y":1995,"s":"peer_reviewed","e":3,"d":"10.1016/0306-4573(95)80035-R","v":"Information Processing &amp; Management","a_":["Byström & Järvelin"],"bl":"https://doi.org/10.1016/0306-4573(95)80035-R","pc":1,"sm":[{"m":"agent draft (claude-sonnet-5-5)","l":"construct separation","tx":"Defines task complexity by a priori determinability: how far the worker can know, before starting, the outcome, the process and the information needed. Five task categories from automatic processing to genuine decision tasks. Higher complexity makes people consult more sources and more general ones."}]},{"k":"cahill2008ssi","a":"foundations-systems","t":"Serializable isolation for snapshot databases","y":2008,"s":"peer_reviewed","e":2,"d":"10.1145/1376616.1376690","v":"Proceedings of the 2008 ACM SIGMOD international conference on Management of data","a_":["Cahill et al."],"bl":"https://doi.org/10.1145/1376616.1376690","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"SSI detects dangerous rw-antidependency structures and aborts with serialization failures. For bundle data: 40001 is an expected, design-level outcome of optimistic concurrency, so a 40001 count is only interpretable as a rate per transaction and alongside retry handling; raw presence (56 of 76 audited cases) says little."}]},{"k":"cai-2022","a":"ai-operations","t":"Distilling the Real Cost of Production Garbage Collectors","y":2022,"s":"peer_reviewed","e":1,"d":"10.1109/ISPASS55109.2022.00005","v":"2022 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS)","a_":["Cai et al."],"bl":"https://doi.org/10.1109/ISPASS55109.2022.00005","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Proposes LBO (lower-bound overhead): compare total CPU/wall cost of a production GC against an idealised baseline derived from the cheapest collector per benchmark, capturing barrier, allocation and mutator-locality costs that pause-time metrics miss. Finds concurrent collectors (ZGC, Shenandoah) carry substantial total-cost overhead at tight heaps."}]},{"k":"callaway2021did","a":"econometrics","t":"Difference-in-Differences with Multiple Time Periods","y":2021,"s":"peer_reviewed","e":10,"d":"10.1016/j.jeconom.2020.12.001","v":"Journal of Econometrics","a_":["Callaway & Sant'Anna"],"bl":"https://doi.org/10.1016/j.jeconom.2020.12.001","pc":2,"sm":[{"m":"agent draft (claude)","l":"support human ai","tx":"DiD estimator for staggered treatment timing with heterogeneous effects. It shows that two-way fixed-effects regressions can be badly biased when already-treated units serve as controls, and it estimates group-time ATTs aggregated into event-study effects."},{"m":"agent draft (claude)","l":"operational translation","tx":"PM seat (council 5): Staggered DiD with heterogeneous effects. The tool for the upgrade card and a within-org 2025.2 vs 2024.2 comparison (9-13 orgs run both series in the same 2026 quarter)."}]},{"k":"cameron2008cluster","a":"ai-operations","t":"Bootstrap-Based Improvements for Inference with Clustered Errors","y":2008,"s":"peer_reviewed","e":4,"d":"10.1162/rest.90.3.414","v":"The Review of Economics and Statistics","a_":["Cameron et al."],"bl":"https://doi.org/10.1162/rest.90.3.414","pc":7,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Wild cluster bootstrap-t gives far better coverage than percentile/pairs cluster bootstrap with few clusters (5-30). Relevant: relrate cells rest on 11-17 orgs."}]},{"k":"cameron2015practitioner","a":"econometrics","t":"A Practitioner’s Guide to Cluster-Robust Inference","y":2015,"s":"peer_reviewed","e":2,"d":"10.3368/jhr.50.2.317","v":"Journal of Human Resources","a_":["Cameron & Miller"],"bl":"https://doi.org/10.3368/jhr.50.2.317","pc":4,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Cluster-robust inference guide: with few or unbalanced clusters, standard errors are too small; use cluster bootstrap / CR3 / wild cluster."}]},{"k":"campbell-1959-mtmm","a":"foundations-other","t":"Convergent and discriminant validation by the multitrait-multimethod matrix.","y":1959,"s":"peer_reviewed","e":2,"d":"10.1037/h0046016","v":"Psychological Bulletin","a_":["Campbell & Fiske"],"bl":"https://doi.org/10.1037/h0046016","pc":1,"sm":[{"m":"agent draft (claude)","l":"fairness measurement","tx":"The multitrait-multimethod matrix. Convergent validity needs agreement across different methods measuring the same trait; method variance shared within a source inflates within-source correlations. For ticket effort, Zendesk, Slack, Jira and PagerDuty are methods."}]},{"k":"campbell-1988","a":"foundations-other","t":"Task Complexity: A Review and Analysis","y":1988,"s":"peer_reviewed","e":5,"d":"10.5465/amr.1988.4306775","v":"The Academy of Management Review","a_":["Campbell"],"bl":"https://doi.org/10.5465/amr.1988.4306775","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Objective task-complexity attributes: multiple paths, multiple outcomes, conflicting interdependence, uncertainty about path-outcome links. Supports grading support tickets on objective characteristics rather than perceived difficulty."}]},{"k":"campbell1979planned","a":"foundations-se","t":"Assessing the impact of planned social change","y":1979,"s":"peer_reviewed","e":2,"d":"10.1016/0149-7189(79)90048-X","v":"Evaluation and Program Planning","a_":["Campbell"],"bl":"https://doi.org/10.1016/0149-7189(79)90048-X","pc":1,"sm":[{"m":"agent draft (claude)","l":"dashboard fix validation","tx":"Source of Campbell's law: the more a quantitative indicator is used for social decision-making, the more it is subject to corruption pressures and the more it distorts the processes it monitors. Underpins pairing every SLA attainment metric with a counter-metric and an unadjusted series, and against tying individual reward to ack speed."}]},{"k":"canay-2021","a":"foundations-stats","t":"The Wild Bootstrap with a “Small” Number of “Large” Clusters","y":2021,"s":"peer_reviewed","e":2,"d":"10.1162/rest_a_00887","v":"The Review of Economics and Statistics","a_":["Canay et al."],"bl":"https://doi.org/10.1162/rest_a_00887","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"The wild bootstrap is valid with a small number of large clusters under homogeneity conditions, via an approximate randomization-test argument."}]},{"k":"card-2016","a":"foundations-other","t":"The problem with ‘5 whys’","y":2016,"s":"peer_reviewed","e":2,"d":"10.1136/bmjqs-2016-005849","v":"BMJ Quality &amp; Safety","a_":["Card"],"bl":"https://doi.org/10.1136/bmjqs-2016-005849","pc":1,"sm":[{"m":"agent draft (claude)","l":"contrasting view","tx":"Card argues the 5 Whys forces a single linear causal chain, depends on the investigator's prior knowledge, is not reproducible across teams, and stops at a convenient 'root'."}]},{"k":"carmel2010followsun","a":"foundations-se","t":"\"Follow the Sun\" Workflow in Global Software Development","y":2010,"s":"peer_reviewed","e":2,"d":"10.2753/MIS0742-1222270102","v":"Journal of Management Information Systems","a_":["Carmel et al."],"bl":"https://doi.org/10.2753/MIS0742-1222270102","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Common-good lens: in follow-the-sun support the agent note is a handoff artefact present at minute 0 regardless of shift. Measure whether tickets opened in thin-staffing hours get a usable working diagnosis before the next shift's first touch, and whether the note reduces handoff re-work."},{"m":"agent draft (claude)","l":"dashboard fix validation","tx":"Defines follow-the-sun workflow (daily handoffs across time-zone sites to shorten calendar duration) and finds handoff coordination costs are the main obstacle; reports a quasi-experiment on handoff practices."}]},{"k":"carroll2006measurementerror","a":"foundations-stats","t":"Measurement Error in Nonlinear Models","y":2006,"s":"peer_reviewed","e":2,"d":"10.1201/9781420010138","v":"Chapman and Hall/CRC","a_":["Carroll et al."],"bl":"https://doi.org/10.1201/9781420010138","pc":1,"sm":[{"m":"agent draft (claude)","l":"uncertainty quantification","tx":"Reference text on classical vs Berkson error, regression calibration, SIMEX and Bayesian measurement-error models, including designs with a validation subsample. The log-derived hours are a surrogate for true effort;"}]},{"k":"casper2024blackbox","a":"security-governance","t":"Black-Box Access is Insufficient for Rigorous AI Audits","y":2024,"s":"peer_reviewed","e":4,"d":"10.1145/3630106.3659037","v":"The 2024 ACM Conference on Fairness, Accountability, and Transparency","a_":["Casper et al."],"bl":"https://doi.org/10.1145/3630106.3659037","pc":1,"sm":[{"m":"agent draft (claude)","l":"ai platform review","tx":"FAccT 2024: audits limited to black-box query access are weak; rigorous audits need white-box and 'outside-the-box' access (training/deployment documentation, evaluation methodology, internal records)."}]},{"k":"castro2008bugreporting","a":"security-governance","t":"Better bug reporting with better privacy","y":2008,"s":"peer_reviewed","e":4,"d":"10.1145/1346281.1346322","v":"Proceedings of the 13th international conference on Architectural support for programming languages and operating systems","a_":["Castro et al."],"bl":"https://doi.org/10.1145/1346281.1346322","pc":4,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Castro, Costa, and Martin (ASPLOS 2008) formalize privacy-preserving bug reporting by computing symbolic path conditions and replacing private user data (SQL literals, payloads, file contents) with synthetic or redacted values that preserve control-flow equivalence for debugging."},{"m":"agent draft (gemini)","l":"security and privacy","tx":"Proves the necessity of source-side privacy scrubbing before diagnostic reports leave customer hosts; operationalized in yb-collector's on-node AhoCorasick + Regex ybc-1.0.0 redactor and strict HMAC-SHA256 IPv4 pseudonymization."}]},{"k":"caveness2020tfdv","a":"foundations-systems","t":"TensorFlow Data Validation: Data Analysis and Validation in Continuous ML Pipelines","y":2020,"s":"peer_reviewed","e":2,"d":"10.1145/3318464.3384707","v":"Proceedings of the 2020 ACM SIGMOD International Conference on Management of Data","a_":["Caveness et al."],"bl":"https://doi.org/10.1145/3318464.3384707","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"TFDV: schema inferred from data, each batch validated for missing features, type/domain drift and span skew."}]},{"k":"chaparro2017missing","a":"ai-operations","t":"Detecting Missing Information in Bug Descriptions","y":2017,"s":"peer_reviewed","e":4,"d":"10.1145/3106237.3106285","v":"Proceedings of the 2017 11th Joint Meeting on Foundations of Software Engineering (ESEC/FSE '17)","a_":["Chaparro et al."],"bl":"https://doi.org/10.1145/3106237.3106285","pc":2,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Defines discourse patterns for observed behaviour (OB), expected behaviour (EB) and steps to reproduce (S2R) in bug descriptions and detects missing EB/S2R with pattern-based and ML detectors. Directly operationalises 'detect missing information at report time and ask for it'."},{"m":"agent draft (gemini)","l":"Automated detection of missing Observed Behavior (OB), Expected Behavior (EB), and Context","tx":"Discourse-pattern analysis of 2,912 bug descriptions demonstrating that reporters frequently omit structured behavioral and reproduction context, and showing that automated discourse rules can detect missing factual slots at submission time without guessing content."}]},{"k":"charlson-1987","a":"foundations-other","t":"A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation","y":1987,"s":"peer_reviewed","e":3,"d":"10.1016/0021-9681(87)90171-8","v":"Journal of Chronic Diseases","a_":["Charlson et al."],"bl":"https://doi.org/10.1016/0021-9681(87)90171-8","pc":1,"sm":[{"m":"agent draft (claude)","l":"support effort scoring","tx":"Derives a weighted comorbidity index (integer weights 1, 2, 3, 6 from relative risks in a Cox model) on one cohort and validates it on another for 10-year mortality. The method, regression-derived integer points validated out of sample, is the classical precursor of points scorecards such as RiskSLIM and of our ex-ante points card for expected ticket effort."}]},{"k":"chaudhuri2002compressing","a":"foundations-systems","t":"Compressing SQL workloads","y":2002,"s":"peer_reviewed","e":3,"d":"10.1145/564691.564747","v":"Proceedings of the 2002 ACM SIGMOD international conference on Management of data","a_":["Chaudhuri et al."],"bl":"https://doi.org/10.1145/564691.564747"},{"k":"chen-2016","a":"ai-operations","t":"AutoFDO: automatic feedback-directed optimization for warehouse-scale applications","y":2016,"s":"peer_reviewed","e":4,"d":"10.1145/2854038.2854044","v":"Proceedings of the 2016 International Symposium on Code Generation and Optimization","a_":["Chen et al."],"bl":"https://doi.org/10.1145/2854038.2854044","pc":1,"sm":[{"m":"agent draft (claude)","l":"bibliographic context","tx":"AutoFDO: sampling-based feedback-directed optimization from production profiles, with iterative and source stability concerns. Go PGO (default.pgo, -pgo=auto since Go 1.21) explicitly follows this AutoFDO-style loop (build, collect production CPU pprof, rebuild); Go docs cite 2-14% gains on Go 1.22 benchmarks."}]},{"k":"chen-2019-incident-triage","a":"foundations-telemetry","t":"An Empirical Investigation of Incident Triage for Online Service Systems","y":2019,"s":"peer_reviewed","e":3,"d":"10.1109/icse-seip.2019.00020","v":"2019 IEEE/ACM 41st International Conference on Software Engineering: Software Engineering in Practice (ICSE-SEIP)","a_":["Chen et al."],"bl":"https://doi.org/10.1109/icse-seip.2019.00020","pc":1,"sm":[{"m":"agent draft (claude-sonnet-5-5)","l":"complexity measurement without effort","tx":"Microsoft empirical study of incident triage in online service systems: incidents are frequently reassigned between teams, and this triage cost is substantial."}]},{"k":"chen-2021","a":"foundations-telemetry","t":"Graph-based Incident Aggregation for Large-Scale Online Service Systems","y":2021,"s":"peer_reviewed","e":7,"d":"10.1109/ASE51524.2021.9678746","v":"2021 36th IEEE/ACM International Conference on Automated Software Engineering (ASE)","a_":["Chen et al."],"bl":"https://doi.org/10.1109/ASE51524.2021.9678746","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"GRLIA: incident aggregation by learning a graph of cascading-failure relations among alert types, evaluated on Huawei Cloud with linked-incident labels. A template for multi-plane fingerprinting over bundles, but its labels came from rich on-call link records;"}]},{"k":"chen-2022","a":"foundations-viz","t":"VizLinter: A Linter and Fixer Framework for Data Visualization","y":2022,"s":"peer_reviewed","e":3,"d":"10.1109/TVCG.2021.3114804","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Chen et al."],"bl":"https://doi.org/10.1109/TVCG.2021.3114804","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"VizLinter encodes visualization design rules (expressiveness, effectiveness, misleading-scale rules) as lint checks over Vega-Lite specs and proposes minimal fixes."}]},{"k":"chen-2025","a":"foundations-viz","t":"VisEval: A Benchmark for Data Visualization in the Era of Large Language Models","y":2025,"s":"peer_reviewed","e":5,"d":"10.1109/TVCG.2024.3456320","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Chen et al."],"bl":"https://doi.org/10.1109/TVCG.2024.3456320","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"VisEval: NL2VIS benchmark of 2,524 queries over 146 databases with automated checkers for validity (code runs, chart renders), legality (data/chart type correct for the query) and readability (layout, overlap, scales)."}]},{"k":"chen1998metamorphic","a":"foundations-se","t":"Metamorphic Testing: A New Approach for Generating Next Test Cases","y":1998,"s":"preprint","e":1,"x":"2002.12543","v":"arXiv:2002.12543","a_":["Chen et al."],"bl":"https://arxiv.org/abs/2002.12543","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Original metamorphic testing proposal: derive follow-up tests from successful ones through expected input/output relations."}]},{"k":"chen2014ormantipatterns","a":"foundations-se","t":"Detecting performance anti-patterns for applications developed using object-relational mapping","y":2014,"s":"peer_reviewed","e":5,"d":"10.1145/2568225.2568259","v":"Proceedings of the 36th International Conference on Software Engineering","a_":["Chen et al."],"bl":"https://doi.org/10.1145/2568225.2568259","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Static-analysis framework that detects ORM performance anti-patterns (one-by-one processing i.e. N+1, excessive data retrieval) in Java/Hibernate code and ranks them by measured performance impact, evaluated on open-source and an industrial system."}]},{"k":"chen2016redundantdataaccess","a":"foundations-se","t":"Finding and Evaluating the Performance Impact of Redundant Data Access for Applications that are Developed Using Object-Relational Mapping Frameworks","y":2016,"s":"peer_reviewed","e":5,"d":"10.1109/TSE.2016.2553039","v":"IEEE Transactions on Software Engineering","a_":["Chen et al."],"bl":"https://doi.org/10.1109/TSE.2016.2553039"},{"k":"chen2019incidenttriage","a":"foundations-telemetry","t":"An Empirical Investigation of Incident Triage for Online Service Systems","y":2019,"s":"peer_reviewed","e":10,"d":"10.1109/ICSE-SEIP.2019.00020","v":"2019 IEEE/ACM 41st International Conference on Software Engineering: Software Engineering in Practice (ICSE-SEIP)","a_":["Chen et al."],"bl":"https://doi.org/10.1109/ICSE-SEIP.2019.00020","pc":3,"sm":[{"m":"agent draft (claude)","l":"coordination","tx":"Incident data from 20 large Microsoft online services (20K+ teams): 4.11% to 91.58% of incidents per service were reassigned at least once, and reassignment increased triage time by up to 10.16x on average, with heavier impact on high-severity incidents. Gives a range for hand-off prevalence and shows that each hand-off adds substantial elapsed time;"},{"m":"agent draft (claude)","l":"empirical findings","tx":"Chen et al. (ICSE-SEIP'19, Microsoft) find incident reassignment/triage is frequent and costly; many incidents are misrouted before reaching the responsible team. For Yugabyte: the gallery does not measure Support->CE->Engineering routing hops or reassignments; time from CE link to correct owner (DB vs PLAT vs CSP) is a feasible metric from Jira history."},{"m":"agent draft (claude)","l":"operational translation","tx":"Empirical study of incident triage at Microsoft: reassignment between teams is common and costly, and triage time is a large share of time to mitigate. For support on-call, the measurable outcome of bundle evidence is fewer reassignments and shorter time to the first correct owner and first human diagnosis, not bot-note latency."}]},{"k":"chen2020incidental","a":"foundations-telemetry","t":"How incidental are the incidents?","y":2020,"s":"peer_reviewed","e":6,"d":"10.1145/3324884.3416624","v":"Proceedings of the 35th IEEE/ACM International Conference on Automated Software Engineering","a_":["Chen et al."],"bl":"https://doi.org/10.1145/3324884.3416624","pc":2,"sm":[{"m":"agent draft (claude)","l":"empirical findings","tx":"Microsoft, 18 online services: a large share of incidents are 'incidental', meaning engineers deprioritise them after diagnosis. DeepIP ranks incidents by the predicted probability of being incidental (AUC 0.808 vs 0.624). For CE escalation analytics: not every CE link is a real defect escalation."},{"m":"agent draft (claude)","l":"operational translation","tx":"Many incidents at Microsoft recur or relate to earlier ones, so retrieving similar historical incidents and their fixes speeds resolution. Supports a known-issue retrieval measure at CE handoff: was the resolving issue findable from history before the CE opened?"}]},{"k":"chen2020lidar","a":"foundations-telemetry","t":"Identifying linked incidents in large-scale online service systems","y":2020,"s":"peer_reviewed","e":14,"d":"10.1145/3368089.3409768","v":"Proceedings of the 28th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering","a_":["Chen et al."],"bl":"https://doi.org/10.1145/3368089.3409768","pc":3,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Chen et al. (LiDAR, FSE'20) identify linked incidents across services using text plus component-dependency features; motivated by duplicate/related incidents wasting engineer effort. For Yugabyte CE: many disk-stall CEs are linked (same customer, same AZ, same cloud volume);"},{"m":"agent draft (claude)","l":"operational translation","tx":"Microsoft LiDAR identifies linked incidents by combining incident text with component dependency information. For Support: link tickets that share a bundle signature, version and topology pattern, so managers see clusters of tickets with a common cause (candidate CE/bug) instead of isolated cases."},{"m":"agent draft (claude)","l":"operational translation","tx":"LiDAR links incidents across services using text plus component dependency graphs. For YB, known-issue auto-matching (bundle signature to Jira) is the linking problem; KFin's 11 write-id FATAL cases were never linked to DB-20218/21822, which a matcher on status class + check site would have done."}]},{"k":"chen2020strangler","a":"foundations-systems","t":"Microservice Migration Using Strangler Fig Pattern: A Case Study on the Green Button System","y":2020,"s":"peer_reviewed","e":1,"d":"10.1109/ICS51289.2020.00107","v":"2020 International Computer Symposium (ICS)","a_":["Li et al."],"bl":"https://doi.org/10.1109/ICS51289.2020.00107","pc":1,"sm":[{"m":"agent draft (gemini)","l":"strangler fig migration","tx":"Demonstrates Fowler's Strangler Fig pattern combined with Domain-Driven Design bounded contexts and an Anti-Corruption Layer (ACL): legacy callers remain untouched while individual bounded modules are extracted behind a routing facade and validated quantitatively against the unmigrated baseline."}]},{"k":"chen2024frugalgpt","a":"ai-operations","t":"FrugalGPT: How to Use Large Language Models While Reducing Cost and Improving Performance","y":2024,"s":"external","e":2,"v":"Transactions on Machine Learning Research (TMLR)","a_":["Chen et al."],"pc":2,"sm":[{"m":"agent draft (qwen)","l":"operational translation","tx":"FrugalGPT's cascade beats single-model GPT-4 on their benchmark at up to 98% cost reduction on some tasks, with scoring-calibrated escalation — the canonical cost argument for a cheap first pass."}]},{"k":"chen2024rcacopilot","a":"ai-operations","t":"Automatic Root Cause Analysis via Large Language Models for Cloud Incidents","y":2024,"s":"peer_reviewed","e":11,"d":"10.1145/3627703.3650072","v":"Proceedings of the Nineteenth European Conference on Computer Systems (EuroSys '24)","a_":["Chen et al."],"bl":"https://doi.org/10.1145/3627703.3650072","pc":12,"sm":[{"m":"agent draft (gemini)","l":"systems performance","tx":"Chen et al. (EuroSys 2024) present RCACopilot for automated cloud incident root-cause analysis at Microsoft, matching incidents to diagnostic handlers and predicting root-cause categories with explanatory narratives."},{"m":"agent draft (claude)","l":"operational translation","tx":"RCACopilot (Microsoft, EuroSys 2024): handler-driven diagnostic collection per alert type, then an LLM predicts the root-cause category and explains it, using retrieved similar historical incidents in context."},{"m":"agent draft (qwen)","l":"operational translation","tx":"RCACopilot never abstains by design (classification), so it inherits every hallucination risk the OpenRCA/ITBench line documents."}]},{"k":"chhugani2008sorting","a":"foundations-systems","t":"Efficient implementation of sorting on multi-core SIMD CPU architecture","y":2008,"s":"peer_reviewed","e":3,"d":"10.14778/1454159.1454171","v":"Proceedings of the VLDB Endowment","a_":["Chhugani et al."],"bl":"https://doi.org/10.14778/1454159.1454171","pc":1,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Chhugani et al. (VLDB '08) show that sorting on multi-core SIMD CPUs becomes memory-bandwidth-bound once data exceeds L2/LLC cache unless blocks are sized to $C / (2E)$ and merged via register-resident SIMD bitonic merge networks and multi-way merge trees."}]},{"k":"chidambaram-2013","a":"foundations-telemetry","t":"Optimistic crash consistency","y":2013,"s":"peer_reviewed","e":3,"d":"10.1145/2517349.2522726","v":"Proceedings of the Twenty-Fourth ACM Symposium on Operating Systems Principles","a_":["Chidambaram et al."],"bl":"https://doi.org/10.1145/2517349.2522726","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Optimistic crash consistency: separates ordering from durability (osync vs dsync) to avoid cache flushes; shows flushes are the dominant cost of journaling file systems. Background for why fsync cost is device-cache-flush cost."}]},{"k":"choetkiertikul-2019","a":"foundations-stats","t":"A Deep Learning Model for Estimating Story Points","y":2019,"s":"peer_reviewed","e":6,"d":"10.1109/TSE.2018.2792473","v":"IEEE Transactions on Software Engineering","a_":["Choetkiertikul et al."],"bl":"https://doi.org/10.1109/TSE.2018.2792473"},{"k":"chu-2021","a":"foundations-telemetry","t":"Prefix-Graph: A Versatile Log Parsing Approach Merging Prefix Tree with Probabilistic Graph","y":2021,"s":"peer_reviewed","e":7,"d":"10.1109/icde51399.2021.00274","v":"2021 IEEE 37th International Conference on Data Engineering (ICDE)","a_":["Chu et al."],"bl":"https://doi.org/10.1109/icde51399.2021.00274","pc":1,"sm":[{"m":"agent draft (gemini)","l":"log template mining","tx":"Chu et al. (ICDE 2021, Prefix-Graph) address the rigid token-length and prefix assumptions of prefix-tree parsers (like Drain) by merging a prefix trie with a probabilistic token transition graph and iteratively finding densest subgraphs (cliques/branches) to separate templates that share prefixes or diverge mid-message."}]},{"k":"claessen2000quickcheck","a":"foundations-se","t":"QuickCheck","y":2000,"s":"peer_reviewed","e":3,"d":"10.1145/351240.351266","v":"Proceedings of the fifth ACM SIGPLAN international conference on Functional programming","a_":["Claessen & Hughes"],"bl":"https://doi.org/10.1145/351240.351266","pc":2,"sm":[{"m":"agent draft (gemini)","l":"property based differential testing","tx":"Introduces executable specifications and composable random generators for property-based testing."}]},{"k":"clayton1987ebayes","a":"foundations-stats","t":"Empirical Bayes Estimates of Age-Standardized Relative Risks for Use in Disease Mapping","y":1987,"s":"peer_reviewed","e":2,"d":"10.2307/2532003","v":"Biometrics","a_":["Clayton & Kaldor"],"bl":"https://doi.org/10.2307/2532003","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Council 5 check: fs_bands now implements Poisson-gamma EB by org (shared frailty across bands, alpha by moments, EM-style updates) with org-bootstrap intervals. Remaining gaps: a case-level binary per node-day saturates in large clusters, and 'no trend' rests on overlapping intervals."}]},{"k":"cleveland1984graphical","a":"foundations-viz","t":"Graphical Perception: Theory, Experimentation, and Application to the Development of Graphical Methods","y":1984,"s":"peer_reviewed","e":12,"d":"10.1080/01621459.1984.10478080","v":"Journal of the American Statistical Association","a_":["Cleveland & McGill"],"bl":"https://doi.org/10.1080/01621459.1984.10478080","pc":3,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Cleveland-McGill ranking (position on common scale > length > angle/area > colour) justifies the dot/bar forms used throughout the Bundle Insights gallery; the gallery mostly honours it, but adds colour flags on point estimates (fs_infra 'more than twice the fleet rate') that ignore the drawn CI."},{"m":"agent draft (claude)","l":"dashboard design review","tx":"Ranks elementary perceptual tasks: position on a common scale beats length on unaligned baselines, which beats angle, area and hue. Applied to the support suite: stacked bars (issue category mix, 100%-stacked resolution-time shares, per-assignee stacks) leave only the bottom segment aligned."},{"m":"agent draft (gemini)","l":"methodology","tx":"Establishes the foundational hierarchy of graphical perception accuracy: position along a common aligned scale (horizontal bar charts and aligned dumbbell/dot plots) yields the lowest perceptual judgment error, outperforming length on non-aligned scales (stacked bars), angle (pie charts), area (bubbles), and color saturation."}]},{"k":"cohen1960kappa","a":"ai-operations","t":"A Coefficient of Agreement for Nominal Scales","y":1960,"s":"peer_reviewed","e":5,"d":"10.1177/001316446002000104","v":"Educational and Psychological Measurement","a_":["Cohen"],"bl":"https://doi.org/10.1177/001316446002000104","pc":2},{"k":"cole-2004","a":"foundations-stats","t":"Adjusted survival curves with inverse probability weights","y":2004,"s":"peer_reviewed","e":1,"d":"10.1016/j.cmpb.2003.10.004","v":"Computer Methods and Programs in Biomedicine","a_":["Cole & Hernán"],"bl":"https://doi.org/10.1016/j.cmpb.2003.10.004","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Adjusted survival curves by inverse-probability-of-exposure weighting. The natural next step for the outcomes card (evidence-ready vs not), with org-cluster bootstrap; only as good as the measured confounders, which the bias card's failed CE check says are insufficient."}]},{"k":"collins-2024","a":"foundations-stats","t":"TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods","y":2024,"s":"peer_reviewed","e":1,"d":"10.1136/bmj-2023-078378","v":"BMJ","a_":["Collins et al."],"bl":"https://doi.org/10.1136/bmj-2023-078378"},{"k":"cook1998complex","a":"foundations-telemetry","t":"How Complex Systems Fail","y":1998,"s":"external","e":1,"a_":["Cook"],"bl":"https://worrydream.com/booklets/Cook_1998_-_How_Complex_Systems_Fail_(booklet).pdf","pc":1,"sm":[{"m":"agent draft (claude)","l":"contrasting view","tx":"Cook's 18 points: complex systems run degraded, catastrophe needs multiple failures, post-accident attribution to a 'root cause' is fundamentally wrong, hindsight biases assessment."}]},{"k":"corbett2012spanner","a":"distributed-systems","t":"Spanner: Google's Globally-Distributed Database","y":2012,"s":"external","e":7,"v":"Proceedings of the 10th USENIX Symposium on Operating Systems Design and Implementation (OSDI '12)","a_":["Corbett et al."],"pc":10,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Spanner's leader reads, TrueTime-bounded uncertainty and bounded-staleness snapshot reads map to YugabyteDB tablet-leader reads, max_clock_skew-driven read restarts (40001), and yb_read_from_followers + yb_follower_read_staleness_ms. Profiler implication: cross-region latency is a leader-placement problem first;"}]},{"k":"correll-2014","a":"foundations-viz","t":"Error Bars Considered Harmful: Exploring Alternate Encodings for Mean and Error","y":2014,"s":"peer_reviewed","e":6,"d":"10.1109/TVCG.2014.2346298","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Correll & Gleicher"],"bl":"https://doi.org/10.1109/TVCG.2014.2346298"},{"k":"correll-2020","a":"foundations-viz","t":"Truncating the Y-Axis: Threat or Menace?","y":2020,"s":"peer_reviewed","e":6,"d":"10.1145/3313831.3376222","v":"Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems","a_":["Correll et al."],"bl":"https://doi.org/10.1145/3313831.3376222"},{"k":"costa-2021","a":"ai-operations","t":"What's Wrong with My Benchmark Results? Studying Bad Practices in JMH Benchmarks","y":2021,"s":"peer_reviewed","e":4,"d":"10.1109/TSE.2019.2925345","v":"IEEE Transactions on Software Engineering","a_":["Costa et al."],"bl":"https://doi.org/10.1109/TSE.2019.2925345","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Mines open-source JMH suites and finds common bad practices that bias results: dead-code elimination (results not returned or sunk into a Blackhole), constant folding of inputs (non-@State finals), loops inside @Benchmark methods, and setup at the wrong @Setup Level (Invocation). Measures the resulting distortion."}]},{"k":"cox-1972","a":"foundations-stats","t":"Regression Models and Life-Tables","y":1972,"s":"peer_reviewed","e":6,"d":"10.1111/j.2517-6161.1972.tb00899.x","v":"Journal of the Royal Statistical Society Series B: Statistical Methodology","a_":["Cox"],"bl":"https://doi.org/10.1111/j.2517-6161.1972.tb00899.x","pc":1,"sm":[{"m":"agent draft (gemini)","l":"methodology","tx":"Introduces the proportional hazards regression model and partial likelihood estimation for right-censored failure-time data without parametric baseline hazard assumptions."}]},{"k":"cox2015libyear","a":"foundations-se","t":"Measuring Dependency Freshness in Software Systems","y":2015,"s":"peer_reviewed","e":2,"d":"10.1109/ICSE.2015.140","v":"2015 IEEE/ACM 37th IEEE International Conference on Software Engineering","a_":["Cox et al."],"bl":"https://doi.org/10.1109/ICSE.2015.140","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Defines dependency freshness as the time (libyears) or version distance between the used and the newest available release, and shows version-count distance is distorted by release cadence."}]},{"k":"cragg-1971","a":"econometrics","t":"Some Statistical Models for Limited Dependent Variables with Application to the Demand for Durable Goods","y":1971,"s":"peer_reviewed","e":1,"d":"10.2307/1909582","v":"Econometrica","a_":["Cragg"],"bl":"https://doi.org/10.2307/1909582"},{"k":"crdb2026diskstall","a":"foundations-telemetry","t":"Troubleshoot cluster setup: disk stalls","y":2026,"s":"external","e":4,"a_":["Cockroach Labs"],"bl":"https://docs.cockroachlabs.com/docs/stable/cluster-setup-troubleshooting","pc":2,"sm":[{"m":"agent draft (claude)","l":"contrasting view","tx":"CockroachDB states transient multi-second disk stalls are common in cloud, warns at 10 s and terminates the node at 20 s of WAL sync stall, and recommends WAL failover to a second disk. A vendor precedent for a bounded-stall policy (T1) rather than leaving leadership on the stalled node."},{"m":"agent draft (claude)","l":"operational translation","tx":"CockroachDB documents that multi-second transient disk stalls are common in cloud and bounds them: warn at 10 s, terminate the node at 20 s, WAL failover. Vendor precedent for a bounded-stall policy and a default alert, which YBA lacks."}]},{"k":"crdb2026loadsplit","a":"distributed-systems","t":"Load-Based Splitting (CockroachDB documentation)","y":2026,"s":"external","e":1,"a_":["Cockroach Labs"],"bl":"https://docs.cockroachlabs.com/docs/stable/load-based-splitting"},{"k":"crdb2026walfailover","a":"distributed-systems","t":"WAL Failover (CockroachDB documentation)","y":2026,"s":"external","e":2,"a_":["Cockroach Labs"],"bl":"https://www.cockroachlabs.com/docs/stable/wal-failover","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"CockroachDB treats disk stalls by magnitude: a stalled store blocks writes for its leases until the stall clears or the node is crashed, by default after 20 s (max sync duration); WAL failover lets it ride out ~40 s. A single-stall-length threshold, not a volume budget, is the vendor's failover trigger."}]},{"k":"crdb_diagreporting","a":"ai-operations","t":"Diagnostics Reporting","y":null,"s":"external","e":1,"a_":["Cockroach Labs"],"bl":"https://docs.cockroachlabs.com/docs/stable/diagnostics-reporting","pc":3,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"CockroachDB diagnostics reporting sends anonymised cluster and feature-usage counters by default (opt-out) and documents the payload. The installed-base feature frame that bundle-based feature cards lack."}]},{"k":"cronin-1992","a":"foundations-other","t":"Measuring Service Quality: A Reexamination and Extension","y":1992,"s":"peer_reviewed","e":1,"d":"10.1177/002224299205600304","v":"Journal of Marketing","a_":["Cronin & Taylor"],"bl":"https://doi.org/10.1177/002224299205600304","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"SERVPERF: performance-only measures predict overall service quality better than SERVQUAL gap scores. For support dashboards: measure delivered performance (resolution, reopen, effort proxies) rather than expectation-gap surveys."}]},{"k":"croskerry2009universal","a":"foundations-other","t":"A Universal Model of Diagnostic Reasoning","y":2009,"s":"peer_reviewed","e":1,"d":"10.1097/ACM.0b013e3181ace703","v":"Academic Medicine","a_":["Croskerry"],"bl":"https://doi.org/10.1097/ACM.0b013e3181ace703","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Dual-process model of diagnostic reasoning: fast pattern recognition (System 1) for familiar presentations, analytic reasoning (System 2) for unfamiliar ones, with calibration and override between them. Maps to known-issue matching (System 1) versus hypothesis-driven investigation plans (System 2) in DB support triage."}]},{"k":"cui2025genaidevs","a":"foundations-stats","t":"The Effects of Generative AI on High-Skilled Work: Evidence from Three Field Experiments with Software Developers","y":2026,"s":"peer_reviewed","e":2,"d":"10.1287/mnsc.2025.00535","v":"Management Science","a_":["Cui et al."],"bl":"https://doi.org/10.1287/mnsc.2025.00535","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Three RCTs (4,867 developers): +26% completed tasks pooled, noisy per site, gains larger for juniors. Operational translation: even thousands of randomized users gave noisy per-site estimates; ~80 bundled non-P1 tickets a month cannot power a CE-rate holdout of the AI note, so measure engineer-level process outcomes instead."}]},{"k":"curino2008prism","a":"foundations-systems","t":"Graceful database schema evolution","y":2008,"s":"peer_reviewed","e":3,"d":"10.14778/1453856.1453939","v":"Proceedings of the VLDB Endowment","a_":["Curino et al."],"bl":"https://doi.org/10.14778/1453856.1453939"},{"k":"curino2008wikipedia","a":"foundations-systems","t":"SCHEMA EVOLUTION IN WIKIPEDIA - Toward a Web Information System Benchmark","y":2008,"s":"peer_reviewed","e":2,"d":"10.5220/0001713003230332","v":"Proceedings of the Tenth International Conference on Enterprise Information Systems","bl":"https://doi.org/10.5220/0001713003230332"},{"k":"curtsinger-2013","a":"ai-operations","t":"STABILIZER","y":2013,"s":"peer_reviewed","e":6,"d":"10.1145/2451116.2451141","v":"Proceedings of the eighteenth international conference on Architectural support for programming languages and operating systems","a_":["Curtsinger & Berger"],"bl":"https://doi.org/10.1145/2451116.2451141","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Stabilizer re-randomizes code/stack/heap layout during execution so layout bias becomes averageable noise, enabling sound tests. For profiler agents: small deltas (a few percent) are within layout noise unless randomized or heavily repeated."}]},{"k":"curtsinger-2015","a":"ai-operations","t":"Coz: Finding Code that Counts with Causal Profiling","y":2015,"s":"peer_reviewed","e":5,"d":"10.1145/2815400.2815409","v":"Proceedings of the 25th Symposium on Operating Systems Principles","a_":["Curtsinger & Berger"],"bl":"https://doi.org/10.1145/2815400.2815409","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Coz virtual speedups answer which line, if faster, speeds the program end to end, for concurrent code where the hottest frame is not the bottleneck. In rust-profiler: use coz-rs progress points only when already in the code; Linux-only; complements samply/perf hotspot ranking."},{"m":"agent draft (gemini)","l":"causal profiling","tx":"Demonstrates via Coz causal profiling (virtual speedups) that flat CPU share != optimization potential in multi-threaded/distributed systems. Optimizing a high-CPU stage that feeds a downstream serial bottleneck (such as mutex contention, RPC queueing, or WAL fsync) merely increases downstream queueing delay, explains sub-additive compounding, and…"}]},{"k":"cutler2024cedar","a":"foundations-systems","t":"Cedar: A New Language for Expressive, Fast, Safe, and Analyzable Authorization","y":2024,"s":"preprint","e":2,"d":"10.1145/3649835","x":"2403.04651","v":"Proceedings of the ACM on Programming Languages","a_":["Cutler et al."],"bl":"https://doi.org/10.1145/3649835","pc":1,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Describes AWS Cedar (OOPSLA 2024), the flagship industrial deployment of Verification-Guided Development pairing an executable Lean 4 formal model and machine-checked soundness proofs with differential testing against a production Rust implementation."}]},{"k":"dabrowski2022appreviews","a":"foundations-se","t":"Analysing app reviews for software engineering: a systematic literature review","y":2022,"s":"peer_reviewed","e":3,"d":"10.1007/s10664-021-10065-7","v":"Empirical Software Engineering","a_":["Dąbrowski et al."],"bl":"https://doi.org/10.1007/s10664-021-10065-7","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"PM seat (council 5): Systematic review of app-review analysis for SE (elicitation, bug/feature classification, prioritisation); flags weak evaluation and representativeness. The same caution applies to support tickets as a roadmap signal."}]},{"k":"dai2020logram","a":"foundations-telemetry","t":"Logram: Efficient Log Parsing Using n-Gram Dictionaries","y":2020,"s":"peer_reviewed","e":9,"d":"10.1109/TSE.2020.3007554","v":"IEEE Transactions on Software Engineering","a_":["Dai et al."],"bl":"https://doi.org/10.1109/TSE.2020.3007554","pc":1,"sm":[{"m":"agent draft (gemini)","l":"log template mining","tx":"Dai et al. (IEEE TSE 2020/2022, Logram) observe that static template words form high-frequency 2-grams and 3-grams across a log corpus whereas dynamic parameters form low-frequency n-grams, enabling fast dictionary-based template extraction without tree traversal."}]},{"k":"dang2012rebucket","a":"foundations-telemetry","t":"ReBucket: A method for clustering duplicate crash reports based on call stack similarity","y":2012,"s":"peer_reviewed","e":6,"d":"10.1109/icse.2012.6227111","v":"2012 34th International Conference on Software Engineering (ICSE)","a_":["Dang et al."],"bl":"https://doi.org/10.1109/icse.2012.6227111","pc":3,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"ReBucket clusters crash reports by stack similarity. Council-5 DocDB lens: WTL's termination plane now records check_site (file:line) and raw FATAL text per pid, so the tests crash panel can key buckets on check_site plus status text;"},{"m":"agent draft (gemini)","l":"operational translation","tx":"Dang et al. (ICSE 2012) introduce ReBucket, a position-dependent dynamic-programming stack alignment algorithm enacted across Microsoft Windows Error Reporting (WER) that models both (1) the distance of a frame from the top of the stack (exponential decay weight $e^{-c \\cdot i}$) and (2) the alignment offset penalty between two call stacks ($e^{-o…"}]},{"k":"david-2018-arxiv","a":"foundations-stats","t":"Categorizing Variants of Goodhart's Law","y":2018,"s":"preprint","e":2,"x":"1803.04585","v":"arXiv:1803.04585","a_":["Manheim & Garrabrant"],"bl":"https://arxiv.org/abs/1803.04585","pc":4,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Manheim and Garrabrant split Goodhart's law into four mechanisms: regressional (optimizing a noisy proxy selects for noise), extremal (the proxy-goal relation breaks at the tails), causal (intervening on a non-causal correlate), and adversarial (agents game the measure)."},{"m":"agent draft (claude)","l":"metric statistics review","tx":"Splits Goodhart failures into regressional, extremal, causal and adversarial variants. In a support SLA dashboard: ack-as-first-response invites adversarial optimisation (fast ack, slow substance; relay templating), the paged-only P1 scope invites selection (only ~49% of urgent tickets page), and a ready-to-close flag is optimisable by closing early."},{"m":"agent draft (claude)","l":"support human ai","tx":"Manheim & Garrabrant categorize Goodhart failures as regressional, extremal, causal and adversarial. Support implication for AI-platform metrics: acceptance rate (causal: pressure to accept), tool-call volume (adversarial: inflated calls), close-suggestion uptake (extremal: closes pushed past where the precision held)."}]},{"k":"davis-2024","a":"foundations-viz","t":"The Risks of Ranking: Revisiting Graphical Perception to Model Individual Differences in Visualization Performance","y":2024,"s":"peer_reviewed","e":5,"d":"10.1109/TVCG.2022.3226463","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Davis et al."],"bl":"https://doi.org/10.1109/TVCG.2022.3226463","pc":1,"sm":[{"m":"agent draft (claude)","l":"contrasting view","tx":"Replicates Cleveland & McGill's position/length/angle/area comparison with Bayesian multilevel models and shows the canonical encoding ranking is an average-observer artefact: many individuals deviate from it, so 'position beats area' is a population prior, not a law."}]},{"k":"davis2006prroc","a":"foundations-stats","t":"The relationship between Precision-Recall and ROC curves","y":2006,"s":"peer_reviewed","e":2,"d":"10.1145/1143844.1143874","v":"Proceedings of the 23rd international conference on Machine learning - ICML '06","a_":["Davis & Goadrich"],"bl":"https://doi.org/10.1145/1143844.1143874","pc":1,"sm":[{"m":"agent draft (gemini)","l":"methodology","tx":"Establishes the formal one-to-one theorem between ROC curves and Precision-Recall (PR) curves (a curve dominates in ROC space iff it dominates in PR space) and proves that linear interpolation in PR space is erroneous; directly underpins hagen-ai's step-wise interpolation-free AUPRC-error metric (forensic_metrics.py) under class imbalance."}]},{"k":"davis2018redos","a":"ai-operations","t":"The impact of regular expression denial of service (ReDoS) in practice: an empirical study at the ecosystem scale","y":2018,"s":"peer_reviewed","e":2,"d":"10.1145/3236024.3236027","v":"Proceedings of the 2018 26th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering","a_":["Davis et al."],"bl":"https://doi.org/10.1145/3236024.3236027","pc":3,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"ESEC/FSE 2018 ecosystem-scale empirical study of Regular Expression Denial of Service (ReDoS) across npm and PyPI, demonstrating that backtracking regex engines suffer catastrophic super-linear CPU stalls in production and motivating linear-time automata engines (RE2, Go regexp, Rust regex) in production telemetry agents."},{"m":"agent draft (qwen)","l":"operational translation","tx":"Davis et al.'s ecosystem-scale ReDoS study motivated our format-engine regex review: wtl compiles all attacker-facing and user-supplied format patterns exclusively with the linear-time Rust regex crate (PikeVM/NFA), rejects lookaround and backreferences at load time, and adds a 38-literal Aho-Corasick prefilter, so classic catastrophic-backtracking…"}]},{"k":"davis2019rethinking","a":"foundations-se","t":"Rethinking Regex engines to address ReDoS","y":2019,"s":"peer_reviewed","e":3,"d":"10.1145/3338906.3342509","v":"Proceedings of the 2019 27th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering","a_":["Davis"],"bl":"https://doi.org/10.1145/3338906.3342509","pc":1,"sm":[{"m":"agent draft (gemini)","l":"security governance","tx":"Demonstrates empirically that ~95% of regular expressions used in production software in Python, Java, and JavaScript do not require non-regular features and can be evaluated in guaranteed O(N) linear time using Thompson NFA/DFA simulation."}]},{"k":"dawid1979maximum","a":"ai-operations","t":"Maximum Likelihood Estimation of Observer Error-Rates Using the EM Algorithm","y":1979,"s":"peer_reviewed","e":1,"d":"10.2307/2346806","v":"Journal of the Royal Statistical Society: Series C (Applied Statistics)","a_":["Dawid & Skene"],"bl":"https://doi.org/10.2307/2346806","pc":2},{"k":"dawson2007referees","a":"sports-analytics","t":"Are Football Referees Really Biased and Inconsistent? Evidence on the Incidence of Disciplinary Sanction in the English Premier League","y":2007,"s":"peer_reviewed","e":1,"d":"10.1111/j.1467-985X.2006.00451.x","v":"Journal of the Royal Statistical Society: Series A","a_":["Dawson et al."],"bl":"https://doi.org/10.1111/j.1467-985X.2006.00451.x","pc":4,"sm":[{"m":"agent draft (gemini)","l":"methodology","tx":"Models home and away disciplinary sanctions jointly via bivariate Poisson regression, testing whether referee card-issuance variance reflects team playing style or officiating inconsistency."}]},{"k":"dayan-2017","a":"foundations-other","t":"Monkey","y":2017,"s":"peer_reviewed","e":9,"d":"10.1145/3035918.3064054","v":"Proceedings of the 2017 ACM International Conference on Management of Data","a_":["Dayan et al."],"bl":"https://doi.org/10.1145/3035918.3064054","pc":1,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"Monkey proves that assigning uniform bits-per-key to Bloom filters across all LSM levels is asymptotically suboptimal because the largest level dominates filter memory while smaller levels dominate false-positive SST probes."}]},{"k":"dayan-2018","a":"foundations-other","t":"Dostoevsky","y":2018,"s":"peer_reviewed","e":10,"d":"10.1145/3183713.3196927","v":"Proceedings of the 2018 International Conference on Management of Data","a_":["Dayan & Idreos"],"bl":"https://doi.org/10.1145/3183713.3196927","pc":1,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"Dostoevsky shows that standard leveled LSM-trees perform superfluous merging at the largest level (where almost no lookups terminate early and obsolete entries are already bounded by the capacity ratio T), introducing Fluid LSM-Tree (Lazy Leveling) that uses tiering at L1..L_{L-1} and leveling at L_L."}]},{"k":"dayan-2019","a":"foundations-telemetry","t":"The Log-Structured Merge-Bush &amp; the Wacky Continuum","y":2019,"s":"peer_reviewed","e":4,"d":"10.1145/3299869.3319903","v":"Proceedings of the 2019 International Conference on Management of Data","a_":["Dayan & Idreos"],"bl":"https://doi.org/10.1145/3299869.3319903","pc":1,"sm":[{"m":"agent draft (gemini)","l":"methodology","tx":"Introduces LSM-Bush and the Wacky continuum, setting doubly-exponential capacity ratios across smaller LSM levels so write amplification scales as O(log log N) rather than O(log N) while keeping point-read and range-read amplification navigable."}]},{"k":"dayan-2022","a":"foundations-systems","t":"Spooky","y":2022,"s":"peer_reviewed","e":6,"d":"10.14778/3551793.3551853","v":"Proceedings of the VLDB Endowment","a_":["Dayan et al."],"bl":"https://doi.org/10.14778/3551793.3551853","pc":1,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"Spooky demonstrates that partial compaction in RocksDB suffers from boundary-skewed overlapping files (inflating WAF) and uncompacted dead data in cold ranges (inflating SAF)."}]},{"k":"dean-2013","a":"ai-operations","t":"The tail at scale","y":2013,"s":"peer_reviewed","e":4,"d":"10.1145/2408776.2408794","v":"Communications of the ACM","a_":["Dean & Barroso"],"bl":"https://doi.org/10.1145/2408776.2408794","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Dean and Barroso argue that in fan-out services, rare per-server latency hiccups dominate end-to-end latency: if 1 in 100 requests to one server is slow, a request that fans out to 100 servers is slow 63% of the time."},{"m":"agent draft (claude)","l":"operational translation","tx":"Fan-out amplification: with k sub-requests the probability of touching a slow component is 1-(1-p)^k. A single stalled tserver leading 1/N of tablets hits 1-(1-1/N)^k of multi-tablet YSQL statements (e.g. 41% at N=6, k=3), plus every transaction whose status tablet it leads. Magnitude of the stall, not its frequency, sets how slow those requests are."}]},{"k":"decan2021semver","a":"foundations-se","t":"What Do Package Dependencies Tell Us About Semantic Versioning?","y":2021,"s":"peer_reviewed","e":1,"d":"10.1109/TSE.2019.2918315","v":"IEEE Transactions on Software Engineering","a_":["Decan & Mens"],"bl":"https://doi.org/10.1109/TSE.2019.2918315","pc":1,"sm":[{"m":"agent draft (gemini)","l":"semver and 0yz dependency traps","tx":"Empirically studies SemVer compliance across Cargo, npm, Packagist, and RubyGems, highlighting the perils of 0.y.z ('zero-space') permissive ranges vs lockfile drift."}]},{"k":"deep2020workloadcompression","a":"foundations-systems","t":"Comprehensive and efficient workload compression","y":2020,"s":"peer_reviewed","e":4,"d":"10.14778/3430915.3430931","v":"Proceedings of the VLDB Endowment","a_":["Deep et al."],"bl":"https://doi.org/10.14778/3430915.3430931","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Formalizes workload compression with two goals, representativity (feature-distribution match) and coverage (keep common queries and outliers), proves the optimum hard and gives a greedy algorithm with approximation guarantees, beating sampling and clustering baselines."}]},{"k":"defraeye2016staffing","a":"foundations-systems","t":"Staffing and scheduling under nonstationary demand for service: A literature review","y":2016,"s":"peer_reviewed","e":6,"d":"10.1016/j.omega.2015.04.002","v":"Omega","a_":["Defraeye & Nieuwenhuyse"],"bl":"https://doi.org/10.1016/j.omega.2015.04.002","pc":2,"sm":[{"m":"agent draft (claude)","l":"support ops review","tx":"Reviews staffing and scheduling under time-varying demand (SIPP, lagged/offered-load methods). For follow-the-sun support with shift handoffs, staffing should track offered load per hour band; P1 misses clustering at handoffs and APJ taking most night pages point to schedule design, not individual performance."},{"m":"agent draft (claude)","l":"vp support dashboard","tx":"Literature review of staffing and shift scheduling when demand varies over time (time-varying arrival rates), covering stationary approximations and their failures, lag-based and simulation methods. Relevant to follow-the-sun coverage: staffing should follow hour-of-week demand and shift hand-off windows, which is where our P1 misses cluster."}]},{"k":"dehghani-2022","a":"ai-operations","t":"Data Mesh: Delivering Data-Driven Value at Scale","y":2022,"s":"external","e":2,"a_":["Dehghani"],"pc":1,"sm":[{"m":"vault note","l":"summary","tx":"Dehghani, Z. (2022). _Data Mesh: Delivering Data-Driven Value at Scale_. Sebastopol, CA: O'Reilly Media."}]},{"k":"dejong2017zerodowntime","a":"distributed-systems","t":"Zero-Downtime SQL Database Schema Evolution for Continuous Deployment","y":2017,"s":"peer_reviewed","e":3,"d":"10.1109/ICSE-SEIP.2017.5","v":"2017 IEEE/ACM 39th International Conference on Software Engineering: Software Engineering in Practice Track (ICSE-SEIP)","a_":["Jong et al."],"bl":"https://doi.org/10.1109/ICSE-SEIP.2017.5"},{"k":"dellacqua2023jagged","a":"foundations-stats","t":"Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of Artificial Intelligence on Knowledge Worker Productivity and Quality","y":2023,"s":"peer_reviewed","e":5,"d":"10.2139/ssrn.4573321","v":"Elsevier BV","a_":["Dell'Acqua et al."],"bl":"https://doi.org/10.2139/ssrn.4573321","pc":3,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"BCG field experiment: GPT-4 raised consultant quality and speed on tasks inside the AI frontier but lowered correctness by about 19 pp on a task outside it. Warns that AI analysis notes can mislead engineers on out-of-distribution cases; the gallery needs a correctness score on the note, stratified by case type, not only speed."},{"m":"agent draft (claude)","l":"support human ai","tx":"Pre-registered field experiment with 758 BCG consultants. On tasks inside the AI frontier, consultants with GPT-4 finished more tasks (~12%), faster (~25%) and with higher quality (~40%); bottom-half performers gained most. On a task outside the frontier, AI users were ~19 points less likely to reach the correct answer."},{"m":"agent draft (claude)","l":"vp support dashboard","tx":"BCG field experiment with 758 consultants: GPT-4 raised productivity and quality on tasks inside the AI capability frontier, but consultants using AI were less likely to reach correct answers on a task outside it."}]},{"k":"deng2013cuped","a":"ai-operations","t":"Improving the Sensitivity of Online Controlled Experiments by Utilizing Pre-Experiment Data","y":2013,"s":"peer_reviewed","e":3,"d":"10.1145/2433396.2433413","v":"Proceedings of the Sixth ACM International Conference on Web Search and Data Mining (WSDM '13)","a_":["Deng et al."],"bl":"https://doi.org/10.1145/2433396.2433413","pc":3,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"CUPED reduces the variance of a treatment-effect estimate by regressing the outcome on a pre-period covariate (typically the same metric measured before exposure); variance falls by a factor of 1-rho^2, so the gain is entirely governed by how well pre-period data predicts the outcome."},{"m":"agent draft (claude)","l":"support human ai","tx":"CUPED uses pre-experiment covariates to reduce variance in controlled experiments, often substantially. Support implication: with small n, adjust ticket-level TTR comparisons for pre-period engineer and customer baselines, severity, product and category to shrink confidence intervals. showing the ready-to-close suggestion on a random half)."}]},{"k":"deng2016metricdev","a":"foundations-stats","t":"Data-Driven Metric Development for Online Controlled Experiments","y":2016,"s":"peer_reviewed","e":2,"d":"10.1145/2939672.2939700","v":"Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining","a_":["Deng & Shi"],"bl":"https://doi.org/10.1145/2939672.2939700","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Deng and Shi codify metric development as a data-driven process: metrics are judged on directionality (moves the right way when the product improves, validated on a corpus of past experiments with known outcomes) and sensitivity (probability of detecting a real effect, the product of how often a treatment moves it and statistical power)."}]},{"k":"devisser2020longitudinal","a":"calibration-conformal","t":"Towards a Theory of Longitudinal Trust Calibration in Human–Robot Teams","y":2019,"s":"peer_reviewed","e":4,"d":"10.1007/s12369-019-00596-x","v":"International Journal of Social Robotics","a_":["Visser et al."],"bl":"https://doi.org/10.1007/s12369-019-00596-x","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Longitudinal model of trust calibration with trust violations and repair as explicit events; accumulated relationship equity buffers violations, and repair acts (explanation, demonstrated change) are needed after a violation."}]},{"k":"dey2020usageindependent","a":"foundations-se","t":"Deriving a Usage-Independent Software Quality Metric","y":2020,"s":"preprint","e":2,"x":"2002.09989","v":"arXiv:2002.09989","a_":["Dey & Mockus"],"bl":"https://arxiv.org/abs/2002.09989","pc":2,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Post-release defect counts scale with usage; normalising by usage gives a quality metric independent of adoption. Relevant to relrate and sigrel: release comparisons must use exposure (node-months or call-home usage)."},{"m":"agent draft (claude)","l":"operational translation","tx":"Field defect counts scale with usage; a usage-independent quality metric divides defects by an exposure model of adoption. For Bundle Insights: tickets per bystander universe-quarter is an exposure proxy but must be aligned to release age to be fair between 2024.2 and 2025.2."}]},{"k":"dhaliwal2011crash","a":"foundations-telemetry","t":"Classifying field crash reports for fixing bugs: A case study of Mozilla Firefox","y":2011,"s":"peer_reviewed","e":6,"d":"10.1109/icsm.2011.6080800","v":"2011 27th IEEE International Conference on Software Maintenance (ICSM)","a_":["Dhaliwal et al."],"bl":"https://doi.org/10.1109/icsm.2011.6080800","pc":2,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Dhaliwal et al. (ICSM 2011) analyze Mozilla Firefox Socorro field crash reports and demonstrate that single-level bucketing by the top crashing signature frequently mixes distinct bugs (when crashes terminate inside generic memory/string/RPC routines) and that sub-clustering buckets using Levenshtein distance over the top 10 stack frames dramatically…"}]},{"k":"dibia-2023","a":"foundations-viz","t":"LIDA: A Tool for Automatic Generation of Grammar-Agnostic Visualizations and Infographics using Large Language Models","y":2023,"s":"peer_reviewed","e":2,"d":"10.18653/v1/2023.acl-demo.11","v":"Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 3: System Demonstrations)","a_":["Dibia"],"bl":"https://doi.org/10.18653/v1/2023.acl-demo.11"},{"k":"dietvorst2015aversion","a":"foundations-stats","t":"Algorithm aversion: People erroneously avoid algorithms after seeing them err.","y":2015,"s":"peer_reviewed","e":3,"d":"10.1037/xge0000033","v":"Journal of Experimental Psychology: General","a_":["Dietvorst et al."],"bl":"https://doi.org/10.1037/xge0000033","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Five experiments: people who see an algorithm err lose confidence in it faster than in an erring human and choose the human forecaster even when the algorithm is more accurate. For support AI notes: one visible fabrication can drive engineers to disuse the whole note, so fabrication must be measured and reported separately from accuracy."}]},{"k":"ding-2014","a":"foundations-systems","t":"Mining Historical Issue Repositories to Heal Large-Scale Online Service Systems","y":2014,"s":"peer_reviewed","e":4,"d":"10.1109/DSN.2014.39","v":"2014 44th Annual IEEE/IFIP International Conference on Dependable Systems and Networks","a_":["Ding et al."],"bl":"https://doi.org/10.1109/DSN.2014.39","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"DSN 2014: mines historical issue repositories (signatures plus healing actions) to suggest mitigation for new incidents; retrieval by symptom signature beats free text. Supports keying past-case lookup on WTL template ids and resolution actions."}]},{"k":"ding2020desensitization","a":"security-governance","t":"DESENSITIZATION: Privacy-Aware and Attack-Preserving Crash Report","y":2020,"s":"peer_reviewed","e":2,"d":"10.14722/ndss.2020.24428","v":"Proceedings 2020 Network and Distributed System Security Symposium","a_":["Ding et al."],"bl":"https://doi.org/10.14722/ndss.2020.24428","pc":3,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Ding et al. (NDSS 2020) present DESENSITIZATION, showing that crash reports can be sanitized of sensitive user inputs while preserving the exact crashing instruction, signal (`si_signo`), register state, and call-stack fault path."}]},{"k":"dintyala2020sqlcheck","a":"foundations-systems","t":"SQLCheck: Automated Detection and Diagnosis of SQL Anti-Patterns","y":2020,"s":"peer_reviewed","e":3,"d":"10.1145/3318464.3389754","v":"Proceedings of the 2020 ACM SIGMOD International Conference on Management of Data","a_":["Dintyala et al."],"bl":"https://doi.org/10.1145/3318464.3389754","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"SQLCheck ranks anti-patterns by impact using query, schema and data context. Our regex flags on WTL templates lack schema/data context, so they are candidate markers, not diagnoses; ORDER BY without LIMIT and SELECT * fire on most OLTP shapes."}]},{"k":"dinur2003revealing","a":"security-governance","t":"Revealing information while preserving privacy","y":2003,"s":"peer_reviewed","e":5,"d":"10.1145/773153.773173","v":"Proceedings of the twenty-second ACM SIGMOD-SIGACT-SIGART symposium on Principles of database systems","a_":["Dinur & Nissim"],"bl":"https://doi.org/10.1145/773153.773173","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Proves that answering too many subset-sum queries too accurately allows reconstruction of the database. The gallery publishes ~60 exact cross-cuts of the same ~67-org population; this result is why differencing across cards is a real, not theoretical, risk for any external release."}]},{"k":"dixon2010ces","a":"ai-operations","t":"Stop Trying to Delight Your Customers","y":2010,"s":"external","e":2,"v":"Harvard Business Review","a_":["Dixon et al."],"pc":1,"sm":[{"m":"agent draft (claude)","l":"support ops review","tx":"Service loyalty is driven more by reducing customer effort (repeat contacts, channel switching, re-explaining) than by delight. For B2B support dashboards: the operational proxies of effort are reopen/re-contact rate, number of customer replies to resolution, and handoffs/reassignments; a CSAT tile with n of a few dozen per year is not a substitute."}]},{"k":"dmitriev2016measuringmetrics","a":"foundations-other","t":"Measuring Metrics","y":2016,"s":"peer_reviewed","e":3,"d":"10.1145/2983323.2983356","v":"Proceedings of the 25th ACM International on Conference on Information and Knowledge Management","a_":["Dmitriev & Wu"],"bl":"https://doi.org/10.1145/2983323.2983356","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Metrics need validated directionality and sensitivity before they steer decisions."}]},{"k":"do-2013","a":"ai-operations","t":"Limplock","y":2013,"s":"peer_reviewed","e":3,"d":"10.1145/2523616.2523627","v":"Proceedings of the 4th annual Symposium on Cloud Computing","a_":["Do et al."],"bl":"https://doi.org/10.1145/2523616.2523627","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"fsync T7: a limping component stalls a scale-out system because detectors see only fail-stop. Matches DB-22970 (the stalled leader keeps its lease); 'peer detected' and 'node lost leadership' columns measure that gap."}]},{"k":"dohmen2008social","a":"sports-analytics","t":"The Influence of Social Forces: Evidence from the Behavior of Football Referees","y":2008,"s":"peer_reviewed","e":2,"d":"10.1111/j.1465-7295.2007.00112.x","v":"Economic Inquiry","a_":["Dohmen"],"bl":"https://doi.org/10.1111/j.1465-7295.2007.00112.x","pc":3,"sm":[{"m":"vault note","l":"summary","tx":"Bundesliga replication of the stoppage-time result, with the running-track/stadium-geometry variation used to separate crowd proximity from crowd size."}]},{"k":"dong-2021","a":"distributed-systems","t":"RocksDB: Evolution of Development Priorities in a Key-Value Store Serving Large-Scale Applications.","y":2021,"s":"peer_reviewed","e":6,"d":"10.1145/3483840","v":"ACM Transactions on Storage","a_":["Dong & Stumm"],"bl":"https://doi.org/10.1145/3483840","pc":3,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"DocDB is a RocksDB fork; RocksDB's evolution toward space efficiency, compaction tuning and write-stall control explains YB tserver CPU/IO hot spots (compaction, flush, block cache) and YB's own SST-file-count throttling (sst_files_soft_limit 24/hard 48) that replaces RocksDB L0 stall triggers."},{"m":"agent draft (claude)","l":"operational translation","tx":"RocksDB priorities (space amp, CPU, compaction) inform DocDB tuning; for query profiling the operational takeaway is that block-cache residency and SST file count drive cold vs warm read latency, so a profiler must report warm/cold runs separately and check SST counts before blaming the plan."}]},{"k":"doshi1986vacations","a":"foundations-telemetry","t":"Queueing systems with vacations ? A survey","y":1986,"s":"peer_reviewed","e":1,"d":"10.1007/BF01149327","v":"Queueing Systems","a_":["Doshi"],"bl":"https://doi.org/10.1007/BF01149327","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Survey of queues with server vacations. Key result used for fsync stalls: stochastic decomposition, waiting time = ordinary M/G/1 wait + residual vacation time E[V^2]/(2E[V]). A long stall leaves a backlog that drains at rate (mu - lambda), so the after-effect lasts B*rho/(1-rho);"}]},{"k":"drupsteen2014learning","a":"foundations-telemetry","t":"What Is Learning? A Review of the Safety Literature to Define Learning from Incidents, Accidents and Disasters","y":2014,"s":"peer_reviewed","e":1,"d":"10.1111/1468-5973.12039","v":"Journal of Contingencies and Crisis Management","a_":["Drupsteen & Guldenmund"],"bl":"https://doi.org/10.1111/1468-5973.12039","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Review of the safety literature defining learning from incidents as a cycle: report, analyse, plan and implement actions, then evaluate whether the change worked. Most organisations stop at analysis. For CE: an RCA document is the analysis step only;"}]},{"k":"du2016spell","a":"foundations-telemetry","t":"Spell: Streaming Parsing of System Event Logs","y":2016,"s":"peer_reviewed","e":9,"d":"10.1109/ICDM.2016.0103","v":"2016 IEEE 16th International Conference on Data Mining (ICDM)","a_":["Du & Li"],"bl":"https://doi.org/10.1109/ICDM.2016.0103","pc":3,"sm":[{"m":"agent draft (gemini)","l":"log template mining","tx":"Du and Li (ICDM 2016, Spell) formulate online log parsing as Longest Common Subsequence (LCS) matching, maintaining a prefix tree of LCSObjects alongside an inverted index of tokens so candidate templates can be pruned before running O(|m| * |t|) dynamic programming LCS alignment."},{"m":"agent draft (gemini)","l":"methodology","tx":"Introduces Spell (Streaming Parsing of System Event Logs), an online longest-common-subsequence (LCS) prefix-tree log parser that processes log streams in a single pass. Demonstrates that online template extraction must bound state per log signature so high-velocity log bursts can be aggregated in O(1) amortized time per event."}]},{"k":"duan-1983-demand","a":"foundations-stats","t":"A Comparison of Alternative Models for the Demand for Medical Care","y":1983,"s":"peer_reviewed","e":3,"d":"10.1080/07350015.1983.10509330","v":"Journal of Business &amp; Economic Statistics","a_":["Duan et al."],"bl":"https://doi.org/10.1080/07350015.1983.10509330","pc":1,"sm":[{"m":"agent draft (claude)","l":"support effort scoring","tx":"The RAND Health Insurance Experiment two-part model: P(any use) times E[spend | use], with a log-scale second part retransformed by smearing. It beat one-equation models on split-sample forecast error. This is the template for a staged ticket-effort model, P(CE escalation) times E[hours | CE] plus (1 - P) times E[hours | no CE]."}]},{"k":"duan-1983-smearing","a":"foundations-stats","t":"Smearing Estimate: A Nonparametric Retransformation Method","y":1983,"s":"peer_reviewed","e":1,"d":"10.1080/01621459.1983.10478017","v":"Journal of the American Statistical Association","a_":["Duan"],"bl":"https://doi.org/10.1080/01621459.1983.10478017"},{"k":"durner2023cloud","a":"ai-operations","t":"Exploiting Cloud Object Storage for High-Performance Analytics","y":2023,"s":"peer_reviewed","e":5,"d":"10.14778/3611479.3611486","v":"Proceedings of the VLDB Endowment","a_":["Durner et al."],"bl":"https://doi.org/10.14778/3611479.3611486","pc":1,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Durner, Leis, and Neumann (PVLDB '23) establish the quantitative performance blueprint for analytical queries over Cloud Object Storage (S3/GCS): object stores exhibit high per-request TTFB latency (~15-50 ms per HTTP GET) and reach peak bandwidth only at request sizes >= 8-16 MiB across high concurrency."}]},{"k":"dwork2006calibrating","a":"foundations-stats","t":"Calibrating Noise to Sensitivity in Private Data Analysis","y":2006,"s":"peer_reviewed","e":4,"d":"10.1007/11681878_14","v":"Lecture Notes in Computer Science","a_":["Dwork et al."],"bl":"https://doi.org/10.1007/11681878_14","sm":[{"m":"vault note","l":"summary","tx":"Foundational epsilon-differential privacy for aggregate telemetry and histogram release"}]},{"k":"dwork2014algorithmic","a":"security-governance","t":"The Algorithmic Foundations of Differential Privacy","y":2014,"s":"peer_reviewed","e":2,"d":"10.1561/0400000042","v":"Foundations and Trends® in Theoretical Computer Science","a_":["Dwork & Roth"],"bl":"https://doi.org/10.1561/0400000042","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Reference text for (eps,delta)-DP, Laplace/Gaussian mechanisms, composition and sparse vector. For fleet benchmarks: org-level (user-level) sensitivity bounding plus Laplace noise on counts, with a fixed privacy budget per release, is the standard way to publish small-cell rates externally; composition across 70 cards is the practical constraint."}]},{"k":"dzindolet2003reliance","a":"foundations-se","t":"The role of trust in automation reliance","y":2003,"s":"peer_reviewed","e":1,"d":"10.1016/S1071-5819(03)00038-7","v":"International Journal of Human-Computer Studies","a_":["Dzindolet et al."],"bl":"https://doi.org/10.1016/S1071-5819(03)00038-7","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Experiments showing people distrust an automated aid after seeing it err, even when it outperforms them, unless told why it errs; explaining errors restored appropriate reliance. Supports a visible error taxonomy and changelog for the triage note."}]},{"k":"edmondson-1999","a":"security-governance","t":"Psychological Safety and Learning Behavior in Work Teams","y":1999,"s":"peer_reviewed","e":1,"d":"10.2307/2666999","v":"Administrative Science Quarterly","a_":["Edmondson"],"bl":"https://doi.org/10.2307/2666999","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Team psychological safety predicts learning behaviour (asking for help, discussing errors), which mediates team performance; coaching-oriented leadership builds it. A blameless review list and team-level aggregates support it; name-bearing load charts visible to peers can undermine it."}]},{"k":"elavsky-2022","a":"foundations-viz","t":"How accessible is my visualization? Evaluating visualization accessibility with Chartability","y":2022,"s":"peer_reviewed","e":6,"d":"10.1111/cgf.14522","v":"Computer Graphics Forum","a_":["Elavsky et al."],"bl":"https://doi.org/10.1111/cgf.14522","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Chartability (POUR+CAF, 50 heuristics) applied to the gallery: passes keyboard navigation of marks, a table alternative and text contrast; fails a bypass link, label-in-name on the theme button, and concise descriptions (aria-describedby median 525 chars, duplicated on article and chart). Touch users get no tooltips."}]},{"k":"elavsky-2023","a":"foundations-viz","t":"Data Navigator: An Accessibility-Centered Data Navigation Toolkit","y":2023,"s":"peer_reviewed","e":4,"d":"10.1109/TVCG.2023.3327393","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Elavsky et al."],"bl":"https://doi.org/10.1109/TVCG.2023.3327393"},{"k":"elliott2017nonprob","a":"foundations-stats","t":"Inference for Nonprobability Samples","y":2017,"s":"peer_reviewed","e":3,"d":"10.1214/16-STS598","v":"Statistical Science","a_":["Elliott & Valliant"],"bl":"https://doi.org/10.1214/16-STS598","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"PM seat (council 5): Bundles are a nonprobability sample of the YB fleet. Elliott and Valliant set out quasi-randomisation (pseudo-inclusion weights from a reference probability sample) and superpopulation modelling as the two routes to fleet-level estimates."}]},{"k":"en-18286","a":"security-governance","t":"EN 18286:2026 (prEN 18286) Artificial intelligence - Quality management system for EU AI Act regulatory purposes","y":2026,"s":"external","e":2,"v":"CEN-CENELEC / Official Journal of the European Union (Decision (EU) 2026/2274)","a_":["CEN/CENELEC"],"bl":"https://eur-lex.europa.eu/eli/dec_impl/2026/2274/oj/eng","pc":1,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"EN 18286:2026 (developed by CEN/CENELEC JTC 21 as prEN 18286, passed formal vote July 10, 2026, and cited in the Official Journal via Commission Implementing Decision (EU) 2026/2274 on October 19, 2026) is the first harmonised European standard conferring legal presumption of conformity with Article 17 (Quality Management System) of the EU AI Act…"}]},{"k":"enders-2001","a":"foundations-stats","t":"The Relative Performance of Full Information Maximum Likelihood Estimation for Missing Data in Structural Equation Models","y":2001,"s":"peer_reviewed","e":1,"d":"10.1207/S15328007SEM0803_5","v":"Structural Equation Modeling: A Multidisciplinary Journal","a_":["Enders & Bandalos"],"bl":"https://doi.org/10.1207/S15328007SEM0803_5"},{"k":"es2024ragas","a":"ai-operations","t":"Ragas: Automated Evaluation of Retrieval Augmented Generation","y":2023,"s":"preprint","e":2,"x":"2309.15217","v":"arXiv:2309.15217","a_":["Es et al."],"bl":"https://arxiv.org/abs/2309.15217","pc":2,"sm":[{"m":"agent draft (claude)","l":"llmops","tx":"Reference-free RAG metrics computed by an LLM: faithfulness (share of answer claims supported by retrieved context), answer relevance, and context relevance; later library versions add context precision/recall."},{"m":"agent draft (claude)","l":"eval kpis","tx":"Ragas defines reference-free RAG metrics computed by an LLM: faithfulness (fraction of answer statements supported by retrieved context), answer relevance, and context relevance. cited log spans/ticket text), but scores must be validated against engineer labels before use as KPIs; Langfuse ships Ragas-based managed evaluators."}]},{"k":"etcd2026hardware","a":"foundations-systems","t":"Hardware recommendations (etcd v3.5 operations guide)","y":2026,"s":"external","e":1,"a_":["etcd authors"],"bl":"https://etcd.io/docs/v3.5/op-guide/hardware/","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"etcd: consensus is very sensitive to disk write latency; slow fsync times out heartbeats and causes elections. The vendor guidance assumes the slow node loses leadership, the opposite of YB's lease-holding leader on a stalled disk."}]},{"k":"euaiact2024","a":"security-governance","t":"Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonised rules on artificial intelligence (Artificial Intelligence Act)","y":2024,"s":"external","e":2,"a_":["EU Parliament & Council"],"bl":"https://eur-lex.europa.eu/eli/reg/2024/1689/oj/eng","pc":1,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Regulation (EU) 2024/1689 (EU AI Act) establishes binding product-safety obligations for high-risk AI systems (Chapter III, Section 2: Art. 9 Risk management system, Art. 10 Data and data governance, Art. 11 Technical documentation, Art. 12 Record-keeping/logging, Art. 13 Transparency to deployers, Art. 14 Human oversight, Art."}]},{"k":"fang2022onlineschemaevo","a":"foundations-systems","t":"Online schema evolution is (almost) free for snapshot databases","y":2022,"s":"peer_reviewed","e":3,"d":"10.14778/3565816.3565818","v":"Proceedings of the VLDB Endowment","a_":["Hu et al."],"bl":"https://doi.org/10.14778/3565816.3565818"},{"k":"farrington-1996","a":"foundations-stats","t":"A Statistical Algorithm for the Early Detection of Outbreaks of Infectious Disease","y":1996,"s":"peer_reviewed","e":4,"d":"10.2307/2983331","v":"Journal of the Royal Statistical Society. Series A (Statistics in Society)","a_":["Farrington et al."],"bl":"https://doi.org/10.2307/2983331","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Quasi-Poisson regression on comparable past weeks gives a threshold for this week's count; alarms when the count exceeds it, downweighting past outbreaks. For YB fleet surveillance the analogue is per-template org counts per month;"}]},{"k":"farrington2009sccsevent","a":"foundations-stats","t":"Case series analysis for censored, perturbed, or curtailed post-event exposures","y":2008,"s":"peer_reviewed","e":1,"d":"10.1093/biostatistics/kxn013","v":"Biostatistics","a_":["Farrington et al."],"bl":"https://doi.org/10.1093/biostatistics/kxn013","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Extends the self-controlled case series to exposures whose occurrence depends on prior events (censored, perturbed or curtailed post-event exposure). Relevant for upgrades as exposure and tickets as events: tickets prompt upgrades, violating standard SCCS."}]},{"k":"feinstein1990kappa","a":"ai-operations","t":"High Agreement but Low Kappa: I. The Problems of Two Paradoxes","y":1990,"s":"peer_reviewed","e":2,"d":"10.1016/0895-4356(90)90158-L","v":"Journal of Clinical Epidemiology","a_":["Feinstein & Cicchetti"],"bl":"https://doi.org/10.1016/0895-4356(90)90158-L","pc":2},{"k":"fernandes2018quantiledotplots","a":"foundations-viz","t":"Uncertainty Displays Using Quantile Dotplots or CDFs Improve Transit Decision-Making","y":2018,"s":"peer_reviewed","e":9,"d":"10.1145/3173574.3173718","v":"Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems","a_":["Fernandes et al."],"bl":"https://doi.org/10.1145/3173574.3173718","pc":1,"sm":[{"m":"agent draft (claude)","l":"dashboard design review","tx":"Incentivised experiment: CDF and quantile-dotplot displays of a waiting-time distribution beat point estimates and intervals for threshold decisions. Directly applicable to SLA feasibility: an empirical CDF of P1 response minutes with vertical lines at 5/15/30 answers \"what % within X\" for any X. It does this better than a bar chart of four percentiles."}]},{"k":"few2013dashboard","a":"foundations-viz","t":"Information Dashboard Design: Displaying Data for At-a-Glance Monitoring","y":2013,"s":"external","e":1,"v":"Analytics Press","a_":["Few"],"pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Few: a dashboard is a single-screen, at-a-glance display of the few measures needed to act; failures include exceeding one screen, burying the key measure, excessive detail, and decoration. Applied to Bundle Insights: the 49-card page is a report, not a dashboard; it needs a one-screen executive layer above it."}]},{"k":"fisch2024stratppi","a":"ai-operations","t":"Stratified Prediction-Powered Inference for Effective Hybrid Evaluation of Language Models","y":2024,"s":"peer_reviewed","e":2,"v":"Advances in Neural Information Processing Systems 37 (NeurIPS 2024)","a_":["Fisch et al."],"bl":"https://proceedings.neurips.cc/paper_files/paper/2024/hash/c9fcd02e6445c7dfbad6986abee53d0d-Abstract-Conference.html","pc":5,"sm":[{"m":"vault note","l":"summary","tx":"arXiv 2406.04291 carries the shorter title ``Stratified Prediction-Powered Inference for Hybrid Language Model Evaluation''"}]},{"k":"forsgren-2021","a":"foundations-se","t":"The SPACE of Developer Productivity","y":2021,"s":"peer_reviewed","e":2,"d":"10.1145/3454122.3454124","v":"Queue","a_":["Forsgren et al."],"bl":"https://doi.org/10.1145/3454122.3454124","pc":1,"sm":[{"m":"agent draft (claude)","l":"support human ai","tx":"SPACE framework: Satisfaction and well-being, Performance, Activity, Communication and collaboration, Efficiency and flow. Productivity cannot be captured by one metric. Use at least three dimensions, mix perceptual (survey) and system data, and avoid activity counts as performance proxies or for individual ranking."}]},{"k":"foukarakis2009deep","a":"security-governance","t":"Deep packet anonymization","y":2009,"s":"peer_reviewed","e":1,"d":"10.1145/1519144.1519147","v":"Proceedings of the Second European Workshop on System Security","a_":["Foukarakis et al."],"bl":"https://doi.org/10.1145/1519144.1519147","pc":2,"sm":[{"m":"agent draft (gemini)","l":"format driven log redaction","tx":"Classic EuroSec '09 work on protocol-parser-driven anonymization: instead of running expensive unanchored regexes over raw payloads, the parser decomposes records into named protocol fields and applies field-specific anonymization policies."}]},{"k":"franconeri2021science","a":"foundations-viz","t":"The Science of Visual Data Communication: What Works","y":2021,"s":"peer_reviewed","e":6,"d":"10.1177/15291006211051956","v":"Psychological Science in the Public Interest","a_":["Franconeri et al."],"bl":"https://doi.org/10.1177/15291006211051956","pc":2,"sm":[{"m":"agent draft (claude)","l":"dashboard design review","tx":"Synthesis of what works in visual communication. Direct labels beat legends; redundant encoding helps colour-deficient readers; a takeaway title or annotation focuses attention; uncertainty should be shown; dual axes and unaligned comparisons mislead. Used as the general benchmark for the support suite's text, colour and annotation practice."},{"m":"agent draft (gemini)","l":"operational translation","tx":"Synthesizes perceptual psychology into concrete chart engineering rules: replace detached legend boxes with direct on-bar or beside-bar labels to eliminate saccadic eye-movement overhead; enforce colorblind-safe contrast (Okabe-Ito Navy #0072B2, Teal #009E73, Vermilion #D55E00, Amber #E69F00); group comparisons into 1-4 core visual chunks;"}]},{"k":"frangakis2002principal","a":"foundations-stats","t":"Principal Stratification in Causal Inference","y":2002,"s":"peer_reviewed","e":3,"d":"10.1111/j.0006-341X.2002.00021.x","v":"Biometrics","a_":["Frangakis & Rubin"],"bl":"https://doi.org/10.1111/j.0006-341X.2002.00021.x","sm":[{"m":"vault note","l":"summary","tx":"Principal stratification on post-treatment variables"}]},{"k":"fu-2023","a":"foundations-stats","t":"GPT2SP: A Transformer-Based Agile Story Point Estimation Approach","y":2023,"s":"peer_reviewed","e":2,"d":"10.1109/TSE.2022.3158252","v":"IEEE Transactions on Software Engineering","a_":["Fu & Tantithamthavorn"],"bl":"https://doi.org/10.1109/TSE.2022.3158252"},{"k":"fu2009execution","a":"foundations-telemetry","t":"Execution Anomaly Detection in Distributed Systems through Unstructured Log Analysis","y":2009,"s":"peer_reviewed","e":14,"d":"10.1109/icdm.2009.60","v":"2009 Ninth IEEE International Conference on Data Mining","a_":["Fu et al."],"bl":"https://doi.org/10.1109/icdm.2009.60","pc":2,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Fu, Lou et al. (ICDM 2009) convert unstructured distributed system logs into template keys and learn finite-state execution automata and temporal transition invariants across log sequences."}]},{"k":"gallifant-2025","a":"ai-operations","t":"The TRIPOD-LLM reporting guideline for studies using large language models","y":2025,"s":"peer_reviewed","e":2,"d":"10.1038/s41591-024-03425-5","v":"Nature Medicine","a_":["Gallifant et al."],"bl":"https://doi.org/10.1038/s41591-024-03425-5","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"TRIPOD-LLM: reporting checklist for studies of LLMs incl. evaluation-only studies: model identity/version, prompts, leakage/contamination, human evaluation protocol and agreement, LLM-as-evaluator validation."}]},{"k":"ganesan2017redundancy","a":"ai-operations","t":"Redundancy Does Not Imply Fault Tolerance: Analysis of Distributed Storage Reactions to Single Errors and Corruptions","y":2017,"s":"peer_reviewed","e":2,"v":"15th USENIX Conference on File and Storage Technologies (FAST 17)","a_":["Ganesan et al."],"bl":"https://www.usenix.org/conference/fast17/technical-sessions/presentation/ganesan","pc":3,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Injecting single-node file-system faults (corruption, read/write errors) into 8 distributed stores shows redundancy rarely saves them: local faults propagate, crash nodes or lose data. Supports a YB DST campaign that injects WAL fsync delay and EIO per node, not only process kills."}]},{"k":"ganesan2021cad","a":"ai-operations","t":"Strong and Efficient Consistency with Consistency-aware Durability","y":2021,"s":"peer_reviewed","e":2,"d":"10.1145/3423138","v":"ACM Transactions on Storage","a_":["Ganesan et al."],"bl":"https://doi.org/10.1145/3423138","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Consistency-aware durability lets replicas acknowledge before fsync while reads stay strong. Council-5 DocDB lens: YugabyteDB's default buffered WAL (async sync at 0.6 s, forced at 1 s) is a related latency-for-durability trade;"}]},{"k":"gans-2003","a":"foundations-stats","t":"Telephone Call Centers: Tutorial, Review, and Research Prospects","y":2003,"s":"peer_reviewed","e":11,"d":"10.1287/msom.5.2.79.16071","v":"Manufacturing &amp; Service Operations Management","a_":["Gans et al."],"bl":"https://doi.org/10.1287/msom.5.2.79.16071","pc":3,"sm":[{"m":"agent draft (claude)","l":"metric statistics review","tx":"Call-centre tutorial: service level (fraction answered within a threshold) must be read together with abandonment, arrival-rate variability (doubly stochastic, overdispersed arrivals with strong time-of-day/day-of-week seasonality) and staffing."},{"m":"agent draft (claude)","l":"support ops dashboard","tx":"Canonical call-center OR tutorial: Erlang C (M/M/c) and Erlang A (with abandonment), service-level metrics (TSF: % answered within t), forecasting, staffing/scheduling, skills-based routing."},{"m":"agent draft (claude)","l":"support ops review","tx":"Canonical tutorial on call-centre operations: arrival forecasting, staffing (square-root safety staffing, Erlang C/A), service-level metrics (TSF, ASA, abandonment) and the forecasting-staffing-scheduling chain."}]},{"k":"gao2018crashrecovery","a":"foundations-telemetry","t":"An empirical study on crash recovery bugs in large-scale distributed systems","y":2018,"s":"peer_reviewed","e":7,"d":"10.1145/3236024.3236030","v":"Proceedings of the 2018 26th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering","a_":["Gao et al."],"bl":"https://doi.org/10.1145/3236024.3236030","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"103 crash-recovery bugs in 4 distributed systems: most stem from incorrect handling of crash/reboot timing and persisted state; many need only one crash. Relevant to YB FATAL and ERROR buckets around tablet bootstrap / remote bootstrap ('Failed to destroy regular DB', 'Not changing its role ... remote bootstrap')."}]},{"k":"garfinkel2019reconstruction","a":"foundations-viz","t":"Understanding database reconstruction attacks on public data","y":2019,"s":"peer_reviewed","e":3,"d":"10.1145/3287287","v":"Communications of the ACM","a_":["Garfinkel et al."],"bl":"https://doi.org/10.1145/3287287","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Explains, from the US Census Bureau's experience, how many consistent published aggregates can be solved jointly to reconstruct underlying records, motivating formal privacy (differential privacy) for 2020. Operational lesson: each extra cross-tab or margin raises reconstruction risk even when every cell passes a threshold rule;"}]},{"k":"garnett2002erlanga","a":"foundations-stats","t":"Designing a Call Center with Impatient Customers","y":2002,"s":"peer_reviewed","e":4,"d":"10.1287/msom.4.3.208.7753","v":"Manufacturing &amp; Service Operations Management","a_":["Garnett et al."],"bl":"https://doi.org/10.1287/msom.4.3.208.7753","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Erlang-A adds exponential patience; staffing at offered load plus a square-root safety term is robust once abandonment is modelled. In async B2B support the abandonment analogue is the customer going quiet when asked for a support bundle (44% of agent bundle requests in a May-Oct 2026 Yugabyte cohort never produced a downloaded bundle), so…"}]},{"k":"gartner-aiready-2026","a":"ai-operations","t":"Lack of AI-Ready Data Puts AI Projects at Risk (60% Abandonment Forecast Through 2026) & Gartner Data and Analytics Summit 2026 Findings on Governed Semantic Layers for AI Agents","y":2025,"s":"external","e":2,"a_":["Gartner"],"pc":1,"sm":[{"m":"vault note","l":"summary","tx":"Gartner (2025-2026). _Lack of AI-Ready Data Puts AI Projects at Risk (60% Abandonment Forecast Through 2026) & Gartner Data and Analytics Summit 2026 Findings on Governed Semantic Layers for AI Agents_. Stamford, CT: Gartner, Inc."}]},{"k":"gaver1962interrupted","a":"foundations-stats","t":"A Waiting Line with Interrupted Service, Including Priorities","y":1962,"s":"peer_reviewed","e":1,"d":"10.1111/j.2517-6161.1962.tb00438.x","v":"Journal of the Royal Statistical Society Series B: Statistical Methodology","a_":["Gaver"],"bl":"https://doi.org/10.1111/j.2517-6161.1962.tb00438.x","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Classic M/G/1 with service interruptions (server breakdowns). The completion time of a job grows with the second moment of interruption length, so for a fixed fraction of time interrupted, a few long interruptions inflate waiting far more than many short ones."}]},{"k":"gebru2021datasheets","a":"security-governance","t":"Datasheets for Datasets","y":2021,"s":"peer_reviewed","e":4,"d":"10.1145/3458723","v":"Communications of the ACM","a_":["Gebru et al."],"bl":"https://doi.org/10.1145/3458723","pc":1,"sm":[{"m":"agent draft (claude)","l":"iso42001 audit","tx":"Datasheets document a dataset's motivation, composition, collection process, preprocessing, uses, distribution and maintenance. In ISO 42001 terms this is the evidence for A.7.2-A.7.6 (data acquisition, quality, provenance, preparation) and A.4.3 data resources."}]},{"k":"geifman2017selective","a":"calibration-conformal","t":"Selective Classification for Deep Neural Networks","y":2017,"s":"external","e":1,"v":"Advances in Neural Information Processing Systems 30 (NeurIPS)","a_":["Geifman & El-Yaniv"],"bl":"https://papers.nips.cc/paper/7073-selective-classification-for-deep-neural-networks","pc":3,"sm":[{"m":"agent draft (gemini)","l":"ai calibration s13","tx":"Across the 30-case Round-6 cohort (439 findings), post-hoc logit-space Bayesian fusion under the Prompt-Exclusion Law achieves finding-level AUROC=0.9424, AURC=0.6950 (E-AURC=0.2923, Brier=0.0918) vs."}]},{"k":"geiger2013editsessions","a":"foundations-other","t":"Using edit sessions to measure participation in wikipedia","y":2013,"s":"peer_reviewed","e":1,"d":"10.1145/2441776.2441873","v":"Proceedings of the 2013 conference on Computer supported cooperative work","a_":["Geiger & Halfaker"],"bl":"https://doi.org/10.1145/2441776.2441873","pc":1,"sm":[{"m":"agent draft (claude)","l":"effort estimation","tx":"Closest methodological template for converting a public activity trace into labor-hours. Edits are grouped into sessions with a 1-hour inter-edit cutoff (vetted with EM-fitted within/between-session inter-edit distributions), session duration is last-minus-first edit plus a pad of 430 s (the mean within-session inter-edit time) to account for work…"}]},{"k":"gelman-1999","a":"foundations-stats","t":"All maps of parameter estimates are misleading","y":1999,"s":"peer_reviewed","e":2,"d":"10.1002/(SICI)1097-0258(19991215)18:23<3221::AID-SIM312>3.0.CO;2-M","v":"Statistics in Medicine","a_":["Gelman & Price"],"bl":"https://doi.org/10.1002/(SICI)1097-0258(19991215)18:23<3221::AID-SIM312>3.0.CO;2-M","pc":1,"sm":[{"m":"agent draft (claude)","l":"metric statistics review","tx":"Maps or tables of raw unit-level rates are dominated at both extremes by the smallest units, so the best and worst look the most extreme purely through sampling noise; partial pooling (empirical/hierarchical Bayes) shrinks small units toward the mean."}]},{"k":"gelman-2006","a":"foundations-stats","t":"Prior distributions for variance parameters in hierarchical models (comment on article by Browne and Draper)","y":2006,"s":"peer_reviewed","e":1,"d":"10.1214/06-BA117A","v":"Bayesian Analysis","a_":["Gelman"],"bl":"https://doi.org/10.1214/06-BA117A","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Hierarchical variance parameters with few groups are poorly identified; moment and ML estimates collapse to zero; weakly informative half-t priors recommended. Relevant to fs_bands (1/alpha by moments, clamped at 0) and the collapsed funnel shrinkage."}]},{"k":"georges2007statistically","a":"foundations-stats","t":"Statistically rigorous java performance evaluation","y":2007,"s":"peer_reviewed","e":9,"d":"10.1145/1297027.1297033","v":"Proceedings of the 22nd annual ACM SIGPLAN conference on Object-oriented programming systems, languages and applications","a_":["Georges et al."],"bl":"https://doi.org/10.1145/1297027.1297033","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Common Java performance methodologies (best-of, single run) mislead; use confidence intervals across multiple VM invocations and separate startup from steady state. For profiler agents: report CIs, not just spread, and keep cold vs warm distinct."},{"m":"agent draft (gemini)","l":"benchmark methodology","tx":"Establishes statistically rigorous start-up vs. steady-state performance evaluation: warmup iterations must be discounted before recording steady-state measurements, and confidence intervals must be reported across multiple independent invocations to separate execution noise from real algorithmic speedups."}]},{"k":"ghosh2022fight","a":"ai-operations","t":"How to Fight Production Incidents? An Empirical Study on a Large-Scale Cloud Service","y":2022,"s":"peer_reviewed","e":9,"d":"10.1145/3542929.3563482","v":"Proceedings of the 13th Symposium on Cloud Computing (SoCC)","a_":["Ghosh et al."],"bl":"https://doi.org/10.1145/3542929.3563482","pc":2,"sm":[{"m":"agent draft (claude)","l":"coordination","tx":"Postmortem study of ~150 high-severity Microsoft Teams incidents across detection, root-causing and mitigation, identifying process gaps (manual steps, cross-team dependencies) that delay response."},{"m":"agent draft (claude)","l":"operational translation","tx":"Empirical study of 152 high-severity incidents in a large Microsoft cloud service: root causes split across code bugs, dependency failures, infrastructure and config; many incidents lacked the telemetry needed to mitigate quickly, and a sizeable share were not product bugs."}]},{"k":"giger-2010","a":"foundations-telemetry","t":"Predicting the fix time of bugs","y":2010,"s":"peer_reviewed","e":8,"d":"10.1145/1808920.1808933","v":"Proceedings of the 2nd International Workshop on Recommendation Systems for Software Engineering","a_":["Giger et al."],"bl":"https://doi.org/10.1145/1808920.1808933"},{"k":"gilardi2023chatgpt","a":"ai-operations","t":"ChatGPT Outperforms Crowd Workers for Text-Annotation Tasks","y":2023,"s":"peer_reviewed","e":3,"d":"10.1073/pnas.2305016120","v":"Proceedings of the National Academy of Sciences","a_":["Gilardi et al."],"bl":"https://doi.org/10.1073/pnas.2305016120","pc":3,"sm":[{"m":"agent draft (claude-sonnet-5-5)","l":"complexity measurement without effort","tx":"ChatGPT zero-shot outperformed MTurk crowd workers on several text-annotation tasks in accuracy and intercoder agreement, at a fraction of the cost. Supports using an LLM as a primary rater for rubric items, but the tasks were relatively simple classification and a 2023 model; it does not show reliability on expert-judgement constructs like complexity."}]},{"k":"glerum2009wer","a":"ai-operations","t":"Debugging in the (very) large","y":2009,"s":"peer_reviewed","e":16,"d":"10.1145/1629575.1629586","v":"Proceedings of the ACM SIGOPS 22nd symposium on Operating systems principles","a_":["Glerum et al."],"bl":"https://doi.org/10.1145/1629575.1629586","pc":8,"sm":[{"m":"agent draft (claude)","l":"contrasting view","tx":"WER is the canonical industrial field-failure corpus: automatic opt-in crash reports bucketed by heuristic signatures at billion-report scale, with explicit treatment of bucketing error and selective data collection."},{"m":"agent draft (claude)","l":"crash bucketing for release quality","tx":"Windows Error Reporting buckets billions of crash reports by a deterministic signature (module, offset, exception code) so engineers can rank bugs by hit count and track fixes across releases; bucketing heuristics trade over-splitting (one bug, many buckets) against over-condensing."},{"m":"agent draft (claude)","l":"operational translation","tx":"Common-good lens: fleet-wide value comes from aggregating the same failure signature across installations and ranking by reach."}]},{"k":"goddard-2012","a":"foundations-other","t":"Automation bias: a systematic review of frequency, effect mediators, and mitigators","y":2012,"s":"peer_reviewed","e":5,"d":"10.1136/amiajnl-2011-000089","v":"Journal of the American Medical Informatics Association","a_":["Goddard et al."],"bl":"https://doi.org/10.1136/amiajnl-2011-000089","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Systematic review of automation bias in clinical decision support: users accept wrong advice (commission) and miss unflagged problems (omission); mitigators include accountability, training and displaying confidence. For human-in-the-loop AI notes: engineer review is a control only if overrides and corrections are recorded."}]},{"k":"goel-1979","a":"foundations-stats","t":"Time-Dependent Error-Detection Rate Model for Software Reliability and Other Performance Measures","y":1979,"s":"peer_reviewed","e":1,"d":"10.1109/TR.1979.5220566","v":"IEEE Transactions on Reliability","a_":["Goel & Okumoto"],"bl":"https://doi.org/10.1109/TR.1979.5220566","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"NHPP reliability growth model with mean value function a(1-exp(-bt)): field failure intensity decays with exposure time after release as defects are found and fixed. Implies a newer release always looks worse in calendar-pooled windows; release comparisons must be day-0 aligned or model the growth curve."}]},{"k":"goeminne2014coevolving","a":"foundations-systems","t":"Co-evolving code-related and database-related changes in a data-intensive software system","y":2014,"s":"peer_reviewed","e":1,"d":"10.1109/CSMR-WCRE.2014.6747193","v":"2014 Software Evolution Week - IEEE Conference on Software Maintenance, Reengineering, and Reverse Engineering (CSMR-WCRE)","a_":["Goeminne et al."],"bl":"https://doi.org/10.1109/CSMR-WCRE.2014.6747193"},{"k":"gonzalez2004multitasking","a":"foundations-se","t":"\"Constant, constant, multi-tasking craziness\"","y":2004,"s":"peer_reviewed","e":3,"d":"10.1145/985692.985707","v":"Proceedings of the SIGCHI Conference on Human Factors in Computing Systems","a_":["González & Mark"],"bl":"https://doi.org/10.1145/985692.985707","pc":1,"sm":[{"m":"agent draft (claude)","l":"effort estimation","tx":"Shadowing study of information workers (analysts, developers, managers): about 3 minutes on an event before switching and roughly 10.5-12 minutes in a 'working sphere' (a thematically connected unit of work) before switching sphere."}]},{"k":"goodhart-1984","a":"foundations-stats","t":"Problems of Monetary Management: The UK Experience","y":1984,"s":"peer_reviewed","e":3,"d":"10.1007/978-1-349-17295-5_4","v":"Monetary Theory and Practice","a_":["Goodhart"],"bl":"https://doi.org/10.1007/978-1-349-17295-5_4"},{"k":"goodmanbacon2021decomp","a":"econometrics","t":"Difference-in-differences with variation in treatment timing","y":2021,"s":"peer_reviewed","e":7,"d":"10.1016/j.jeconom.2021.03.014","v":"Journal of Econometrics","a_":["Goodman-Bacon"],"bl":"https://doi.org/10.1016/j.jeconom.2021.03.014","pc":1,"sm":[{"m":"agent draft (claude)","l":"limitations caveats","tx":"TWFE DiD with staggered timing is a weighted mix of all 2x2 comparisons, including already-treated units as controls with possibly negative weights. Warns against pooled before/after-upgrade rate ratios across cohorts."}]},{"k":"graber2005diagerror","a":"foundations-other","t":"Diagnostic Error in Internal Medicine","y":2005,"s":"peer_reviewed","e":1,"d":"10.1001/archinte.165.13.1493","v":"Archives of Internal Medicine","a_":["Graber et al."],"bl":"https://doi.org/10.1001/archinte.165.13.1493","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"100 diagnostic-error cases: system-related and cognitive factors co-occur; premature closure (stopping at the first plausible diagnosis) is the dominant cognitive failure. For support triage: the key failure to score is a confident single cause that stops the investigation, not merely a wrong list."}]},{"k":"graefe2006sorting","a":"foundations-systems","t":"Implementing sorting in database systems","y":2006,"s":"peer_reviewed","e":4,"d":"10.1145/1132960.1132964","v":"ACM Computing Surveys","a_":["Graefe"],"bl":"https://doi.org/10.1145/1132960.1132964","pc":1,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Graefe's foundational survey (ACM CSUR '06) formalizes external merge sort invariants in database engines: replacement selection (generating runs ~2x memory buffer size on partially ordered inputs), optimal merge fan-in $F$ bounded by per-run read buffer size so seek/syscall overhead is amortized without thrashing memory, and forecasting/prefetching…"}]},{"k":"gramazio-2017","a":"foundations-viz","t":"Colorgorical: Creating discriminable and preferable color palettes for information visualization","y":2017,"s":"peer_reviewed","e":3,"d":"10.1109/TVCG.2016.2598918","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Gramazio et al."],"bl":"https://doi.org/10.1109/TVCG.2016.2598918"},{"k":"gray2005krspeed","a":"foundations-other","t":"The Role of Knowledge Repositories in Technical Support Environments: Speed Versus Learning in User Performance","y":2005,"s":"peer_reviewed","e":1,"d":"10.2753/MIS0742-1222220306","v":"Journal of Management Information Systems","a_":["GRAY & DURCIKOVA"],"bl":"https://doi.org/10.2753/MIS0742-1222220306","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Common-good lens: in technical support, repository use trades learning for speed. A triage note that surfaces the known fix maximises speed for the current ticket; whether engineers learn from it is a separate outcome. Pair known-issue reuse metrics with a learning metric (later unassisted handling of the same signature)."}]},{"k":"green-2007","a":"foundations-se","t":"Coping with Time‐Varying Demand When Setting Staffing Requirements for a Service System","y":2007,"s":"peer_reviewed","e":8,"d":"10.1111/j.1937-5956.2007.tb00164.x","v":"Production and Operations Management","a_":["Green et al."],"bl":"https://doi.org/10.1111/j.1937-5956.2007.tb00164.x","pc":1,"sm":[{"m":"agent draft (claude)","l":"support ops dashboard","tx":"Reviews stationary-independent-period-by-period (SIPP) and lagged/pointwise-stationary staffing for time-varying demand, showing when stationary Erlang approximations fail. Relevance: shift hand-offs and event windows (e.g., Black Friday) need time-varying staffing analysis rather than annual-average λ."}]},{"k":"greenland2005multiplebias","a":"foundations-stats","t":"Multiple-Bias Modelling for Analysis of Observational Data","y":2005,"s":"peer_reviewed","e":5,"d":"10.1111/j.1467-985X.2004.00349.x","v":"Journal of the Royal Statistical Society Series A: Statistics in Society","a_":["Greenland"],"bl":"https://doi.org/10.1111/j.1467-985X.2004.00349.x","pc":1,"sm":[{"m":"agent draft (claude)","l":"uncertainty quantification","tx":"Probabilistic (Monte Carlo) bias analysis: place prior distributions on bias parameters that the data cannot identify (here: session gap, per-comment minutes, unlogged-work multiplier, missing-channel fraction), draw them jointly, recompute the estimate, and add sampling error by bootstrap. Exactly the right tool when there is no ground truth;"}]},{"k":"gregg-2016","a":"ai-operations","t":"The flame graph","y":2016,"s":"peer_reviewed","e":4,"d":"10.1145/2909476","v":"Communications of the ACM","a_":["Gregg"],"bl":"https://doi.org/10.1145/2909476","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Flame graphs are the presentation layer for perf/eBPF stacks collected from yb-tserver and postgres backends; they are only as good as the stack walker: YB release binaries are LTO+PGO(+BOLT on x86) clang builds, so fp vs dwarf unwinding and BOLT-split .cold fragments change frame merging."}]},{"k":"gregg2012use","a":"ai-operations","t":"Thinking Methodically about Performance","y":2012,"s":"peer_reviewed","e":1,"d":"10.1145/2405116.2413037","v":"Queue","a_":["Gregg"],"bl":"https://doi.org/10.1145/2405116.2413037","pc":3,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Gregg's ACM Queue article contrasts anti-methods (streetlight, random change, blame-someone-else) with methods. The main one is the USE method: for every resource, check utilization, saturation and errors, using a resource list built from a functional block diagram (CPUs, memory, network interfaces, storage devices, controllers, interconnects)."},{"m":"agent draft (claude)","l":"methodology","tx":"USE maps onto bundle metrics: CPU via node_cpu_seconds/cpu_utime_irate; memory via generic_current_allocated_bytes vs server_memory_soft_limit and *_memory_pressure_rejections; disk via node_disk_io_time and log_sync_latency; RPC queues via rpcs_in_queue. Per-device disk saturation is blocked because the ETL drops the device label."},{"m":"agent draft (claude)","l":"sre platform","tx":"The USE method: for every resource, check utilization, saturation and errors. It applies to resources such as the Cloud SQL connection pool, the Prefect work-pool slots, LLM token-per-minute quotas and third-party API rate-limit headroom. RED, or the four golden signals, covers request-serving services such as the MCP server and the API."}]},{"k":"groen2017qualityexperts","a":"foundations-se","t":"Users — The Hidden Software Product Quality Experts?: A Study on How App Users Report Quality Aspects in Online Reviews","y":2017,"s":"peer_reviewed","e":2,"d":"10.1109/RE.2017.73","v":"2017 IEEE 25th International Requirements Engineering Conference (RE)","a_":["Groen et al."],"bl":"https://doi.org/10.1109/RE.2017.73","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"PM seat (council 5): Users report quality aspects (reliability, performance, usability) in feedback channels, i.e. feedback reveals non-functional requirements. Support tickets are a richer non-functional channel; supports treating fsync and upgrade cards as roadmap input."}]},{"k":"gu2023acto","a":"ai-operations","t":"Acto: Automatic End-to-End Testing for Operation Correctness of Cloud System Management","y":2023,"s":"peer_reviewed","e":1,"d":"10.1145/3600006.3613161","v":"Proceedings of the 29th Symposium on Operating Systems Principles","a_":["Gu et al."],"bl":"https://doi.org/10.1145/3600006.3613161","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Acto tests cloud-system operators end to end by generating state transitions and checking operation correctness, finding many bugs in popular operators. A method for YBA operator/helm QA that field data could seed with the real upgrade paths and storage classes seen in bundles."}]},{"k":"gulati2009parda","a":"ai-operations","t":"PARDA: Proportional Allocation of Resources for Distributed Storage Access","y":2009,"s":"peer_reviewed","e":1,"v":"7th USENIX Conference on File and Storage Technologies (FAST 09)","a_":["Gulati et al."],"bl":"https://www.usenix.org/conference/fast-09/parda-proportional-allocation-resources-distributed-storage-access","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"PARDA (VMware): hosts sharing a SAN array adapt their queue depth to observed latency to get proportional shares, the basis of vSphere Storage I/O Control. Implies that on a shared SAN, latency is a fabric-wide property: one host's burst raises every host's fsync latency, matching the Softbank 20107 all-hosts w_await pattern."}]},{"k":"gunawi2014cloudbugs","a":"foundations-telemetry","t":"What Bugs Live in the Cloud? A Study of 3000+ Issues in Cloud Systems","y":2014,"s":"peer_reviewed","e":9,"d":"10.1145/2670979.2670986","v":"Proceedings of the ACM Symposium on Cloud Computing","a_":["Gunawi et al."],"bl":"https://doi.org/10.1145/2670979.2670986","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"3000+ cloud-system issues: upgrade, configuration and hardware-failure handling bugs are major classes with severe outcomes; supports keeping upgrade and environment-triggered crashes separate in release QA metrics."}]},{"k":"gunawi2016cloudstop","a":"ai-operations","t":"Why Does the Cloud Stop Computing? Lessons from Hundreds of Service Outages","y":2016,"s":"peer_reviewed","e":9,"d":"10.1145/2987550.2987583","v":"Proceedings of the Seventh ACM Symposium on Cloud Computing (SoCC)","a_":["Gunawi et al."],"bl":"https://doi.org/10.1145/2987550.2987583","pc":4,"sm":[{"m":"agent draft (claude)","l":"empirical findings","tx":"597 public outage reports from 32 cloud services (2009-2015): storage and upgrade/config faults are leading root causes, redundancy often fails to mask them, and many outages last hours. Context for treating cloud volume freezes as expected events that the database must tolerate."},{"m":"agent draft (claude)","l":"operational translation","tx":"Gunawi et al. analysed 597 unplanned outages across 32 internet services (2009-2015) from public post-mortems and news, classifying root causes, impacts and fixes; upgrades, configuration and unknown causes are prominent and redundancy often fails to mask faults."}]},{"k":"gunawi2018failslow","a":"ai-operations","t":"Fail-Slow at Scale: Evidence of Hardware Performance Faults in Large Production Systems","y":2018,"s":"peer_reviewed","e":9,"v":"16th USENIX Conference on File and Storage Technologies (FAST 18)","a_":["Gunawi et al."],"bl":"https://www.usenix.org/conference/fast18/presentation/gunawi","pc":10,"sm":[{"m":"agent draft (claude)","l":"empirical findings","tx":"Fail-slow at scale (FAST'18): 101 incidents; slow hardware ranges from transient blips to permanent slowdowns, and stop-start/transient stalls are common. Frequency matters operationally because transient blips precede permanent faults, so counts of short stalls are a leading indicator, not an impact measure."},{"m":"agent draft (claude)","l":"operational translation","tx":"Fail-slow at scale: slow hardware is common and propagates. Council-5 DocDB lens: fsync/VOLUME.md shows the software side decides the impact: buffered WAL hides sub-second stalls (verified in log.cc: appends take no mutex; forced fsync at 1 s/1 MB blocks), so per-stall magnitude, not volume, drives write unavailability on default flags."},{"m":"agent draft (claude)","l":"operational translation","tx":"101 fail-slow hardware incidents in production; faults convert and cascade and are hard to localise because nodes keep responding. Supports the per-node environmental-hazard reading of the fsync POC and the ask for per-drive sync telemetry and host evidence in YBA bundles."}]},{"k":"guo2017calibration","a":"calibration-conformal","t":"On Calibration of Modern Neural Networks","y":2017,"s":"peer_reviewed","e":4,"v":"Proceedings of the 34th International Conference on Machine Learning (ICML 2017)","a_":["Guo et al."],"bl":"https://proceedings.mlr.press/v70/guo17a.html","pc":4,"sm":[{"m":"agent draft (gemini)","l":"r6 stakeholder s13 calibration","tx":"In Doctor v2.5 Round-6 (439 findings across 30 enterprise cases), post-hoc logit-space Bayesian fusion with temperature scaling (T=1.15) and FLEET_RULE_PRIORS (99.8% coverage, 438/439 findings) achieved finding-level AUROC=0.9424, E-AURC=0.2923 (-47.8% vs in-prompt prior injection), and Brier=0.0918 (-62.7%), with case-level Hits@1=70.0% (21/30) and…"},{"m":"agent draft (gemini)","l":"systems performance","tx":"Combined with the Prompt-Exclusion Law (assert_prompt_excludes_priors), logit-space fusion prevents the rho=+0.75 in-prompt base-rate inversion while guaranteeing strict Bayesian monotonicity with respect to fleet fire rate f_r."}]},{"k":"guzman2014feature","a":"foundations-se","t":"How Do Users Like This Feature? A Fine Grained Sentiment Analysis of App Reviews","y":2014,"s":"peer_reviewed","e":4,"d":"10.1109/RE.2014.6912257","v":"2014 IEEE 22nd International Requirements Engineering Conference (RE)","a_":["Guzman & Maalej"],"bl":"https://doi.org/10.1109/RE.2014.6912257","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"PM seat (council 5): Feature-level sentiment mining of app reviews ties feedback to named features. The support analogue maps tickets and IDEA links to named product features so pain can be ranked per feature, next to featadopt usage shares."}]},{"k":"haerem-2015","a":"foundations-other","t":"Task Complexity: Extending a Core Concept","y":2015,"s":"peer_reviewed","e":3,"d":"10.5465/amr.2013.0350","v":"Academy of Management Review","a_":["Hærem et al."],"bl":"https://doi.org/10.5465/amr.2013.0350","pc":1,"sm":[{"m":"agent draft (claude-sonnet-5-5)","l":"measurement","tx":"Extends Wood/Campbell: complexity as the number of pathways through the task's event network, so it grows non-additively with components and links. Reviews over 700 citing studies; nearly all use the Wood/Campbell typology without a network indicator."}]},{"k":"hainmueller-2012","a":"foundations-stats","t":"Entropy Balancing for Causal Effects: A Multivariate Reweighting Method to Produce Balanced Samples in Observational Studies","y":2012,"s":"peer_reviewed","e":1,"d":"10.1093/pan/mpr025","v":"Political Analysis","a_":["Hainmueller"],"bl":"https://doi.org/10.1093/pan/mpr025","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Entropy balancing finds unit weights that exactly match chosen covariate moments between groups, while staying as close as possible to uniform weights (minimum KL divergence). Unlike logistic propensity weighting, no balance iteration is needed, and extreme weights are less common."}]},{"k":"halfaker2015sessions","a":"foundations-other","t":"User Session Identification Based on Strong Regularities in Inter-activity Time","y":2015,"s":"peer_reviewed","e":3,"d":"10.1145/2736277.2741117","v":"Proceedings of the 24th International Conference on World Wide Web","a_":["Halfaker et al."],"bl":"https://doi.org/10.1145/2736277.2741117","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"fsync T7: sessions cut at the valley of a mixture over log inter-activity times. On 4 cached POC cases the stall-line gap histogram is unimodal (peak ~30 s, chronic cases) and episode counts fall steadily from gap 0 to 60 s with no plateau, while acute EA is one 43.4 s episode at any gap >= 1 s."},{"m":"agent draft (claude)","l":"effort estimation","tx":"Shows that inter-activity times in many user-initiated systems (Wikipedia, search, games, OSM, MovieLens) form a mixture of a within-session and a between-session mode, and that a rule-of-thumb inactivity threshold of about one hour separates them across domains. Recommends fitting the mixture to the local data rather than importing a global constant."}]},{"k":"halfin-1981","a":"foundations-stats","t":"Heavy-Traffic Limits for Queues with Many Exponential Servers","y":1981,"s":"peer_reviewed","e":9,"d":"10.1287/opre.29.3.567","v":"Operations Research","a_":["Halfin & Whitt"],"bl":"https://doi.org/10.1287/opre.29.3.567","pc":1,"sm":[{"m":"agent draft (claude)","l":"support ops dashboard","tx":"Establishes the Quality-and-Efficiency-Driven (QED) many-server heavy-traffic regime: with c ≈ R + β√R servers, P(wait) converges to a nontrivial limit. Relevance: square-root staffing applies to large pools;"}]},{"k":"hao2016tailatstore","a":"ai-operations","t":"The Tail at Store: A Revelation from Millions of Hours of Disk and SSD Deployments","y":2016,"s":"peer_reviewed","e":3,"v":"14th USENIX Conference on File and Storage Technologies (FAST 16)","a_":["Hao et al."],"bl":"https://www.usenix.org/conference/fast16/technical-sessions/presentation/hao","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"fsync T5-T7: a drive is slow when its latency exceeds 2x the median of its RAID-group peers at the same time; disks were slow 0.2% and SSDs 0.6% of the time; 1.5%/2.2% of RAID groups had at least one slow drive."},{"m":"agent draft (claude)","l":"operational translation","tx":"Millions of drive-hours: slow I/O is common, mostly short, concentrated in a minority of drives and recurrent. Mirrors YB bundle data: frequent moderate WAL stalls (1-5 s, 80% of episodes) cluster in chronic orgs, while the long tail (>=15 s, 6% of episodes) carries most blocked seconds and timeouts."}]},{"k":"hauer2020meaningful","a":"foundations-systems","t":"Meaningful Availability","y":2020,"s":"peer_reviewed","e":3,"v":"17th USENIX Symposium on Networked Systems Design and Implementation (NSDI 20)","a_":["Hauer et al."],"bl":"https://www.usenix.org/conference/nsdi20/presentation/hauer","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"PM seat (council 5): Windowed user-uptime separates one long outage from many short ones; directly supports VOLUME.md's GF1 reporting recommendation."}]},{"k":"hayes-2007","a":"foundations-viz","t":"Answering the Call for a Standard Reliability Measure for Coding Data","y":2007,"s":"peer_reviewed","e":2,"d":"10.1080/19312450709336664","v":"Communication Methods and Measures","a_":["Hayes & Krippendorff"],"bl":"https://doi.org/10.1080/19312450709336664","pc":1,"sm":[{"m":"agent draft (claude-sonnet-5-5)","l":"complexity measurement without effort","tx":"Argues for Krippendorff's alpha as a general reliability coefficient (any number of raters, missing data, nominal to ratio scales) and gives bootstrap CIs and guidance on interpreting alpha. Use ordinal alpha for the LLM-vs-human and LLM-vs-LLM comparisons on rubric items since it handles incomplete rating designs; report CI, not just a point estimate."}]},{"k":"hazem-2026-arxiv","a":"foundations-other","t":"Evaluating Decision Models for Text Annotation in Computational Social Science","y":2026,"s":"preprint","e":2,"x":"2609.24574","v":"arXiv:2609.24574","a_":["Ibrahim & Zaki"],"bl":"https://arxiv.org/abs/2609.24574","pc":1,"sm":[{"m":"agent draft (claude)","l":"system one calibration","tx":"Independent eval of a hosted decision model on the Ziems et al. CSS annotation tasks vs 19 LLMs. Decision model trails per-task best LLM on 14/15 tasks (median -11.6 macro-F1) at ~44x lower cost. Its confidence is better calibrated than verbalized confidence of 16/19 LLMs, but three frontier models have lower median calibration error (0.066 vs 0.157)."}]},{"k":"he-2018","a":"foundations-telemetry","t":"Identifying impactful service system problems via log analysis","y":2018,"s":"peer_reviewed","e":7,"d":"10.1145/3236024.3236083","v":"Proceedings of the 2018 26th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering","a_":["He et al."],"bl":"https://doi.org/10.1145/3236024.3236083","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"He et al. (Log3C, FSE'18, Microsoft) cluster log sequences cascadingly and correlate cluster counts with KPI degradation to surface impactful problems."}]},{"k":"he2016experience","a":"foundations-stats","t":"Experience Report: System Log Analysis for Anomaly Detection","y":2016,"s":"peer_reviewed","e":9,"d":"10.1109/issre.2016.21","v":"2016 IEEE 27th International Symposium on Software Reliability Engineering (ISSRE)","a_":["He et al."],"bl":"https://doi.org/10.1109/issre.2016.21","pc":2,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"He et al. (ISSRE 2016) evaluate supervised and unsupervised log anomaly detection methods on production system logs, showing that accurate log parsing and fixed-window/sliding-window event count matrices are prerequisite foundations before any downstream anomaly scoring."}]},{"k":"he2017drain","a":"ai-operations","t":"Drain: An Online Log Parsing Approach with Fixed Depth Tree","y":2017,"s":"peer_reviewed","e":25,"d":"10.1109/ICWS.2017.13","v":"Proceedings of the 2017 IEEE 24th International Conference on Web Services (ICWS '17)","a_":["He et al."],"bl":"https://doi.org/10.1109/ICWS.2017.13","pc":11,"sm":[{"m":"agent draft (claude)","l":"log template mining","tx":"fsync T5: Drain masks digit-bearing tokens and groups by length and leading tokens, so a variable-length op phrase is masked or splits templates. WTL's source-derived catalog with manual-force rows (e.g. 'Time spent sync call for <*>') is the lever if the S1 op is ever split at template level."},{"m":"agent draft (gemini)","l":"systems performance","tx":"Establishes online fixed-depth tree log template mining, which underpins WTL v6's template_text/template_id columns consumed by workload_signatures.py and wtl_v6.py."},{"m":"agent draft (gemini)","l":"log template mining","tx":"He et al. (ICWS 2017, Drain) introduce an online log parser built on a fixed-depth parse tree: root -> Layer 1 (message token length) -> Layer 2..d (leading static tokens, routing digit-bearing tokens to a wildcard branch) -> leaf cluster list matched by token-wise similarity simSeq >= st and updated by replacing differing tokens with '<*>'."}]},{"k":"heckman1979selection","a":"econometrics","t":"Sample Selection Bias as a Specification Error","y":1979,"s":"peer_reviewed","e":4,"d":"10.2307/1912352","v":"Econometrica","a_":["Heckman"],"bl":"https://doi.org/10.2307/1912352","pc":2,"sm":[{"m":"agent draft (claude)","l":"limitations caveats","tx":"Sample selection bias is an omitted-variable problem: estimates from a self-selected sample are biased unless the selection mechanism is modelled."},{"m":"agent draft (claude)","l":"operational translation","tx":"PM seat (council 5): featadopt and uppath are both selected samples: the fleet visible through ticketed YBAs, and upgrades visible only when a bundle reached a ticket. Bystanders reduce but do not remove selection."}]},{"k":"heer2010crowdsourcing","a":"foundations-viz","t":"Crowdsourcing graphical perception","y":2010,"s":"peer_reviewed","e":12,"d":"10.1145/1753326.1753357","v":"Proceedings of the SIGCHI Conference on Human Factors in Computing Systems","a_":["Heer & Bostock"],"bl":"https://doi.org/10.1145/1753326.1753357","pc":1,"sm":[{"m":"agent draft (gemini)","l":"methodology","tx":"Replicates and extends Cleveland & McGill's graphical perception hierarchy on web displays, quantifying log-error across position, length, angle, area, and horizon-graph encodings."}]},{"k":"heinrich2018requirements","a":"foundations-se","t":"Requirements for Data Quality Metrics","y":2017,"s":"peer_reviewed","e":1,"d":"10.1145/3148238","v":"Journal of Data and Information Quality","a_":["Heinrich et al."],"bl":"https://doi.org/10.1145/3148238","pc":1,"sm":[{"m":"agent draft (claude)","l":"data quality","tx":"Proposes five requirements for DQ metrics (defined min/max, interval-scaled values, quality of configuration parameters, sound aggregation, economic efficiency) and evaluates existing metrics incl. currency metrics against them. expected cadence) and aggregate only with a defensible rule (weakest-link rather than averaging)."}]},{"k":"helske-2021","a":"foundations-viz","t":"Can Visualization Alleviate Dichotomous Thinking? Effects of Visual Representations on the Cliff Effect","y":2021,"s":"peer_reviewed","e":6,"d":"10.1109/TVCG.2021.3073466","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Helske et al."],"bl":"https://doi.org/10.1109/TVCG.2021.3073466","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Cliff effect around p=0.05 persists across CI, gradient and violin displays; gradient CIs reduce it somewhat. Significance-only colouring in a gallery should carry a redundant continuous cue."}]},{"k":"herbsleb2003distributed","a":"foundations-systems","t":"An empirical study of speed and communication in globally distributed software development","y":2003,"s":"peer_reviewed","e":1,"d":"10.1109/TSE.2003.1205177","v":"IEEE Transactions on Software Engineering","a_":["Herbsleb & Mockus"],"bl":"https://doi.org/10.1109/TSE.2003.1205177","pc":1,"sm":[{"m":"agent draft (claude)","l":"coordination","tx":"Modification-request data at Lucent: work items spanning sites took about 2.5x longer than single-site items, and the slowdown was explained largely by the larger number of people involved, not by size or complexity alone."}]},{"k":"hernan2004selection","a":"ai-operations","t":"A Structural Approach to Selection Bias","y":2004,"s":"peer_reviewed","e":5,"d":"10.1097/01.ede.0000135174.63482.43","v":"Epidemiology","a_":["Hernán et al."],"bl":"https://doi.org/10.1097/01.ede.0000135174.63482.43","pc":3,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Selection bias as conditioning on a common effect (collider). Bundles are uploaded because tickets are hard, and universes become visible because a ticket ships a bundle; first-observed universes (unknown source version) carry 32% adverse upgrade outcomes vs 13% for known-source paths, a collider pattern in the uppath card."}]},{"k":"hernán-2010","a":"foundations-stats","t":"The Hazards of Hazard Ratios","y":2010,"s":"peer_reviewed","e":3,"d":"10.1097/EDE.0b013e3181c1ea43","v":"Epidemiology","a_":["Hernán"],"bl":"https://doi.org/10.1097/EDE.0b013e3181c1ea43","pc":1,"sm":[{"m":"agent draft (claude)","l":"limitations caveats","tx":"Hazard ratios are period-averaged and subject to built-in selection, so they do not translate into 'how much faster'. Recommends standardised survival curves / risk differences. Used to justify reporting G-computed RMST and median differences, not only Cox HRs, for intake grade vs TTR."}]},{"k":"hernán-2016","a":"foundations-stats","t":"Using Big Data to Emulate a Target Trial When a Randomized Trial Is Not Available: Table 1.","y":2016,"s":"peer_reviewed","e":3,"d":"10.1093/aje/kwv254","v":"American Journal of Epidemiology","a_":["Hernán & Robins"],"bl":"https://doi.org/10.1093/aje/kwv254","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Target-trial emulation: specify eligibility, strategies, assignment, time zero, follow-up and outcome, then emulate; misaligned time zero yields immortal-time and selection bias. Used to diagnose relrate (series label from the quarter's last snapshot, possibly after the ticket) and the upgrade before/after card (upgrades triggered by tickets)."}]},{"k":"herschel2017provsurvey","a":"foundations-systems","t":"A survey on provenance: What for? What form? What from?","y":2017,"s":"peer_reviewed","e":3,"d":"10.1007/s00778-017-0486-1","v":"The VLDB Journal","a_":["Herschel et al."],"bl":"https://doi.org/10.1007/s00778-017-0486-1","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Survey classifying provenance by purpose, granularity (workflow vs tuple level) and capture mode. Workflow provenance (code version, parameters, input versions) reproduces aggregates only if inputs are immutable or versioned."}]},{"k":"hidalgo2020slo","a":"foundations-telemetry","t":"Implementing Service Level Objectives: A Practical Guide to SLIs, SLOs, and Error Budgets","y":2020,"s":"external","e":2,"v":"O'Reilly Media","a_":["Hidalgo"],"pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Practical SLO/error-budget design: SLIs at the user boundary, windowed and burn-rate alerting, separate cumulative budget from acute paging. For fsync stalls: GF1 = blocked-seconds budget with burn-rate alerts (catches chronic moderate stalls), GF2 = single-stall magnitude page (catches the rare long stall before threshold crossings)."}]},{"k":"hochstein2023csmhealth","a":"foundations-other","t":"Customer success management, customer health, and retention in B2B industries","y":2023,"s":"peer_reviewed","e":1,"d":"10.1016/j.ijresmar.2023.09.002","v":"International Journal of Research in Marketing","a_":["Hochstein et al."],"bl":"https://doi.org/10.1016/j.ijresmar.2023.09.002","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"B2B evidence that customer-success health scores from usage and service signals predict retention. Supports adding support-derived signals (repeat P1, stalls, CE escalations, EOL exposure) to account health once joined to ARR and renewal dates, which the gallery lacks."}]},{"k":"hoefler-2015","a":"ai-operations","t":"Scientific benchmarking of parallel computing systems","y":2015,"s":"peer_reviewed","e":4,"d":"10.1145/2807591.2807644","v":"Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis","a_":["Hoefler & Belli"],"bl":"https://doi.org/10.1145/2807591.2807644","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Twelve rules for reporting performance: report variability and CIs, use medians/nonparametric stats for skewed data, document the full environment, distinguish deterministic from noisy improvements. Basis for a shared report schema across profiler agents."}]},{"k":"hoff2015trust","a":"foundations-se","t":"Trust in Automation","y":2014,"s":"peer_reviewed","e":2,"d":"10.1177/0018720814547570","v":"Human Factors: The Journal of the Human Factors and Ergonomics Society","a_":["Hoff & Bashir"],"bl":"https://doi.org/10.1177/0018720814547570","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Systematic review (127 studies) splitting trust in automation into dispositional, situational and learned layers; learned trust is updated by observed performance, and design features (transparency, feedback) shape it."}]},{"k":"hoffswell2020responsive","a":"foundations-viz","t":"Techniques for Flexible Responsive Visualization Design","y":2020,"s":"peer_reviewed","e":2,"d":"10.1145/3313831.3376777","v":"Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems","a_":["Hoffswell et al."],"bl":"https://doi.org/10.1145/3313831.3376777","sm":[{"m":"vault note","l":"summary","tx":"gallery council 9 (dataviz/communication), 2026-10-28"}]},{"k":"hofman-2020","a":"foundations-viz","t":"How Visualizing Inferential Uncertainty Can Mislead Readers About Treatment Effects in Scientific Results","y":2020,"s":"peer_reviewed","e":5,"d":"10.1145/3313831.3376454","v":"Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems","a_":["Hofman et al."],"bl":"https://doi.org/10.1145/3313831.3376454","pc":1,"sm":[{"m":"agent draft (claude)","l":"empirical findings","tx":"Two experiments: readers shown 95% confidence intervals overestimate treatment effects and pay more than readers shown 95% prediction intervals; axis rescaling helps less than prediction intervals or animated HOPs, and inferential-uncertainty displays make readers underestimate outcome variability."}]},{"k":"holmstrom-1991","a":"ai-operations","t":"Multitask Principal–Agent Analyses: Incentive Contracts, Asset Ownership, and Job Design","y":1991,"s":"peer_reviewed","e":2,"d":"10.1093/jleo/7.special_issue.24","v":"The Journal of Law, Economics, and Organization","a_":["Holmstrom & Milgrom"],"bl":"https://doi.org/10.1093/jleo/7.special_issue.24","pc":1,"sm":[{"m":"agent draft (claude)","l":"gaming risk of per engineer workload metrics","tx":"Multitask principal-agent model. When an agent spreads effort across several tasks and only some tasks are measurable, strong incentives on the measured tasks pull effort away from the unmeasured ones. The optimal response is often weaker incentives, or a job design that groups tasks so that they are equally measurable."}]},{"k":"hooimeijer-2007","a":"foundations-telemetry","t":"Modeling bug report quality","y":2007,"s":"peer_reviewed","e":8,"d":"10.1145/1321631.1321639","v":"Proceedings of the twenty-second IEEE/ACM international conference on Automated software engineering","a_":["Hooimeijer & Weimer"],"bl":"https://doi.org/10.1145/1321631.1321639","pc":1,"sm":[{"m":"agent draft (claude-sonnet-5-5)","l":"complexity measurement without effort","tx":"Models bug report quality: predicts whether a report gets resolved quickly from report-intrinsic features (attachments, comment count at filing time, reporter reputation, severity). Many features available at triage time predict time, showing intrinsic report characteristics and effort-related outcomes are entangled."}]},{"k":"horspool1980practical","a":"foundations-systems","t":"Practical fast searching in strings","y":1980,"s":"peer_reviewed","e":2,"d":"10.1002/spe.4380100608","v":"Software: Practice and Experience","a_":["Horspool"],"bl":"https://doi.org/10.1002/spe.4380100608","pc":1,"sm":[{"m":"agent draft (gemini)","l":"systems performance","tx":"Introduces the Boyer-Moore-Horspool bad-character shift heuristic that powers CPython's C-level fastsearch (str.find and the 'in' operator)."}]},{"k":"hu2026cbstail","a":"foundations-systems","t":"Come Hell or Still Water: Alleviating Tail Latency in Cloud Block Store","y":2026,"s":"external","e":3,"v":"23rd USENIX Symposium on Networked Systems Design and Implementation (NSDI 26)","a_":["Hu et al."],"bl":"https://www.usenix.org/system/files/nsdi26-hu-chaolei.pdf","pc":1,"sm":[{"m":"agent draft (claude)","l":"empirical findings","tx":"Alibaba EBS production: tail spikes (>10 ms) are frequent (649 in ~100 clusters/week), P99999 reaches ~109 ms, caused by short bursts from a few major-client virtual disks that load-balancing spreads across all storage nodes; proxy event-loop scheduling dominates tail latency even underloaded."}]},{"k":"huang-2015","a":"foundations-telemetry","t":"ConfValley","y":2015,"s":"peer_reviewed","e":1,"d":"10.1145/2741948.2741963","v":"Proceedings of the Tenth European Conference on Computer Systems","a_":["Huang et al."],"bl":"https://doi.org/10.1145/2741948.2741963","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"ConfValley: declarative validation of cloud configuration against specs and inferred constraints. Supports encoding named-flag hazards as checks over YBA platform_universe snapshots rather than mining drift after incidents."}]},{"k":"huang2017gray","a":"ai-operations","t":"Gray Failure: The Achilles' Heel of Cloud-Scale Systems","y":2017,"s":"peer_reviewed","e":9,"d":"10.1145/3102980.3103005","v":"Proceedings of the 16th Workshop on Hot Topics in Operating Systems (HotOS)","a_":["Huang et al."],"bl":"https://doi.org/10.1145/3102980.3103005","pc":7,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Gray failure = differential observability: the system's failure detector sees health while clients see failure. Directly matches the fsync POC: leader heartbeat does not touch Log::Sync, so a stalled leader keeps its lease (DB-22970); peers and YBA see a healthy node."},{"m":"agent draft (claude)","l":"yugabytedb fsync grey failure","tx":"Defines gray failure as differential observability: the failure detector (heartbeats, probes) sees a healthy component while the application observes degraded or failed service."},{"m":"agent draft (claude)","l":"operational translation","tx":"fsync T5-T7: gray failure is defined by differential observability. Implication for the stall_episode fact table: record the node's own evidence (S1/S4 durations) and what peers observed (Fail-or-stepdown naming the node) as separate columns; their disagreement is the gray-failure signature."}]},{"k":"huang2018panorama","a":"ai-operations","t":"Capturing and Enhancing In Situ System Observability for Failure Detection","y":2018,"s":"peer_reviewed","e":3,"v":"13th USENIX Symposium on Operating Systems Design and Implementation (OSDI 18)","a_":["Huang et al."],"bl":"https://www.usenix.org/conference/osdi18/presentation/huang","pc":6,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"fsync T5-T7: Panorama turns interacting components into in-situ observers whose timeouts/errors are evidence about the observed component, reported with its identity. Peers' 'Fail or stepdown of leader <uuid> detected' lines are such observer reports and must be joined to the stalled node by its permanent uuid, not by time alone."},{"m":"agent draft (claude)","l":"limitations caveats","tx":"Panorama turns in-process observations into failure-detection evidence, catching gray failures that heartbeats miss. Fleet-scale corollary: the useful gray-failure signal (fsync stall, slow node vs peers) lives in INFO-level glog, which only YB's long-retention glog plane carries; short-retention planes cannot support it."},{"m":"agent draft (claude)","l":"operational translation","tx":"Panorama turns in-situ observations made by a component's peers (timeouts and errors when calling it) into a failure detector for gray failure. For YB: peers' 'Couldn't send request ... Timed out' and denied pre-votes already mark a stalled node to about 1 s; a peer-observation detector is cheaper than host telemetry."}]},{"k":"huang2020tidb","a":"distributed-systems","t":"TiDB: A Raft-based HTAP Database","y":2020,"s":"peer_reviewed","e":5,"d":"10.14778/3415478.3415535","v":"Proceedings of the VLDB Endowment","a_":["Huang et al."],"bl":"https://doi.org/10.14778/3415478.3415535","pc":8,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"TiDB's Placement Driver balances region replicas, leaders and hot regions from store heartbeats as separate objectives. Supports treating YB leader skew and byte skew as two distinct per-universe metrics; hot-tablet detection needs load (ops/s), which tablet reports do not carry."},{"m":"agent draft (gemini)","l":"operational translation","tx":"Presents TiDB/TiKV's multi-Raft LSM architecture, detailing how Raft WAL batching, learner replication, and MVCC key encoding impact storage and network amplification."}]},{"k":"hubert1985ari","a":"foundations-stats","t":"Comparing partitions","y":1985,"s":"peer_reviewed","e":1,"d":"10.1007/BF01908075","v":"Journal of Classification","a_":["Hubert & Arabie"],"bl":"https://doi.org/10.1007/BF01908075"},{"k":"hughes-2016","a":"foundations-se","t":"Experiences with QuickCheck: Testing the Hard Stuff and Staying Sane","y":2016,"s":"peer_reviewed","e":1,"d":"10.1007/978-3-319-30936-1_9","v":"Lecture Notes in Computer Science","a_":["Hughes"],"bl":"https://doi.org/10.1007/978-3-319-30936-1_9","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Property-based testing in industry: generators plus invariants instead of examples, with shrinking to minimal failures. Applied to a figure gallery: generate perturbed card JSON (deleted keys, nulls, empty arrays, out-of-range numbers) and assert page-level properties (all cards present, failing card fails closed with a message, no NaN/undefined in text)."}]},{"k":"hullman-2011","a":"foundations-viz","t":"Visualization Rhetoric: Framing Effects in Narrative Visualization","y":2011,"s":"peer_reviewed","e":5,"d":"10.1109/TVCG.2011.255","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Hullman & Diakopoulos"],"bl":"https://doi.org/10.1109/TVCG.2011.255","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Visualization rhetoric: framing choices (what is coloured, titled, omitted) steer interpretation. A page that colours significance on some cards and categories on others creates conflicting rhetorical codes; one colour contract is needed."}]},{"k":"hullman-2020","a":"foundations-viz","t":"Why Authors Don't Visualize Uncertainty","y":2020,"s":"peer_reviewed","e":6,"d":"10.1109/TVCG.2019.2934287","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Hullman"],"bl":"https://doi.org/10.1109/TVCG.2019.2934287"},{"k":"hullman2015hops","a":"foundations-viz","t":"Hypothetical Outcome Plots Outperform Error Bars and Violin Plots for Inferences about Reliability of Variable Ordering","y":2015,"s":"peer_reviewed","e":3,"d":"10.1371/journal.pone.0142444","v":"PLOS ONE","a_":["Hullman et al."],"bl":"https://doi.org/10.1371/journal.pone.0142444","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"HOPs outperform error bars for judging how reliable an ordering is. Relevant to ranking cards (fs_infra, sigrel) where readers will rank groups whose intervals overlap heavily."}]},{"k":"hullman2019pursuiterror","a":"foundations-viz","t":"In Pursuit of Error: A Survey of Uncertainty Visualization Evaluation","y":2019,"s":"peer_reviewed","e":6,"d":"10.1109/TVCG.2018.2864889","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Hullman et al."],"bl":"https://doi.org/10.1109/TVCG.2018.2864889","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Survey of uncertainty-vis evaluation: readers misjudge intervals; evaluations should test decisions, not only perception. The gallery draws Wilson CIs widely, but several rate cards (fs_bands, fs_leader) show point values only, and colour flags point estimates whose CIs span the reference."},{"m":"agent draft (claude)","l":"dashboard design review","tx":"Surveys 86 uncertainty-visualization evaluations. Omitting uncertainty leads viewers to treat point estimates as exact. Frequency framings and distributional displays support better judgments than bare intervals."}]},{"k":"huo2023semparser","a":"foundations-telemetry","t":"SemParser: A Semantic Parser for Log Analytics","y":2023,"s":"peer_reviewed","e":8,"d":"10.1109/ICSE48619.2023.00082","v":"2023 IEEE/ACM 45th International Conference on Software Engineering (ICSE)","a_":["Huo et al."],"bl":"https://doi.org/10.1109/ICSE48619.2023.00082","pc":1,"sm":[{"m":"agent draft (claude)","l":"log template mining","tx":"fsync T5: SemParser recovers parameter semantics (concept vs instance) that syntax-only parsers lose, improving downstream detection. Supports treating the S1 op string as a semantic concept, recovered from the slot downstream."}]},{"k":"hussey2007steppedwedge","a":"foundations-stats","t":"Design and analysis of stepped wedge cluster randomized trials","y":2007,"s":"peer_reviewed","e":1,"d":"10.1016/j.cct.2006.05.007","v":"Contemporary Clinical Trials","a_":["Hussey & Hughes"],"bl":"https://doi.org/10.1016/j.cct.2006.05.007","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Stepped-wedge cluster randomization: all clusters start in control and cross to treatment at randomized steps; analysis is a mixed model with fixed period effects and cluster random intercepts, so secular trends are separated from treatment."}]},{"k":"hutwagner-2003","a":"foundations-telemetry","t":"The bioterrorism preparedness and response Early Aberration Reporting System (EARS)","y":2003,"s":"peer_reviewed","e":2,"d":"10.1007/PL00022319","v":"Journal of Urban Health","a_":["Hutwagner et al."],"bl":"https://doi.org/10.1007/PL00022319","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"EARS: simple CUSUM-style aberration detection on syndromic counts, judged by timeliness and false-alarm rate. Analogy for support: treat a cross-customer signature (Jira key, log template, differential label) as a syndrome; measure detection lead time and false alarms against the date engineering first linked the bug."}]},{"k":"huyse-1999","a":"foundations-other","t":"“INTERMED”: a method to assess health service needs","y":1999,"s":"peer_reviewed","e":1,"d":"10.1016/s0163-8343(98)00057-7","v":"General Hospital Psychiatry","a_":["Huyse et al."],"bl":"https://doi.org/10.1016/s0163-8343(98)00057-7","pc":1,"sm":[{"m":"agent draft (claude-sonnet-5-5)","l":"complexity measurement without effort","tx":"INTERMED Part I (development and reliability; Crossref title omits the part number). Scores biopsychosocial and health-system dimensions on historical, current, and prognostic axes; reliability tested with double-scored cases (reported ~94% of ratings within 1 point)."}]},{"k":"ibrahim2016forecasting","a":"foundations-telemetry","t":"Modeling and forecasting call center arrivals: A literature survey and a case study","y":2016,"s":"peer_reviewed","e":6,"d":"10.1016/j.ijforecast.2015.11.012","v":"International Journal of Forecasting","a_":["Ibrahim et al."],"bl":"https://doi.org/10.1016/j.ijforecast.2015.11.012","pc":2,"sm":[{"m":"agent draft (claude)","l":"support ops review","tx":"Survey of arrival-forecasting methods (time series, Poisson mixtures, Bayesian, distributional forecasts) plus case study; stresses overdispersion and the need for distributional (not point) forecasts for staffing. A support suite with overdispersed inflow and event surges (BFCM) needs interval forecasts of pages/tickets by shift to size event staffing."},{"m":"agent draft (claude)","l":"vp support dashboard","tx":"Survey of call-arrival modelling and forecasting: arrivals are time-varying, over-dispersed relative to Poisson and correlated within and across days; reviews time-series, doubly-stochastic and mixed-effects approaches and intraday updating, with a case study."}]},{"k":"inoue2015simd","a":"foundations-systems","t":"SIMD- and cache-friendly algorithm for sorting an array of structures","y":2015,"s":"peer_reviewed","e":2,"d":"10.14778/2809974.2809988","v":"Proceedings of the VLDB Endowment","a_":["Inoue & Taura"],"bl":"https://doi.org/10.14778/2809974.2809988","pc":1,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Inoue & Taura (VLDB '15) prove that when sorting an Array of Structures (AoS) such as compile-time log parsing `Vec<Row>` (`sizeof(Row) ~ 300-500B`), the dominant bottleneck is NOT key comparison—it is the random-access cache/TLB thrashing of rearranging large records after sorting packed `(key, index)` pairs (`take_record_batch` in the log parser)."}]},{"k":"ioannidis2005why","a":"foundations-stats","t":"Why Most Published Research Findings Are False","y":2005,"s":"peer_reviewed","e":6,"d":"10.1371/journal.pmed.0020124","v":"PLoS Medicine","a_":["Ioannidis"],"bl":"https://doi.org/10.1371/journal.pmed.0020124","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Low prior odds, small effects, flexible designs and many tested relationships make many claimed findings false. Frames the gallery-wide multiplicity count (about 1,850 explicit tests, mostly in sigrel/sigjira under BH)."}]},{"k":"iqbal2007disruption","a":"foundations-se","t":"Disruption and recovery of computing tasks","y":2007,"s":"peer_reviewed","e":1,"d":"10.1145/1240624.1240730","v":"Proceedings of the SIGCHI Conference on Human Factors in Computing Systems","a_":["Iqbal & Horvitz"],"bl":"https://doi.org/10.1145/1240624.1240730","pc":1,"sm":[{"m":"agent draft (claude)","l":"effort estimation","tx":"Field logging of Microsoft information workers responding to email/IM alerts: after switching to answer an alert, people took on the order of 10 minutes (Mark et al. 2016 cite 'over nine minutes') to return to the suspended task, often drifting to other applications."}]},{"k":"isenberg2011dualscale","a":"foundations-viz","t":"A Study on Dual-Scale Data Charts","y":2011,"s":"peer_reviewed","e":2,"d":"10.1109/TVCG.2011.160","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Isenberg et al."],"bl":"https://doi.org/10.1109/TVCG.2011.160","pc":1,"sm":[{"m":"agent draft (claude)","l":"dashboard design review","tx":"Controlled study of dual-scale charts. Readers make more errors and spurious cross-series comparisons when two quantities share an area with different scales. A separate aligned strip for n is safer."}]},{"k":"iso-42001","a":"security-governance","t":"ISO/IEC 42001:2023 Information technology - Artificial intelligence - Management system","y":2023,"s":"external","e":7,"v":"International Organization for Standardization","a_":["ISO/IEC"],"pc":4,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Auditor's reading for AI-derived operational metrics: clause 9.1 requires determining what is measured, methods that ensure valid results, when, and who analyses them, with retained documented information; 7.5.3 requires control of documented information (identification, versioning, retention, access, change)."},{"m":"agent draft (gemini)","l":"operational translation","tx":"ISO/IEC 42001:2023 (published 2023-12-18) specifies the requirements for an Artificial Intelligence Management System (AIMS) across Clauses 4-10 (Annex SL harmonized structure) and 38 normative reference controls across 9 Annex A control objectives (A.2 Policies [3], A.3 Internal organization [2], A.4 Resources [5], A.5 Impact assessment [4], A.6 AI…"},{"m":"agent draft (gemini)","l":"security governance","tx":"ISO/IEC 42001:2023 (AIMS) Clause 6.1.2 (AI risk assessment), Annex A.6.2.6 (AI system data quality & provenance), and Annex A.8.4 (Information for interested parties) require verifiable controls over training/inference data sanitization, deterministic redaction lineage (`RULESET_VERSION` + cryptographic fingerprint), and explicit disclosure when…"}]},{"k":"iso-5259","a":"security-governance","t":"ISO/IEC 5259-1:2024 through 5259-5:2025 Artificial intelligence - Data quality for analytics and machine learning (ML) (Parts 1-5: Overview, Data Quality Measures, Management Requirements, Process Framework, and Data Quality Governance Framework)","y":2024,"s":"external","e":5,"v":"International Organization for Standardization","a_":["ISO/IEC"],"pc":4,"sm":[{"m":"agent draft (claude)","l":"data quality","tx":"ISO/IEC 5259 series extends ISO/IEC 25012-style DQ characteristics to analytics/ML data, adding measures and lifecycle/governance processes (provenance, representativeness, label quality)."},{"m":"agent draft (gemini)","l":"operational translation","tx":"ISO/IEC 5259-1..5:2024-2025 operationalizes ISO/IEC 42001 Annex A.7 (A.7.2-A.7.6) and EU AI Act Article 10: Part 1 (5259-1:2024, 2024-07) defines the data life cycle (DLC) framework and terminology;"},{"m":"agent draft (gemini)","l":"security governance","tx":"ISO/IEC 5259-1:2024..5259-5:2025 establishes data quality measures (completeness, consistency, accuracy, auditability, and provenance) for analytics and ML pipelines."}]},{"k":"jain2016sqlshare","a":"foundations-other","t":"SQLShare","y":2016,"s":"peer_reviewed","e":2,"d":"10.1145/2882903.2882957","v":"Proceedings of the 2016 International Conference on Management of Data","a_":["Jain et al."],"bl":"https://doi.org/10.1145/2882903.2882957"},{"k":"jalkanen-2009","a":"maritime","t":"A modelling system for the exhaust emissions of marine traffic and its application in the Baltic Sea area","y":2009,"s":"peer_reviewed","e":2,"d":"10.5194/acp-9-9209-2009","v":"Atmospheric Chemistry and Physics","a_":["Jalkanen et al."],"bl":"https://doi.org/10.5194/acp-9-9209-2009","pc":1,"sm":[{"m":"agent draft (gemini)","l":"maritime emissions telemetry","tx":"Jalkanen et al. (2009) STEAM: A modelling system for the exhaust emissions of marine traffic using AIS telemetry"}]},{"k":"jalkanen-2012","a":"maritime","t":"Extension of an assessment model of ship traffic exhaust emissions for particulate matter and carbon monoxide","y":2012,"s":"peer_reviewed","e":2,"d":"10.5194/acp-12-2641-2012","v":"Atmospheric Chemistry and Physics","a_":["Jalkanen et al."],"bl":"https://doi.org/10.5194/acp-12-2641-2012","pc":1,"sm":[{"m":"agent draft (gemini)","l":"maritime emissions telemetry","tx":"Jalkanen et al. (2012) Extension of an assessment model of ship traffic exhaust emissions for particulate matter and carbon monoxide"}]},{"k":"jeong2009bugtossing","a":"foundations-telemetry","t":"Improving bug triage with bug tossing graphs","y":2009,"s":"peer_reviewed","e":7,"d":"10.1145/1595696.1595715","v":"Proceedings of the 7th joint meeting of the European software engineering conference and the ACM SIGSOFT symposium on The foundations of software engineering","a_":["Jeong et al."],"bl":"https://doi.org/10.1145/1595696.1595715","pc":1,"sm":[{"m":"agent draft (claude)","l":"coordination","tx":"Mozilla and Eclipse bug reports: a large fraction of bugs (reported as roughly 37-44%) are 'tossed' to another developer at least once, and tossing lengthens time-to-fix. Analogue for support escalations: re-routing among engineers is common and each reassignment adds latency and re-reading effort; tossing counts are a usable proxy for coordination load."}]},{"k":"jeyakumar2019explainit","a":"foundations-telemetry","t":"ExplainIt! -- A Declarative Root-cause Analysis Engine for Time Series Data","y":2019,"s":"peer_reviewed","e":3,"d":"10.1145/3299869.3314048","v":"Proceedings of the 2019 International Conference on Management of Data","a_":["Jeyakumar et al."],"bl":"https://doi.org/10.1145/3299869.3314048","pc":1,"sm":[{"m":"agent draft (gemini)","l":"declarative rule engine design","tx":"ExplainIt! (SIGMOD 2019) introduces a declarative SQL-like DSL for specifying root-cause hypotheses and causal dependency graphs over massive operational time-series telemetry."}]},{"k":"ji2026selective","a":"foundations-telemetry","t":"Selective encryption of sensitive fields in structured logs: Quantifying security–utility–performance tradeoffs","y":2026,"s":"peer_reviewed","e":2,"d":"10.1016/j.jisa.2026.104550","v":"Journal of Information Security and Applications","a_":["Ji et al."],"bl":"https://doi.org/10.1016/j.jisa.2026.104550","pc":2,"sm":[{"m":"agent draft (gemini)","l":"format driven log redaction","tx":"Demonstrates that parsing structured logs and selectively protecting only sensitive fields (rather than scanning or encrypting entire log records) improves ingestion throughput by 22.4%–54.0% and reduces query latency by up to 42.3%."}]},{"k":"jia2011mutation","a":"foundations-se","t":"An Analysis and Survey of the Development of Mutation Testing","y":2011,"s":"peer_reviewed","e":4,"d":"10.1109/TSE.2010.62","v":"IEEE Transactions on Software Engineering","a_":["Jia & Harman"],"bl":"https://doi.org/10.1109/TSE.2010.62","pc":3,"sm":[{"m":"agent draft (gemini)","l":"methodology","tx":"Comprehensive survey of mutation testing (IEEE TSE 2011), establishing fault-injection mutation kill rates as the rigorous empirical measure of whether a test or conformance suite actually detects semantic defects injected into production code."},{"m":"agent draft (gemini)","l":"mutation testing ci governance","tx":"Comprehensive survey of mutation testing foundations (Competent Programmer Hypothesis and Coupling Effect), mutant generation/reduction, and equivalent mutant filtering."}]},{"k":"jia2024aicreativity","a":"foundations-other","t":"When and How Artificial Intelligence Augments Employee Creativity","y":2024,"s":"peer_reviewed","e":1,"d":"10.5465/amj.2022.0426","v":"Academy of Management Journal","a_":["Jia et al."],"bl":"https://doi.org/10.5465/amj.2022.0426","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Field experiment (telemarketing): AI taking the routine first stage raised creativity and sales, mostly for higher-skilled agents. Operational translation: the bundle note is first-stage triage for skilled engineers; benefit, if any, should show as engineer time-to-hypothesis, not volume, and may concentrate in senior staff."}]},{"k":"jiahao-2023-arxiv","a":"security-governance","t":"GPTFUZZER: Red Teaming Large Language Models with Auto-Generated Jailbreak Prompts","y":2023,"s":"preprint","e":1,"x":"2309.10253","v":"arXiv:2309.10253","a_":["Yu et al."],"bl":"https://arxiv.org/abs/2309.10253","sm":[{"m":"vault note","l":"summary","tx":"GPTFUZZER: automated red-team prompt fuzzing; method ancestor of ybc chaos-sheet adversarial corpus"}]},{"k":"jiang2024lilac","a":"ai-operations","t":"LILAC: Log Parsing using LLMs with Adaptive Parsing Cache","y":2024,"s":"peer_reviewed","e":4,"d":"10.1145/3643733","v":"Proceedings of the ACM on Software Engineering","a_":["Jiang et al."],"bl":"https://doi.org/10.1145/3643733","pc":5,"sm":[{"m":"agent draft (claude)","l":"log template mining","tx":"fsync T5: LLM parsing with an adaptive cache for template consistency; not needed for five source-known stall templates. Contrast to catalog-based parsing."},{"m":"agent draft (gemini)","l":"log template mining","tx":"Jiang et al. (FSE 2024, LILAC) bridge the throughput gap of LLM log parsers by combining an ICL-prompted LLM extractor with an online Adaptive Parsing Cache (a prefix-tree template cache that matches incoming log lines in O(|tokens|) and only queries the LLM on cache misses, refining templates when multiple candidates collide)."},{"m":"agent draft (gemini)","l":"operational translation","tx":"LILAC combines LLM template extraction with an adaptive parsing cache to reach high F1 accuracy at Drain-like speeds, which is well-suited for offline the triage system/wtl2 silver-layer ingestion of ybc bundles rather than on-node collection."}]},{"k":"jiang2024loghub2","a":"ai-operations","t":"A Large-Scale Evaluation for Log Parsing Techniques: How Far Are We?","y":2024,"s":"peer_reviewed","e":5,"d":"10.1145/3650212.3652123","v":"Proceedings of the 33rd ACM SIGSOFT International Symposium on Software Testing and Analysis","a_":["Jiang et al."],"bl":"https://doi.org/10.1145/3650212.3652123","pc":7,"sm":[{"m":"agent draft (gemini)","l":"log template mining","tx":"Jiang et al. (ISSTA 2024, Loghub-2.0) evaluate 15 log parsers across 14 systems totaling 50.4M annotated log lines and 3,488 ground-truth templates using GA, PA, FGA (Fine-grained Grouping Accuracy), and FTA (Fine-grained Template Accuracy), revealing three watershed findings: (1) when scaling from 2k to 50.4M lines, 9 of 15 parsers fail to finish…"},{"m":"agent draft (gemini)","l":"parsing benchmark comparison","tx":"Releases Loghub-2.0 (Zenodo 8275861, 14 systems, 50.4M annotated lines, 3,488 ground-truth templates) and re-evaluates 15 log parsers. Proves that 2k-line sample benchmarks severely overstate parser accuracy: on full 3.6M-line datasets, 9 of 15 parsers fail to finish within 12h, statistical parsers drop below 0.50 PA, and FGA collapses below 0.03 on…"},{"m":"agent draft (gemini)","l":"limitations caveats","tx":"Loghub-2.0 demonstrates that heuristic and ML log parsers perform poorly on infrequent and parameter-intensive lines."}]},{"k":"jimenez2024swebench","a":"ai-operations","t":"SWE-bench: Can Language Models Resolve Real-World GitHub Issues?","y":2024,"s":"peer_reviewed","e":1,"v":"Proceedings of the Twelfth International Conference on Learning Representations (ICLR 2024)","a_":["Jimenez et al."],"bl":"https://openreview.net/forum?id=VTF8yNQM66"},{"k":"joel-2025-arxiv","a":"foundations-se","t":"Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity","y":2025,"s":"preprint","e":2,"x":"2507.09089","v":"arXiv:2507.09089","a_":["Becker et al."],"bl":"https://arxiv.org/abs/2507.09089","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"METR RCT: experienced open-source developers were 19% slower with early-2025 AI tools while believing they were faster. Operational translation: self-reported time saved by the AI bundle note is not evidence; require measured touch time or a blinded comparison."},{"m":"agent draft (claude)","l":"support human ai","tx":"METR RCT: 16 experienced open-source developers, 246 real issues randomized to AI-allowed or AI-disallowed. AI use made them ~19% slower, yet they believed AI made them ~20% faster, before and after the study. Support implication: experts on familiar, complex systems may not gain speed, and self-reported time savings are unreliable."}]},{"k":"johanssen2019feedback","a":"foundations-se","t":"How do Practitioners Capture and Utilize User Feedback During Continuous Software Engineering?","y":2019,"s":"peer_reviewed","e":2,"d":"10.1109/RE.2019.00026","v":"2019 IEEE 27th International Requirements Engineering Conference (RE)","a_":["Johanssen et al."],"bl":"https://doi.org/10.1109/RE.2019.00026","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"PM seat (council 5): Practitioners capture feedback from many channels but rarely link it systematically to requirements or usage data. Motivates linking tickets to IDEA keys and to fleet feature usage, which the gallery does not do yet."}]},{"k":"johansson-2017","a":"maritime","t":"Global assessment of shipping emissions in 2015 on a high spatial and temporal resolution","y":2017,"s":"peer_reviewed","e":4,"d":"10.1016/j.atmosenv.2017.08.042","v":"Atmospheric Environment","a_":["Johansson et al."],"bl":"https://doi.org/10.1016/j.atmosenv.2017.08.042","pc":1,"sm":[{"m":"agent draft (gemini)","l":"maritime emissions telemetry","tx":"Johansson, Jalkanen & Kukkonen (2017) Global assessment of shipping emissions in 2015 on a high spatial and temporal resolution"}]},{"k":"jordan2016souffle","a":"foundations-se","t":"Soufflé: On Synthesis of Program Analyzers","y":2016,"s":"peer_reviewed","e":1,"d":"10.1007/978-3-319-41540-6_23","v":"Lecture Notes in Computer Science","a_":["Jordan et al."],"bl":"https://doi.org/10.1007/978-3-319-41540-6_23","pc":1,"sm":[{"m":"agent draft (gemini)","l":"declarative rule engine design","tx":"Soufflé (CAV 2016) demonstrates how declarative Datalog rule specifications can be both interpreted at runtime for rapid rule prototyping and ahead-of-time compiled/embedded into high-performance native C++ binaries via staged specialization."}]},{"k":"joyner-2022","a":"foundations-viz","t":"Visualization Accessibility in the Wild: Challenges Faced by Visualization Designers","y":2022,"s":"peer_reviewed","e":3,"d":"10.1145/3491102.3517630","v":"CHI Conference on Human Factors in Computing Systems","a_":["Joyner et al."],"bl":"https://doi.org/10.1145/3491102.3517630","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Interviews with visualization designers: accessibility is skipped under time pressure and for lack of tooling. Operationally: put a11y checks (reflow at 320 px, focus-not-obscured, colour-only encodings) into an automated pre-publish script rather than relying on review rounds."}]},{"k":"jung-2022","a":"foundations-viz","t":"Communicating Visualizations without Visuals: Investigation of Visualization Alternative Text for People with Visual Impairments","y":2022,"s":"peer_reviewed","e":5,"d":"10.1109/TVCG.2021.3114846","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Jung et al."],"bl":"https://doi.org/10.1109/TVCG.2021.3114846"},{"k":"jung2021safe","a":"ai-operations","t":"Safe systems programming in Rust","y":2021,"s":"peer_reviewed","e":1,"d":"10.1145/3418295","v":"Communications of the ACM","a_":["Jung et al."],"bl":"https://doi.org/10.1145/3418295","pc":2,"sm":[{"m":"agent draft (gemini)","l":"architectural synthesis","tx":"Communications of the ACM (2021) synthesis of how Rust resolves the longstanding systems programming trade-off between safety (historically requiring GC) and low-level resource control (zero-cost abstractions, deterministic drop, no runtime GC pauses)."}]},{"k":"just2014mutants","a":"foundations-telemetry","t":"Are mutants a valid substitute for real faults in software testing?","y":2014,"s":"peer_reviewed","e":1,"d":"10.1145/2635868.2635929","v":"Proceedings of the 22nd ACM SIGSOFT International Symposium on Foundations of Software Engineering","a_":["Just et al."],"bl":"https://doi.org/10.1145/2635868.2635929","pc":1,"sm":[{"m":"agent draft (gemini)","l":"empirical mutation adequacy","tx":"Empirically validates across 357 real faults in 5 open-source systems (321k LOC) that mutant detection is statistically correlated with real fault detection independently of code coverage, justifying cargo-mutants as a mandatory CI adequacy gate for Yugabyte diagnostic rulepacks."}]},{"k":"k8s2026probes","a":"foundations-systems","t":"Configure Liveness, Readiness and Startup Probes","y":2026,"s":"external","e":2,"a_":["Kubernetes authors"],"bl":"https://kubernetes.io/docs/tasks/configure-pod-container/configure-liveness-readiness-startup-probes/","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Probe semantics: timeoutSeconds default 1 s; failureThreshold x periodSeconds bounds detection; one success resets the counter. Explains why the YB helm liveness probe misses grey disk stalls (Comply tservers stalled 34-97 s never restarted) and over-fires on throttled volumes (fsync ticket T8)."}]},{"k":"kale-2019","a":"foundations-viz","t":"Hypothetical Outcome Plots Help Untrained Observers Judge Trends in Ambiguous Data","y":2019,"s":"peer_reviewed","e":10,"d":"10.1109/TVCG.2018.2864909","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Kale et al."],"bl":"https://doi.org/10.1109/TVCG.2018.2864909","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"HOPs help untrained readers judge trend reliability."}]},{"k":"kale-2021","a":"foundations-viz","t":"Visual Reasoning Strategies for Effect Size Judgments and Decisions","y":2021,"s":"peer_reviewed","e":4,"d":"10.1109/TVCG.2020.3030335","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Kale et al."],"bl":"https://doi.org/10.1109/TVCG.2020.3030335"},{"k":"kalibera2013rigorous","a":"foundations-systems","t":"Rigorous benchmarking in reasonable time","y":2013,"s":"peer_reviewed","e":6,"d":"10.1145/2464157.2464160","v":"Proceedings of the 2013 international symposium on memory management","a_":["Kalibera & Jones"],"bl":"https://doi.org/10.1145/2464157.2464160","pc":4,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Multi-level repetition design (iterations, processes, builds) with CIs on effect size, optimizing repetition allocation. Replaces ad hoc '3-5 runs' with multi-process repetition and a CI on the ratio of means."},{"m":"agent draft (claude)","l":"operational translation","tx":"Choose repetition counts per level (process invocations, iterations) from pilot variance; report speedups as ratios with confidence intervals. For profiler agents: replace fixed '3-5 runs' with variance-driven repetition and a CI on the ratio."},{"m":"agent draft (gemini)","l":"benchmark methodology","tx":"Formalizes hierarchical multi-run benchmarking, non-parametric/bootstrap confidence intervals, and uncertainty quantification for systems performance ratios in reasonable execution time."}]},{"k":"kamra2008anomalousaccess","a":"foundations-telemetry","t":"Detecting anomalous access patterns in relational databases","y":2007,"s":"peer_reviewed","e":2,"d":"10.1007/s00778-007-0051-4","v":"The VLDB Journal","a_":["Kamra et al."],"bl":"https://doi.org/10.1007/s00778-007-0051-4","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Builds per-role profiles of SQL access from query logs (command, relations, attributes) and flags deviations with naive Bayes. A direct template for per-universe workload fingerprints from audit/pgaudit statement classes, and for drift detection between bundles of one universe."}]},{"k":"kanev-2015","a":"ai-operations","t":"Profiling a warehouse-scale computer","y":2015,"s":"peer_reviewed","e":7,"d":"10.1145/2749469.2750392","v":"Proceedings of the 42nd Annual International Symposium on Computer Architecture","a_":["Kanev et al."],"bl":"https://doi.org/10.1145/2749469.2750392","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Fleet-wide profiling shows a datacenter tax (malloc, memmove, compression, serialization) spread thin across binaries. For Rust profiling: sum allocator and memcpy frames across the flat profile rather than only the top frame; allocator choice is a fleet-level NOTE."}]},{"k":"kanjee2023gpt4ddx","a":"foundations-other","t":"Accuracy of a Generative Artificial Intelligence Model in a Complex Diagnostic Challenge","y":2023,"s":"peer_reviewed","e":1,"d":"10.1001/jama.2023.8288","v":"JAMA","a_":["Kanjee et al."],"bl":"https://doi.org/10.1001/jama.2023.8288","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"GPT-4 on 70 NEJM CPC cases: correct diagnosis anywhere in the differential 64%, top-ranked 39%, mean Bond quality score 4.2; inter-rater kappa 0.57. Operational lesson for LLM triage audits: report top-1 and any-of-N separately, use an ordinal partial-credit scale, and expect only moderate rater agreement on partial credit, so adjudicate."}]},{"k":"kaplan-1958","a":"foundations-stats","t":"Nonparametric Estimation from Incomplete Observations","y":1958,"s":"peer_reviewed","e":8,"d":"10.1080/01621459.1958.10501452","v":"Journal of the American Statistical Association","a_":["Kaplan & Meier"],"bl":"https://doi.org/10.1080/01621459.1958.10501452","pc":3,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Product-limit estimator for censored time-to-event data; use for time from a universe's adoption of a release to its first ticket or crash, treating the 320 upgrade units excluded as too recent as censored instead of dropped."},{"m":"agent draft (claude)","l":"metric statistics review","tx":"Product-limit estimator of the survival function from right-censored data. For time-to-resolve, open tickets are right-censored at current age; computing p50/p90 on solved tickets only biases recent cohorts low (measured: Q3-2026 cohort p90 33.1 d solved-only, with 32 of 731 still open; treating open as long-tail puts p90 near 38.8 d)."},{"m":"agent draft (claude)","l":"support ops dashboard","tx":"Product-limit estimator for survival curves under right-censoring. Applied to support tickets: open tickets are right-censored resolution times; computing TTR only over solved tickets (survivorship bias) understates TTR in any cohort window."}]},{"k":"kapoor2023leakage","a":"foundations-se","t":"Leakage and the reproducibility crisis in machine-learning-based science","y":2023,"s":"peer_reviewed","e":2,"d":"10.1016/j.patter.2023.100804","v":"Patterns","a_":["Kapoor & Narayanan"],"bl":"https://doi.org/10.1016/j.patter.2023.100804","pc":1,"sm":[{"m":"agent draft (claude)","l":"limitations caveats","tx":"A taxonomy of eight leakage types in ML-based science, including temporal leakage (features measured after the outcome) and non-independence between train and test. It proposes model info sheets."}]},{"k":"kay-2016","a":"foundations-se","t":"When (ish) is My Bus?","y":2016,"s":"peer_reviewed","e":3,"d":"10.1145/2858036.2858558","v":"Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems","a_":["Kay et al."],"bl":"https://doi.org/10.1145/2858036.2858558"},{"k":"kcs2016v6","a":"foundations-other","t":"KCS v6 Practices Guide","y":2016,"s":"external","e":3,"a_":["Consortium for Service Innovation"],"bl":"https://library.serviceinnovation.org/KCS/KCS_v6/KCS_v6_Practices_Guide","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"KCS Evolve Loop: aggregate reuse and new-vs-known analysis feeds root cause and product improvement back to engineering. Bundle signatures + KB coverage + Jira links are raw material for a ranked known-issue-by-org-reach feed to PM."}]},{"k":"keiichi-2016-arxiv","a":"foundations-telemetry","t":"Length Matters: Clustering System Log Messages using Length of Words","y":2016,"s":"preprint","e":5,"x":"1611.03213","v":"arXiv:1611.03213","a_":["Shima"],"bl":"https://arxiv.org/abs/1611.03213","pc":1,"sm":[{"m":"agent draft (gemini)","l":"log template mining","tx":"Shima (arXiv:1611.03213, 2016, LenMa) proposes clustering system log messages by representing each whitespace-tokenized message as a vector of word character-lengths and computing cosine similarity against existing clusters with the same token count."}]},{"k":"keiningham2007nps","a":"foundations-other","t":"A Longitudinal Examination of Net Promoter and Firm Revenue Growth","y":2007,"s":"peer_reviewed","e":2,"d":"10.1509/jmkg.71.3.039","v":"Journal of Marketing","a_":["Keiningham et al."],"bl":"https://doi.org/10.1509/jmkg.71.3.039","pc":2,"sm":[{"m":"agent draft (claude)","l":"support ops review","tx":"Longitudinal test finds no evidence that NPS is a superior predictor of revenue growth versus other satisfaction measures, contradicting Reichheld's 'one number' claim. Lesson for support leadership: do not replace a weak CSAT signal with a single NPS figure;"},{"m":"agent draft (claude)","l":"vp support dashboard","tx":"Longitudinal test of Reichheld's claim that Net Promoter is the single best predictor of firm growth; using Norwegian customer data across industries, the authors could not replicate NPS superiority over the ACSI satisfaction index. For a support dashboard: do not use NPS as the headline customer-outcome metric;"}]},{"k":"kerbiriou2019pugz","a":"foundations-systems","t":"Parallel Decompression of Gzip-Compressed Files and Random Access to DNA Sequences","y":2019,"s":"peer_reviewed","e":2,"d":"10.1109/IPDPSW.2019.00042","v":"2019 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW)","a_":["Kerbiriou & Chikhi"],"bl":"https://doi.org/10.1109/IPDPSW.2019.00042","pc":1,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"pugz (Kerbiriou & Chikhi, IPDPSW '19) breaks the single-core serial ceiling of single-member gzip decompression (`MultiGzDecoder`) via a two-stage parallel algorithm: Stage 1 partitions the compressed bit-stream into chunks, locates a valid Deflate dynamic Huffman block header in each chunk via bit-level trial decoding, and inflates symbols while…"}]},{"k":"kerr1975folly","a":"foundations-other","t":"On the Folly of Rewarding A, While Hoping for B.","y":1975,"s":"peer_reviewed","e":3,"d":"10.5465/255378","v":"Academy of Management Journal","a_":["Kerr"],"bl":"https://doi.org/10.5465/255378","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Reward systems pay for the measurable proxy and get it, not the goal. In support ops: a bundle-coverage target (activity) rewards asking every customer for a bundle, raising customer effort and time-to-relief, while the goal is faster correct diagnosis. KCS v6 states the same rule: no goals on activity metrics (link rate, coverage);"},{"m":"agent draft (claude)","l":"vp support dashboard","tx":"Classic management essay: organizations routinely reward measurable behaviour A while hoping for unmeasured outcome B, producing the opposite of what they want. Basis for pairing every support target (e.g., FRT, tickets closed) with a counter-metric (reopen rate, CSAT, escalation quality) and for not tying individual compensation to activity counts."}]},{"k":"kersten2006mylyn","a":"foundations-se","t":"Using task context to improve programmer productivity","y":2006,"s":"peer_reviewed","e":3,"d":"10.1145/1181775.1181777","v":"Proceedings of the 14th ACM SIGSOFT international symposium on Foundations of software engineering","a_":["Kersten & Murphy"],"bl":"https://doi.org/10.1145/1181775.1181777","pc":1,"sm":[{"m":"agent draft (claude)","l":"effort estimation","tx":"Mylyn captured per-task interaction context in the IDE, the ancestor of task-attributed activity timing (the approach WakaTime-style tools later popularised with heartbeat-plus-timeout heuristics). Relevant as precedent for attributing interaction time to a task id (here, a ticket) only while the task is active; it does not itself validate time estimates."}]},{"k":"khan2022guidelines","a":"foundations-telemetry","t":"Guidelines for assessing the accuracy of log message template identification techniques","y":2022,"s":"peer_reviewed","e":8,"d":"10.1145/3510003.3510101","v":"Proceedings of the 44th International Conference on Software Engineering","a_":["Khan et al."],"bl":"https://doi.org/10.1145/3510003.3510101","pc":2,"sm":[{"m":"agent draft (claude)","l":"log template mining","tx":"fsync T5: over-generalised templates (a constant token treated as variable) silently merge distinct events. WTL's 'Time spent <*>: real...' merges Fsync log, Append to log and Log roll; the op is recoverable from the slot value, so the minimal fix is downstream classification on the slot, avoiding a template-id change that would split silver vintages."},{"m":"agent draft (gemini)","l":"log template mining","tx":"Khan, Shin, Bianculli, and Briand (ICSE 2022) prove that Grouping Accuracy (GA) is fundamentally insufficient for evaluating log parsers because a parser can achieve GA = 1.0 while producing completely wrong template strings (e.g., failing to mask a constant-per-group parameter or over-masking static words), and formalize message-level Parsing Accuracy…"}]},{"k":"khatri2010designing","a":"foundations-other","t":"Designing data governance","y":2010,"s":"peer_reviewed","e":5,"d":"10.1145/1629175.1629210","v":"Communications of the ACM","a_":["Khatri & Brown"],"bl":"https://doi.org/10.1145/1629175.1629210","pc":1,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Adapts Weill & Ross's IT governance framework to information assets, distinguishing data governance (what decisions must be made and who holds the locus of accountability/decision rights) from data management (making and implementing day-to-day decisions)."}]},{"k":"khomh2012fasterreleases","a":"foundations-se","t":"Do faster releases improve software quality? An empirical case study of Mozilla Firefox","y":2012,"s":"peer_reviewed","e":2,"d":"10.1109/MSR.2012.6224279","v":"2012 9th IEEE Working Conference on Mining Software Repositories (MSR)","a_":["Khomh et al."],"bl":"https://doi.org/10.1109/MSR.2012.6224279","pc":1,"sm":[{"m":"agent draft (claude)","l":"limitations caveats","tx":"Firefox rapid release: no more post-release bugs but bugs surface and are fixed faster, and adoption dynamics change what is observed in the field. Warns that release cadence and adoption timing confound field-defect comparisons between releases."}]},{"k":"khomh2015rapidreleases","a":"foundations-se","t":"Understanding the impact of rapid releases on software quality","y":2014,"s":"peer_reviewed","e":2,"d":"10.1007/s10664-014-9308-x","v":"Empirical Software Engineering","a_":["Khomh et al."],"bl":"https://doi.org/10.1007/s10664-014-9308-x","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Firefox rapid vs traditional releases: post-release bug counts similar, but bugs surface and get fixed faster; release comparisons must align on time since release. Supports day-0 alignment in relrate."}]},{"k":"killick2012optimal","a":"ai-operations","t":"Optimal Detection of Changepoints With a Linear Computational Cost","y":2012,"s":"external","e":2,"v":"Journal of the American Statistical Association (JASA)","a_":["Killick et al."],"pc":4,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"fsync T7: PELT considered for episode boundaries and rejected: stall lines are sparse events carrying their own duration, not a sampled series; interval union on [ts - real, ts] is exact and auditable, while changepoints need a sampled signal and a penalty."},{"m":"agent draft (claude)","l":"operational translation","tx":"PELT: exact penalized change-point detection in linear time."},{"m":"agent draft (gemini)","l":"systems performance","tx":"Provides the Pruned Exact Linear Time (PELT) optimal changepoint detection framework that grounds WTL v6's two-tier statistical temporal compression (burst_window_ms, effective_window_ms) and Doctor v2.5's temporal status classification in promdump_query.py (_classify_breach_pattern) and finding_context.py (OBSERVED_ACTIVE vs INTERMITTENT vs…"}]},{"k":"kim-2014","a":"foundations-stats","t":"Are Call Center and Hospital Arrivals Well Modeled by Nonhomogeneous Poisson Processes?","y":2014,"s":"peer_reviewed","e":4,"d":"10.1287/msom.2014.0490","v":"Manufacturing &amp; Service Operations Management","a_":["Kim & Whitt"],"bl":"https://doi.org/10.1287/msom.2014.0490","pc":1,"sm":[{"m":"agent draft (claude)","l":"support ops dashboard","tx":"KS-based tests show call-center and ED arrivals are consistent with NHPP only after accounting for rounding, intra-interval rate variation, and overdispersion from pooling hours across days."}]},{"k":"kim-2018","a":"foundations-viz","t":"Assessing Effects of Task and Data Distribution on the Effectiveness of Visual Encodings","y":2018,"s":"peer_reviewed","e":8,"d":"10.1111/cgf.13409","v":"Computer Graphics Forum","a_":["Kim & Heer"],"bl":"https://doi.org/10.1111/cgf.13409"},{"k":"kim-2024","a":"foundations-se","t":"Erie: A Declarative Grammar for Data Sonification","y":2024,"s":"peer_reviewed","e":2,"d":"10.1145/3613904.3642442","v":"Proceedings of the CHI Conference on Human Factors in Computing Systems","a_":["Kim et al."],"bl":"https://doi.org/10.1145/3613904.3642442"},{"k":"kim2006bugfixtime","a":"foundations-telemetry","t":"How long did it take to fix bugs?","y":2006,"s":"peer_reviewed","e":7,"d":"10.1145/1137983.1138027","v":"Proceedings of the 2006 international workshop on Mining software repositories","a_":["Kim & Whitehead"],"bl":"https://doi.org/10.1145/1137983.1138027","pc":1,"sm":[{"m":"agent draft (claude)","l":"vp support dashboard","tx":"Mines open-source repositories to measure how long bugs take to fix per file; fix times are long and highly skewed. Supports reporting customer-escalation (CE Jira) time-to-fix as a distribution/survival curve with open items censored, not as a mean."}]},{"k":"kim2011crashgraphs","a":"foundations-telemetry","t":"Crash graphs: An aggregated view of multiple crashes to improve crash triage","y":2011,"s":"peer_reviewed","e":5,"d":"10.1109/dsn.2011.5958261","v":"2011 IEEE/IFIP 41st International Conference on Dependable Systems &amp; Networks (DSN)","a_":["Kim et al."],"bl":"https://doi.org/10.1109/dsn.2011.5958261","pc":2,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Kim et al. (DSN 2011) introduce Crash Graphs, aggregating multiple stack traces within a crash bucket into a directed call-transition graph where nodes are functions and edges represent caller-callee adjacency across crashes."}]},{"k":"kim2011topcrashes","a":"foundations-telemetry","t":"Which Crashes Should I Fix First?: Predicting Top Crashes at an Early Stage to Prioritize Debugging Efforts","y":2011,"s":"peer_reviewed","e":3,"d":"10.1109/TSE.2011.20","v":"IEEE Transactions on Software Engineering","a_":["Kim et al."],"bl":"https://doi.org/10.1109/TSE.2011.20","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Top-crash prediction at Mozilla: few crash classes dominate. Council-5 DocDB lens: in YugabyteDB bundles a single field crash class (intents-DB 'Unexpected write id' Corruption, 2025.2.x before 2025.2.7) is a known-fixed cluster; the crash panel should rank by distinct events and versions-with-fix, so support can say 'upgrade' instead of 'unclassified'."}]},{"k":"kim2021accessiblevis","a":"foundations-viz","t":"Accessible Visualization: Design Space, Opportunities, and Challenges","y":2021,"s":"peer_reviewed","e":6,"d":"10.1111/cgf.14298","v":"Computer Graphics Forum","a_":["Kim et al."],"bl":"https://doi.org/10.1111/cgf.14298","pc":2,"sm":[{"m":"agent draft (claude)","l":"limitations caveats","tx":"Correction to my earlier perspective: the abstract says the authors analysed visualization-accessibility papers from the last ~20 years; I did not verify the exact count (56) I stated earlier, so treat that number as unverified. The design-space framing (users, tasks, media) and the gap in interactive/complex-chart access stand."},{"m":"agent draft (claude)","l":"methodology","tx":"Survey of 56 accessible-visualization papers organised into a design space by user group (mostly blind/low-vision), literacy, task, and medium (tactile, sonification, screen-reader text, interaction). Finds the field concentrated on BLV users and static description, with gaps in interactive navigation and in non-visual access to complex charts."}]},{"k":"kim2021captions","a":"foundations-viz","t":"Towards Understanding How Readers Integrate Charts and Captions: A Case Study with Line Charts","y":2021,"s":"peer_reviewed","e":3,"d":"10.1145/3411764.3445443","v":"Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems","a_":["Kim et al."],"bl":"https://doi.org/10.1145/3411764.3445443"},{"k":"kim2021responsivecomm","a":"foundations-viz","t":"Design Patterns and Trade‐Offs in Responsive Visualization for Communication","y":2021,"s":"peer_reviewed","e":2,"d":"10.1111/cgf.14321","v":"Computer Graphics Forum","a_":["Kim et al."],"bl":"https://doi.org/10.1111/cgf.14321","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Analysed 378 responsive visualization pairs from news sites and catalogued the transformation strategies designers use between desktop and mobile (rescale, reposition, transpose axes, simplify/aggregate, remove or move labels and annotations, change encoding)."}]},{"k":"kim2024imnotsure","a":"foundations-other","t":"\"I'm Not Sure, But...\": Examining the Impact of Large Language Models' Uncertainty Expression on User Reliance and Trust","y":2024,"s":"peer_reviewed","e":8,"d":"10.1145/3630106.3658941","v":"The 2024 ACM Conference on Fairness Accountability and Transparency","a_":["Kim et al."],"bl":"https://doi.org/10.1145/3630106.3658941","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Large preregistered experiment: first-person uncertainty phrases from an LLM ('I'm not sure, but...') reduced over-reliance and raised accuracy on hard questions, with a small trust cost; general-perspective phrasing had weaker effects."}]},{"k":"kim2025anarchy","a":"foundations-systems","t":"Anarchy in the Database: A Survey and Evaluation of Database Management System Extensibility","y":2025,"s":"peer_reviewed","e":3,"d":"10.14778/3725688.3725719","v":"Proceedings of the VLDB Endowment","a_":["Kim et al."],"bl":"https://doi.org/10.14778/3725688.3725719","pc":1,"sm":[{"m":"agent draft (claude)","l":"empirical findings","tx":"Taxonomy of DBMS extensibility across PostgreSQL, MySQL, MariaDB, SQLite, Redis, DuckDB plus static/dynamic analysis of 400+ PostgreSQL extensions: about 16.8% conflict with at least one other extension and can cause failures, correlated with extension complexity and hook usage."}]},{"k":"kleinberg2002bursty","a":"foundations-telemetry","t":"Bursty and hierarchical structure in streams","y":2002,"s":"peer_reviewed","e":1,"d":"10.1145/775047.775061","v":"Proceedings of the eighth ACM SIGKDD international conference on Knowledge discovery and data mining","a_":["Kleinberg"],"bl":"https://doi.org/10.1145/775047.775061","pc":2,"sm":[{"m":"agent draft (gemini)","l":"methodology","tx":"Formalizes burst detection in continuous event streams using an infinite-state automaton where burst states correspond to elevated emission rates and state transitions incur cost proportional to the state jump. Establishes the theoretical foundation for separating baseline arrival rates from multi-scale burst onset and peak intervals."}]},{"k":"knespel2023rapidgzip","a":"foundations-systems","t":"Rapidgzip: Parallel Decompression and Seeking in Gzip Files Using Cache Prefetching","y":2023,"s":"peer_reviewed","e":1,"d":"10.1145/3588195.3592992","v":"Proceedings of the 32nd International Symposium on High-Performance Parallel and Distributed Computing","a_":["Knespel & Brunst"],"bl":"https://doi.org/10.1145/3588195.3592992","pc":1,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"rapidgzip (Knespel & Brunst, HPDC '23) generalizes pugz's two-stage speculative parallel gzip decompression from ASCII-only text to arbitrary binary/UTF-8 gzip streams using a 16-bit LUT prefilter for Deflate block-header discovery, two-step marker resolution with cache prefetching, and on-the-fly seek-point indexing."}]},{"k":"knox2014complaints","a":"foundations-other","t":"Customer Complaints and Recovery Effectiveness: A Customer Base Approach","y":2014,"s":"peer_reviewed","e":2,"d":"10.1509/jm.12.0317","v":"Journal of Marketing","a_":["Knox & Oest"],"bl":"https://doi.org/10.1509/jm.12.0317","pc":1,"sm":[{"m":"agent draft (claude)","l":"vp support dashboard","tx":"Customer-base model linking complaints to subsequent defection and spend, and quantifying recovery effectiveness: complaints signal elevated churn risk, and effective recovery offsets much of it. Justifies an account-risk tab that tracks complaint/escalation episodes per account and whether recovery happened, rather than only aggregate satisfaction."}]},{"k":"kolesar-1998","a":"foundations-other","t":"INSIGHTS ON SERVICE SYSTEM DESIGN FROM A NORMAL APPROXIMATION TO ERLANG'S DELAY FORMULA","y":1998,"s":"peer_reviewed","e":4,"d":"10.1111/j.1937-5956.1998.tb00457.x","v":"Production and Operations Management","a_":["KOLESAR & GREEN"],"bl":"https://doi.org/10.1111/j.1937-5956.1998.tb00457.x","pc":1,"sm":[{"m":"agent draft (claude)","l":"support ops dashboard","tx":"Normal approximation to Erlang C yielding square-root staffing rules and insight that delay probability depends on safety capacity relative to sqrt(load). Useful for quick what-if staffing tables on a support dashboard, but approximation degrades at very low offered load (small on-call pools)."}]},{"k":"kong2018titles","a":"foundations-viz","t":"Frames and Slants in Titles of Visualizations on Controversial Topics","y":2018,"s":"peer_reviewed","e":9,"d":"10.1145/3173574.3174012","v":"Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems","a_":["Kong et al."],"bl":"https://doi.org/10.1145/3173574.3174012","pc":2,"sm":[{"m":"agent draft (claude)","l":"dashboard design review","tx":"Shows that chart titles strongly frame what viewers recall and conclude, often more than the data. Relevant to dashboard panel titles: a title that states the scope (\"P1 within target\") without the exclusion lets readers generalise to all urgent tickets. Titles should carry the question and the scope."},{"m":"agent draft (gemini)","l":"empirical findings","tx":"Demonstrates that action-oriented takeaway titles shape over 80% of viewer recall and main-message extraction when charts are viewed briefly. For executive LinkedIn figures, replacing passive metric labels ('Model Routing Comparison') with explicit empirical takeaway titles ('Why Two-Tier Model Routing Fails on Complex Work') ensures the core…"}]},{"k":"kong2019titlemisalignment","a":"foundations-viz","t":"Trust and Recall of Information across Varying Degrees of Title-Visualization Misalignment","y":2019,"s":"peer_reviewed","e":5,"d":"10.1145/3290605.3300576","v":"Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems","a_":["Kong et al."],"bl":"https://doi.org/10.1145/3290605.3300576"},{"k":"koole-2023","a":"foundations-telemetry","t":"A Practice-Oriented Overview of Call Center Workforce Planning","y":2023,"s":"peer_reviewed","e":7,"d":"10.1287/stsy.2021.0008","v":"Stochastic Systems","a_":["Koole & Li"],"bl":"https://doi.org/10.1287/stsy.2021.0008","pc":2,"sm":[{"m":"agent draft (claude)","l":"Support Operations and Workforce Management lead","tx":"Practice-oriented review of the WFM chain: forecast arrivals and handle time per interval, convert to required staff with a queueing model (Erlang C/A or simulation), then schedule and roster, with feedback between the steps. It is honest about which methods work in practice and which theory does not transfer."},{"m":"agent draft (claude)","l":"operational translation","tx":"Practice-oriented review of workforce planning: forecast (with uncertainty), staff via Erlang C/A or simulation, schedule, then intraday control. For Yugabyte support the translation is that a planner needs per-class arrival rates with their over-dispersion and per-class work content, not elapsed resolution time;"}]},{"k":"kosara-2013","a":"foundations-viz","t":"Storytelling: The Next Step for Visualization","y":2013,"s":"peer_reviewed","e":3,"d":"10.1109/MC.2013.36","v":"Computer","a_":["Kosara & Mackinlay"],"bl":"https://doi.org/10.1109/MC.2013.36","pc":1,"sm":[{"m":"agent draft (claude-opus-5.5)","l":"operational translation","tx":"Argues presentation (storytelling) is a distinct visualization goal from exploration: guided sequences, annotation and a stated message. Used in the bundle-insights gallery for takeaway-sentence titles and stakeholder-ordered sections."}]},{"k":"krippendorff-2004","a":"foundations-stats","t":"Reliability in Content Analysis.","y":2004,"s":"peer_reviewed","e":3,"d":"10.1111/j.1468-2958.2004.tb00738.x","v":"Human Communication Research","a_":["Krippendorff"],"bl":"https://doi.org/10.1111/j.1468-2958.2004.tb00738.x","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Reliability conditions for content analysis and alpha thresholds (>=.800 for conclusions; .667-.800 tentative only)."}]},{"k":"kuiper-2025","a":"ai-operations","t":"Saving Private Hash Join","y":2025,"s":"peer_reviewed","e":3,"d":"10.14778/3742728.3742762","v":"Proceedings of the VLDB Endowment","a_":["Kuiper et al."],"bl":"https://doi.org/10.14778/3742728.3742762","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Adaptive external hash join in DuckDB with runtime compression and unified memory mgmt across operators. Spill is graceful; measure peak temp dir size and temp_directory location; memory_limit and threads interact."}]},{"k":"kul2018sqlclustering","a":"distributed-systems","t":"Similarity Metrics for SQL Query Clustering","y":2018,"s":"peer_reviewed","e":2,"d":"10.1109/TKDE.2018.2831214","v":"IEEE Transactions on Knowledge and Data Engineering","a_":["Kul et al."],"bl":"https://doi.org/10.1109/TKDE.2018.2831214","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Compares similarity metrics for clustering SQL query logs and shows structural feature extraction yields usable workload clusters. Operational use for PM workload-shape questions from support bundles: cluster pgaudit/slow-query statement fingerprints (artifact.query_event;"}]},{"k":"kumar2012webcost","a":"foundations-telemetry","t":"Does the Web Reduce Customer Service Cost? Empirical Evidence from a Call Center","y":2012,"s":"peer_reviewed","e":3,"d":"10.1287/isre.1110.0390","v":"Information Systems Research","a_":["Kumar & Telang"],"bl":"https://doi.org/10.1287/isre.1110.0390","pc":1,"sm":[{"m":"agent draft (claude)","l":"contrasting view","tx":"Panel study of a firm's call center after web self-service launched. The web substitutes for simple, information-type calls but not for problem-type calls, and self-service can generate new contacts. For a B2B database vendor this means 'deflection' cannot be read off portal views; it must be measured per issue type as change in assisted contacts."}]},{"k":"kuschewski2023btrblocks","a":"ai-operations","t":"BtrBlocks: Efficient Columnar Compression for Data Lakes","y":2023,"s":"peer_reviewed","e":2,"d":"10.1145/3589263","v":"Proceedings of the ACM on Management of Data","a_":["Kuschewski et al."],"bl":"https://doi.org/10.1145/3589263","pc":3,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"BtrBlocks motivates cascading lightweight type-specific encodings before general compression; in ybc's Parquet writer on the coordinator, we apply the Parquet equivalents (dictionary on labels, DELTA_BINARY_PACKED on timestamps/offsets, BYTE_STREAM_SPLIT on metric doubles) ahead of ZSTD level 3."},{"m":"agent draft (claude)","l":"operational translation","tx":"BtrBlocks: cascaded lightweight encodings chosen by sampling, no heavyweight codec; argues decompression throughput, not ratio, is the bottleneck for cloud data-lake scans at high network bandwidth. For a profiler: on fast networks/NVMe, heavy codecs can be CPU-bound; measure decode CPU vs bytes saved."}]},{"k":"kwon-2023","a":"distributed-systems","t":"Realizing Strong Determinism Contract on Log-Structured Merge Key-Value Stores","y":2023,"s":"peer_reviewed","e":5,"d":"10.1145/3582695","v":"ACM Transactions on Storage","a_":["Kwon et al."],"bl":"https://doi.org/10.1145/3582695","pc":1,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"Vigil-KV co-designs LSM compaction/flush scheduling with NVMe Predictable Latency Mode (PLM) windows, cutting tail latency by 3.19x and average latency by 34% across Facebook and Yahoo workloads."}]},{"k":"laaber2019cloud","a":"foundations-systems","t":"Software microbenchmarking in the cloud. How bad is it really?","y":2019,"s":"peer_reviewed","e":5,"d":"10.1007/s10664-019-09681-1","v":"Empirical Software Engineering","a_":["Laaber et al."],"bl":"https://doi.org/10.1007/s10664-019-09681-1","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Cloud microbenchmark variability differs widely by benchmark/provider; randomized multiple interleaved trials on the same instance reduce false positives. For profiler agents on shared hosts: interleave A/B and report the minimum detectable effect."},{"m":"agent draft (gemini)","l":"cloud variance","tx":"Quantifies cloud VM performance variability across providers and instance types, showing that unpaired multi-VM comparisons suffer severe false positives. Directly corroborates WTL-P2's +40% to 2.1x 'platform lottery' finding on Cloud Batch."}]},{"k":"lamb2024datafusion","a":"foundations-systems","t":"Apache Arrow DataFusion: A Fast, Embeddable, Modular Analytic Query Engine","y":2024,"s":"peer_reviewed","e":8,"d":"10.1145/3626246.3653368","v":"Companion of the 2024 International Conference on Management of Data","a_":["Lamb et al."],"bl":"https://doi.org/10.1145/3626246.3653368","pc":2,"sm":[{"m":"agent draft (gemini)","l":"PyO3 + Arrow C Data Interface FFI Architecture","tx":"Apache Arrow DataFusion demonstrates how a modular Rust analytical engine built natively on the `arrow-rs` and `parquet` crates can serve as the shared execution core across language ecosystems (including `datafusion-python` via PyO3 and the Arrow C Data Interface / PyCapsule zero-copy protocol)."},{"m":"agent draft (gemini)","l":"columnar architecture","tx":"Demonstrates that a modular, Rust-native analytical query engine built directly on Apache Arrow in-memory RecordBatches and Apache Parquet (arrow-rs/parquet crates) achieves OLAP query performance competitive with tightly coupled C++ engines (DuckDB) while exposing zero-copy columnar extension APIs."}]},{"k":"lamkanfi-2010","a":"foundations-telemetry","t":"Predicting the severity of a reported bug","y":2010,"s":"peer_reviewed","e":4,"d":"10.1109/MSR.2010.5463284","v":"2010 7th IEEE Working Conference on Mining Software Repositories (MSR 2010)","a_":["Lamkanfi et al."],"bl":"https://doi.org/10.1109/MSR.2010.5463284","pc":1,"sm":[{"m":"agent draft (claude-sonnet-5-5)","l":"ex ante prediction","tx":"Text-mining classifier predicts bug severity (severe vs non-severe) from report summaries on Mozilla, Eclipse and GNOME, precision and recall about 0.65-0.85 with roughly 500 reports per class. Shows severity is predictable at intake from text alone, so an impact score can be ex-ante. Severity is the defect's technical effect; it is not fix effort."}]},{"k":"landis-1977","a":"foundations-stats","t":"The Measurement of Observer Agreement for Categorical Data","y":1977,"s":"peer_reviewed","e":6,"d":"10.2307/2529310","v":"Biometrics","a_":["Landis & Koch"],"bl":"https://doi.org/10.2307/2529310","pc":1,"sm":[{"m":"agent draft (claude-sonnet-5-5)","l":"complexity measurement without effort","tx":"Source of the widely used kappa benchmark bands (0.61-0.80 substantial, above 0.80 almost perfect). The authors themselves call the cutoffs arbitrary, so use them descriptively only. Report kappa per rubric item with confidence intervals and prevalence context, and set a pre-registered minimum (e.g."}]},{"k":"langdale2019simdjson","a":"foundations-systems","t":"Parsing gigabytes of JSON per second","y":2019,"s":"peer_reviewed","e":4,"d":"10.1007/s00778-019-00578-5","v":"The VLDB Journal","a_":["Langdale & Lemire"],"bl":"https://doi.org/10.1007/s00778-019-00578-5","pc":2,"sm":[{"m":"agent draft (gemini)","l":"systems performance","tx":"Extracting the 13 Prometheus label fields (json_extract_string) in the inner subquery s before unnest(s.vals) reduces JSON object traversal from O(N_series * M_samples) (~190 samples/series) to O(N_series)."},{"m":"agent draft (gemini)","l":"operational translation","tx":"simdjson (VLDBJ '19) proves that branch-heavy byte-by-byte text lexing can be replaced by a two-stage SIMD pipeline: Stage 1 loads 64-byte blocks into SIMD registers, computes bitmasks of structural characters and character classes (newlines, digits, hex characters, quotes, backslashes) using branchless vector comparisons and `pshufb`/`tbl` lookups,…"}]},{"k":"laurie-2014","a":"security-governance","t":"Certificate Transparency","y":2014,"s":"peer_reviewed","e":1,"d":"10.1145/2668152.2668154","v":"Queue","a_":["Laurie"],"bl":"https://doi.org/10.1145/2668152.2668154","pc":1,"sm":[{"m":"agent draft (qwen)","l":"operational translation","tx":"CT is the canonical template for ybc's R16 --anchor design: an evidence producer (CA/bundle builder) must not be the sole authority over its own log; an external append-only chain (Merkle tree/SHA-256 manifest chain) makes silent tampering detectable."}]},{"k":"le-2023","a":"foundations-telemetry","t":"Log Parsing with Prompt-based Few-shot Learning","y":2023,"s":"peer_reviewed","e":10,"d":"10.1109/ICSE48619.2023.00204","v":"2023 IEEE/ACM 45th International Conference on Software Engineering (ICSE)","a_":["Le & Zhang"],"bl":"https://doi.org/10.1109/ICSE48619.2023.00204","pc":1,"sm":[{"m":"agent draft (gemini)","l":"log template mining","tx":"Le and Zhang (ICSE 2023, LogPPT) formulate log parsing as prompt-based few-shot token classification using RoBERTa with virtual label tokens to identify keywords vs parameters from 32 adaptive few-shot training examples."}]},{"k":"le2022logad","a":"ai-operations","t":"Log-based anomaly detection with deep learning","y":2022,"s":"peer_reviewed","e":3,"d":"10.1145/3510003.3510155","v":"Proceedings of the 44th International Conference on Software Engineering","a_":["Le & Zhang"],"bl":"https://doi.org/10.1145/3510003.3510155","pc":2,"sm":[{"m":"agent draft (claude)","l":"limitations caveats","tx":"Le and Zhang (ICSE 2022) re-evaluate deep-learning log anomaly detectors under realistic conditions (chronological splits, parsing errors, class imbalance, unstable logs) and find reported performance largely does not hold up."}]},{"k":"lee-2004","a":"foundations-se","t":"Trust in Automation: Designing for Appropriate Reliance","y":2004,"s":"peer_reviewed","e":7,"d":"10.1518/hfes.46.1.50_30392","v":"Human Factors: The Journal of the Human Factors and Ergonomics Society","a_":["Lee & See"],"bl":"https://doi.org/10.1518/hfes.46.1.50_30392","pc":1,"sm":[{"m":"agent draft (claude)","l":"support human ai","tx":"Defines appropriate reliance as trust that is calibrated to the automation's actual capability, with good resolution (it differentiates situations) and specificity (it applies to particular functions). Trust rests on performance, process and purpose information."}]},{"k":"lee2009bounds","a":"foundations-stats","t":"Training, Wages, and Sample Selection: Estimating Sharp Bounds on Treatment Effects","y":2009,"s":"peer_reviewed","e":4,"d":"10.1111/j.1467-937X.2009.00536.x","v":"Review of Economic Studies","a_":["LEE"],"bl":"https://doi.org/10.1111/j.1467-937X.2009.00536.x","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Sharp bounds on a treatment effect when the outcome is observed only for a selected subsample whose size the treatment changes, under monotonicity of selection. Trim the arm with the larger selected share by the excess fraction from the top or bottom of its outcome distribution."}]},{"k":"lee2023eadro_icse","a":"foundations-telemetry","t":"Eadro: An End-to-End Troubleshooting Framework for Microservices on Multi-source Data","y":2023,"s":"peer_reviewed","e":3,"d":"10.1109/ICSE48619.2023.00150","v":"2023 IEEE/ACM 45th International Conference on Software Engineering (ICSE)","a_":["Lee et al."],"bl":"https://doi.org/10.1109/ICSE48619.2023.00150","pc":1,"sm":[{"m":"agent draft (claude)","l":"applicability to support bundle fleet","tx":"Eadro jointly learns anomaly detection and root-cause localization from three modalities (logs as event counts, KPIs/metrics, and traces) with modality-specific encoders fused through a dependency graph, and shows that multi-source fusion beats single-source methods on microservice benchmarks."}]},{"k":"leeflang-2009","a":"foundations-stats","t":"Diagnostic test accuracy may vary with prevalence: implications for evidence-based diagnosis","y":2009,"s":"peer_reviewed","e":1,"d":"10.1016/j.jclinepi.2008.04.007","v":"Journal of Clinical Epidemiology","a_":["Leeflang et al."],"bl":"https://doi.org/10.1016/j.jclinepi.2008.04.007","pc":1,"sm":[{"m":"agent draft (claude)","l":"limitations caveats","tx":"Sensitivity and specificity are not fixed properties of a test. They shift with prevalence and with the case spectrum, through mechanisms such as patient mix, referral filters and the reference standard."}]},{"k":"leesatapornwongsa2016taxdc","a":"foundations-systems","t":"TaxDC","y":2016,"s":"peer_reviewed","e":4,"d":"10.1145/2872362.2872374","v":"Proceedings of the Twenty-First International Conference on Architectural Support for Programming Languages and Operating Systems","a_":["Leesatapornwongsa et al."],"bl":"https://doi.org/10.1145/2872362.2872374","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Taxonomy of 104 distributed concurrency bugs in Cassandra/HBase/Hadoop/ZooKeeper: timing (message ordering, races with faults/reboots), triggering inputs and fixes. Useful to classify YB field FATALs such as 'Unexpected new leader in same term' that look like Raft message-ordering races rather than environment faults."}]},{"k":"leijen2019mimalloc_aplas","a":"foundations-systems","t":"Mimalloc: Free List Sharding in Action","y":2019,"s":"peer_reviewed","e":1,"d":"10.1007/978-3-030-34175-6_13","v":"Lecture Notes in Computer Science","a_":["Leijen et al."],"bl":"https://doi.org/10.1007/978-3-030-34175-6_13","pc":1,"sm":[{"m":"agent draft (gemini)","l":"allocator optimization","tx":"Introduces free-list sharding per mimalloc page to eliminate multi-threaded allocator lock contention and arena fragmentation. In WTL-P2, switching from glibc malloc to mimalloc reduced wall time to 0.895x (95% CI [0.866, 0.924]), RSS to 0.746x [0.592, 0.957] across 11 bundles, and collapsed Case 19698 (33.6 GB) peak RSS from 52.9 GiB to 12.8 GiB (-75.8%)."}]},{"k":"leinonen2023llmerrors","a":"foundations-se","t":"Using Large Language Models to Enhance Programming Error Messages","y":2023,"s":"peer_reviewed","e":2,"d":"10.1145/3545945.3569770","v":"Proceedings of the 54th ACM Technical Symposium on Computer Science Education V. 1","a_":["Leinonen et al."],"bl":"https://doi.org/10.1145/3545945.3569770"},{"k":"leis-2014","a":"ai-operations","t":"Morsel-driven parallelism","y":2014,"s":"peer_reviewed","e":3,"d":"10.1145/2588555.2610507","v":"Proceedings of the 2014 ACM SIGMOD International Conference on Management of Data","a_":["Leis et al."],"bl":"https://doi.org/10.1145/2588555.2610507","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Morsel-driven scheduling with elastic parallelism. In DuckDB the morsel is effectively a row group (122,880 rows), so scans with fewer row groups than threads under-parallelize; check row groups vs threads."}]},{"k":"leis-2015","a":"ai-operations","t":"How good are query optimizers, really?","y":2015,"s":"peer_reviewed","e":4,"d":"10.14778/2850583.2850594","v":"Proceedings of the VLDB Endowment","a_":["Leis et al."],"bl":"https://doi.org/10.14778/2850583.2850594","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"JOB benchmark on IMDB: all tested cardinality estimators (incl. PostgreSQL) routinely err by orders of magnitude on multi-join queries, errors growing with join count; cardinality errors dominate cost-model errors for plan quality; PostgreSQL non-index nested-loop joins under underestimates cause timeouts (disabling them removed all timeouts)."},{"m":"agent draft (claude)","l":"operational translation","tx":"For a YSQL profiler: cardinality mis-estimation, not cost-model constants, dominates plan quality; nested-loop choices driven by underestimates are the main failure. Operationally: compare estimated vs actual rows per node first, and treat an un-ANALYZEd YB table (1000-row default heuristic) as the YB analogue of the JOB underestimation problem;"}]},{"k":"lepers-2019","a":"foundations-systems","t":"KVell","y":2019,"s":"peer_reviewed","e":3,"d":"10.1145/3341301.3359628","v":"Proceedings of the 27th ACM Symposium on Operating Systems Principles","a_":["Lepers et al."],"bl":"https://doi.org/10.1145/3341301.3359628","pc":1,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"KVell demonstrates that on modern multi-queue NVMe SSDs, traditional LSM/B-tree storage engines become CPU-, synchronization-, and memory-copy-bound rather than disk-I/O-bound, motivating zero-copy shared-nothing I/O and CPU-per-op minimization."}]},{"k":"lex-2014","a":"foundations-viz","t":"UpSet: Visualization of Intersecting Sets","y":2014,"s":"peer_reviewed","e":2,"d":"10.1109/TVCG.2014.2346248","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Lex et al."],"bl":"https://doi.org/10.1109/TVCG.2014.2346248","pc":1,"sm":[{"m":"agent draft (claude-opus-5.5)","l":"operational translation","tx":"UpSet: matrix plus bar layout for set intersections, scaling past Venn/Euler. Rationale for not using a Sankey for audit class to sub-group, where cases belong to many sub-groups (overlapping memberships, not a flow partition)."}]},{"k":"li2007tcloseness","a":"security-governance","t":"t-Closeness: Privacy Beyond k-Anonymity and l-Diversity","y":2007,"s":"peer_reviewed","e":2,"d":"10.1109/ICDE.2007.367856","v":"2007 IEEE 23rd International Conference on Data Engineering","a_":["Li et al."],"bl":"https://doi.org/10.1109/ICDE.2007.367856"},{"k":"li2008outoforder","a":"foundations-systems","t":"Out-of-order processing","y":2008,"s":"peer_reviewed","e":1,"d":"10.14778/1453856.1453890","v":"Proceedings of the VLDB Endowment","a_":["Li et al."],"bl":"https://doi.org/10.14778/1453856.1453890","sm":[{"m":"vault note","l":"summary","tx":"Li et al. (PVLDB 2008) Out-of-Order Processing: A New Architecture for High-Performance Stream Systems"}]},{"k":"li2017mison","a":"foundations-systems","t":"Mison","y":2017,"s":"peer_reviewed","e":3,"d":"10.14778/3115404.3115416","v":"Proceedings of the VLDB Endowment","a_":["Li et al."],"bl":"https://doi.org/10.14778/3115404.3115416","pc":1,"sm":[{"m":"agent draft (gemini)","l":"systems performance","tx":"Demonstrates that analytical queries over nested JSON spend the majority of CPU cycles materializing unused fields and array elements unless projection pushdown and structural filtering precede unnesting."}]},{"k":"li2020gandalf","a":"ai-operations","t":"Gandalf: An Intelligent, End-To-End Analytics Service for Safe Deployment in Cloud-Scale Infrastructure","y":2020,"s":"peer_reviewed","e":2,"v":"17th USENIX Symposium on Networked Systems Design and Implementation (NSDI)","a_":["Li et al."],"bl":"https://www.usenix.org/conference/nsdi20/presentation/li","pc":3,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Gandalf (Microsoft Azure, NSDI 2020) uses fleet-wide telemetry for safe deployment. It correlates failure signals with rollouts in time and space and decides to stop or continue a rollout automatically. For a database vendor it is the peer pattern for the highest-value use of field data: release safety."}]},{"k":"lightman2023verify","a":"foundations-other","t":"Let's Verify Step by Step","y":2023,"s":"preprint","e":1,"x":"2305.20050","v":"arXiv:2305.20050","a_":["Lightman et al."],"bl":"https://arxiv.org/abs/2305.20050","pc":2,"sm":[{"m":"agent draft (qwen)","l":"operational translation","tx":"The headline result — a small specialized PRM guiding best-of-N search outperforms GPT-4's own critic — is the strongest literature support for System-1 gating of a System-2 generator without the small model ever producing the answer itself."}]},{"k":"lin2013agnostic","a":"foundations-stats","t":"Agnostic notes on regression adjustments to experimental data: Reexamining Freedman’s critique","y":2013,"s":"peer_reviewed","e":1,"d":"10.1214/12-AOAS583","v":"The Annals of Applied Statistics","a_":["Lin"],"bl":"https://doi.org/10.1214/12-AOAS583"},{"k":"lin2015cowic","a":"foundations-telemetry","t":"Cowic: A Column-Wise Independent Compression for Log Stream Analysis","y":2015,"s":"peer_reviewed","e":4,"d":"10.1109/CCGrid.2015.45","v":"2015 15th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing","a_":["Lin et al."],"bl":"https://doi.org/10.1109/CCGrid.2015.45","pc":1,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Cowic pioneers column-wise independent compression for log streams so individual log entries or columns (timestamps, IPs, template bodies) can be decompressed and queried independently without inflating entire multi-megabyte archive blocks."}]},{"k":"lin2016logcluster","a":"foundations-telemetry","t":"Log clustering based problem identification for online service systems","y":2016,"s":"peer_reviewed","e":13,"d":"10.1145/2889160.2889232","v":"Proceedings of the 38th International Conference on Software Engineering Companion","a_":["Lin et al."],"bl":"https://doi.org/10.1145/2889160.2889232","pc":3,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Microsoft LogCluster: cluster failure log sequences, keep one representative per cluster, and match new clusters against a knowledge base of previously diagnosed clusters so engineers only inspect novel ones."},{"m":"agent draft (gemini)","l":"operational translation","tx":"Lin, Zhang, Lou et al. (ICSE 2016) introduce LogCluster for Microsoft online service systems, weighting log templates by inverse document/incident frequency (IDF) and contrastively comparing log sequences between anomalous and normal windows/nodes."}]},{"k":"lin2025sleep","a":"foundations-stats","t":"Sleep-time Compute: Beyond Inference Scaling at Test-time","y":2025,"s":"preprint","e":1,"x":"2504.13171","v":"arXiv:2504.13171","a_":["Lin et al."],"bl":"https://arxiv.org/abs/2504.13171","pc":2,"sm":[{"m":"agent draft (qwen)","l":"operational translation","tx":"Sleep-time compute is literally the Security-1 daemon economics: pre-computing inferences over known context before queries arrive cut test-time compute ~5x at equal accuracy and amortized 2.5x across related queries; gains scale with query predictability."}]},{"k":"lingren2014preannotation","a":"foundations-other","t":"Evaluating the impact of pre-annotation on annotation speed and potential bias: natural language processing gold standard development for clinical named entity recognition in clinical trial announcements","y":2014,"s":"peer_reviewed","e":3,"d":"10.1136/amiajnl-2013-001837","v":"Journal of the American Medical Informatics Association","a_":["Lingren et al."],"bl":"https://doi.org/10.1136/amiajnl-2013-001837","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Clinical NER gold-standard study: dictionary pre-annotation shortened annotation time without a measurable bias in their setting, but authors caution that pre-annotation can bias annotators toward the pre-labels; motivates keeping AI labels hidden from human raters when humans are the reference for measuring AI agreement."}]},{"k":"lipsitch-2010","a":"foundations-stats","t":"Negative Controls","y":2010,"s":"peer_reviewed","e":5,"d":"10.1097/EDE.0b013e3181d61eeb","v":"Epidemiology","a_":["Lipsitch et al."],"bl":"https://doi.org/10.1097/EDE.0b013e3181d61eeb","pc":2,"sm":[{"m":"agent draft (claude)","l":"hagen impact causal","tx":"Defines negative-control exposures and outcomes: variables that share the confounding structure of the main analysis but cannot be causally affected, so any association signals bias."},{"m":"agent draft (claude)","l":"operational translation","tx":"Negative-control exposures and outcomes detect confounding and bias when they share the bias structure but not the causal path. Applied to the sigrel release-regression detector: client SQL errors (relation/column does not exist, password authentication failed) are flagged as release rises, so the detector also picks up log-plane and ticket-rate…"}]},{"k":"little-1961","a":"foundations-telemetry","t":"A Proof for the Queuing Formula: L = λ W","y":1961,"s":"peer_reviewed","e":1,"d":"10.1287/opre.9.3.383","v":"Operations Research","a_":["Little"],"bl":"https://doi.org/10.1287/opre.9.3.383","pc":2,"sm":[{"m":"agent draft (claude)","l":"metric statistics review","tx":"L = lambda W holds for any stationary queue. Used as a consistency check across three separately derived dashboard numbers: 55.8 customer tickets/week x mean 18.5 days to resolve implies ~147 open, against 124 observed open now, a ~16% gap that points at the estimated solve time (created + summed status durations) or reopen handling."},{"m":"agent draft (claude)","l":"support ops dashboard","tx":"Proof of L = λW for stationary queues. For support dashboards: backlog (L), inflow (λ) and mean time-in-system (W) are linked, so a fixed backlog target (e.g. 100) is meaningful only relative to inflow; W = L/λ gives a consistency check on reported time-to-resolve. Requires steady state and consistent boundaries (same ticket population for L, λ, W)."}]},{"k":"liu-2012","a":"foundations-other","t":"Task complexity: A review and conceptualization framework","y":2012,"s":"peer_reviewed","e":3,"d":"10.1016/j.ergon.2012.09.001","v":"International Journal of Industrial Ergonomics","a_":["Liu & Li"],"bl":"https://doi.org/10.1016/j.ergon.2012.09.001","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Framework of task-complexity dimensions (size, variety, ambiguity, relationship, variability, unreliability, novelty, incongruity, action complexity, temporal demand), separating objective from subjective complexity; vocabulary for ex-ante (presented) vs ex-post (realised) ticket complexity."}]},{"k":"liu-2018","a":"foundations-viz","t":"Somewhere Over the Rainbow","y":2018,"s":"peer_reviewed","e":5,"d":"10.1145/3173574.3174172","v":"Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems","a_":["Liu & Heer"],"bl":"https://doi.org/10.1145/3173574.3174172","pc":1,"sm":[{"m":"agent draft (claude)","l":"empirical findings","tx":"Relative-distance judgments over four single-hue and five multi-hue quantitative colormaps: perceptually-uniform multi-hue ramps (viridis family) performed at least as well as single-hue luminance ramps, rainbow-type maps were worse, and combining perceptual-space distance with color-naming predicted performance better than either alone."}]},{"k":"liu2015opprentice","a":"foundations-telemetry","t":"Opprentice","y":2015,"s":"peer_reviewed","e":1,"d":"10.1145/2815675.2815679","v":"Proceedings of the 2015 Internet Measurement Conference","a_":["Liu et al."],"bl":"https://doi.org/10.1145/2815675.2815679","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Opprentice ensembles many cheap detectors under supervised operator labels."}]},{"k":"liu2019cloudincidents","a":"ai-operations","t":"What Bugs Cause Production Cloud Incidents?","y":2019,"s":"peer_reviewed","e":4,"d":"10.1145/3317550.3321438","v":"Proceedings of the Workshop on Hot Topics in Operating Systems (HotOS)","a_":["Liu et al."],"bl":"https://doi.org/10.1145/3317550.3321438","pc":3,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Microsoft Azure study of high-severity production incidents: many come from timing, fault handling and inter-component data problems, and detection often relies on customers. Analogue for YB: field incidents surface via tickets, so ticket-derived rates are a lower bound."}]},{"k":"liu2024lost","a":"ai-operations","t":"Lost in the Middle: How Language Models Use Long Contexts","y":2024,"s":"peer_reviewed","e":2,"d":"10.1162/tacl_a_00638","v":"Transactions of the Association for Computational Linguistics (TACL)","a_":["Liu et al."],"bl":"https://doi.org/10.1162/tacl_a_00638","pc":2,"sm":[{"m":"agent draft (gemini)","l":"agentic observability","tx":"Proves that even when LLMs accept 100k+ token inputs, retrieval and multi-hop reasoning degrade sharply in a U-shaped curve when relevant facts sit in the middle of long contexts (>32k tokens)."}]},{"k":"lo-2022","a":"foundations-viz","t":"Misinformed by Visualization: What Do We Learn From Misinformative Visualizations?","y":2022,"s":"peer_reviewed","e":3,"d":"10.1111/cgf.14559","v":"Computer Graphics Forum","a_":["Lo et al."],"bl":"https://doi.org/10.1111/cgf.14559"},{"k":"lopezbernal2017its","a":"foundations-stats","t":"Interrupted time series regression for the evaluation of public health interventions: a tutorial","y":2016,"s":"peer_reviewed","e":5,"d":"10.1093/ije/dyw098","v":"International Journal of Epidemiology","a_":["Bernal et al."],"bl":"https://doi.org/10.1093/ije/dyw098","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Applied as the Wave C outcome figure for the Bundle Insights gallery: segmented regression of the monthly bundled-minus-unbundled gap in median TTR (KM) and 30-day CE rate around the bot start (2026-06-15), phase-in month excluded, Newey-West SEs."}]},{"k":"lopezbernal2018controls","a":"foundations-stats","t":"The use of controls in interrupted time series studies of public health interventions","y":2018,"s":"peer_reviewed","e":4,"d":"10.1093/ije/dyy135","v":"International Journal of Epidemiology","a_":["Bernal et al."],"bl":"https://doi.org/10.1093/ije/dyy135","pc":2,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Controls in ITS: a control series that did not receive the intervention removes co-interventions and secular shocks common to both; model either the difference series or a group x post interaction. Requires the control to be unaffected by the intervention and to share time-varying confounders."},{"m":"agent draft (claude)","l":"operational translation","tx":"Controls in ITS (control series, placebo dates, non-equivalent outcomes) address history threats."}]},{"k":"lou-2013","a":"foundations-telemetry","t":"Software analytics for incident management of online services: An experience report","y":2013,"s":"peer_reviewed","e":3,"d":"10.1109/ASE.2013.6693105","v":"2013 28th IEEE/ACM International Conference on Automated Software Engineering (ASE)","a_":["Lou et al."],"bl":"https://doi.org/10.1109/ASE.2013.6693105","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Microsoft ASE 2013 experience report on software analytics for incident management: on-call engineers need fast retrieval of similar historical incidents plus the suspicious signals for the current one, presented together. Supports a case-open panel pairing vital signs with history."}]},{"k":"lozi-2016","a":"ai-operations","t":"The Linux scheduler","y":2016,"s":"peer_reviewed","e":1,"d":"10.1145/2901318.2901326","v":"Proceedings of the Eleventh European Conference on Computer Systems","a_":["Lozi et al."],"bl":"https://doi.org/10.1145/2901318.2901326","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Lozi et al. show that four bugs in CFS load balancing (group imbalance, scheduling-group construction, overload-on-wakeup, missing domains) leave cores idle while runnable threads wait. They found them with a sanity checker for the invariant 'no idle core while threads wait' and with heatmap visualizations."}]},{"k":"lu-2017","a":"distributed-systems","t":"WiscKey","y":2017,"s":"peer_reviewed","e":4,"d":"10.1145/3033273","v":"ACM Transactions on Storage","a_":["Lu et al."],"bl":"https://doi.org/10.1145/3033273","pc":1,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"WiscKey separates keys (stored sorted in the LSM-tree) from values (stored in an append-only vLog), proving that LSM compaction re-sorts values unnecessarily when only keys are needed for indexing. Reduces write amplification by an order of magnitude and achieves 2.5x-111x faster DB load and 1.6x-14x faster YCSB workloads than LevelDB."}]},{"k":"lu2023perseus","a":"ai-operations","t":"Perseus: A Fail-Slow Detection Framework for Cloud Storage Systems","y":2023,"s":"peer_reviewed","e":6,"v":"21st USENIX Conference on File and Storage Technologies (FAST 23)","a_":["Lu et al."],"bl":"https://www.usenix.org/conference/fast23/presentation/lu","pc":10,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"fsync T5-T7: Perseus (Alibaba, FAST'23) detects fail-slow drives from 15 s latency/throughput samples. A per-node polynomial regression of latency on throughput gives an adaptive upper bound (95%/99.9% prediction bounds); slowdown ratio SR = latency / upper bound;"},{"m":"agent draft (claude)","l":"methodology","tx":"Perseus scores slowdowns by slowdown ratio x duration bands (minutes to hours) from 15 s samples: a volume-style risk score for chronic fail-slow drives. Too coarse for 5-60 s acute freezes, which need per-event duration from logs."},{"m":"agent draft (claude)","l":"operational translation","tx":"fsync POC: Perseus's per-drive latency-vs-throughput outlier model motivates keeping device/model labels in promdump ETL and a relative-tail alert; not yet replicable because node_exporter rollups are empty and log_sync_latency is sum/count only."}]},{"k":"lu2025xinda","a":"ai-operations","t":"One-Size-Fits-None: Understanding and Enhancing Slow-Fault Tolerance in Modern Distributed Systems","y":2025,"s":"peer_reviewed","e":2,"v":"22nd USENIX Symposium on Networked Systems Design and Implementation (NSDI 25)","a_":["Lu et al."],"bl":"https://www.usenix.org/conference/nsdi25/presentation/lu","pc":5,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"fsync T7: Xinda (NSDI'25) finds slow-fault impact non-monotonic in magnitude (a small slow fault can hurt more than a large one; slow followers can cost about 1.5x more than slow leaders) and ships the ADR runtime detector."},{"m":"agent draft (claude)","l":"empirical findings","tx":"Xinda (NSDI'25): slow-fault impact in Raft-like systems is non-monotonic in slowdown magnitude and depends on which timeout it crosses; slow followers can cost more than slow leaders. Supports a threshold (step) model over a linear volume model."}]},{"k":"lundgard-2022","a":"foundations-viz","t":"Accessible Visualization via Natural Language Descriptions: A Four-Level Model of Semantic Content","y":2022,"s":"peer_reviewed","e":12,"d":"10.1109/TVCG.2021.3114770","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Lundgard & Satyanarayan"],"bl":"https://doi.org/10.1109/TVCG.2021.3114770","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Four-level model of chart descriptions (encoding, statistics, trends, domain context); readers value levels 2-3 most. A per-card screen-reader description, a table view and keyboard-navigable marks cover levels 1-3; a takeaway title covers 3-4."}]},{"k":"luo-2020","a":"foundations-systems","t":"Breaking down memory walls","y":2020,"s":"peer_reviewed","e":9,"d":"10.14778/3430915.3430916","v":"Proceedings of the VLDB Endowment","a_":["Luo & Carey"],"bl":"https://doi.org/10.14778/3430915.3430916","pc":3,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"Studies co-tuning memory allocation between write buffers (MemTables) and read caches (block cache + Bloom filters) in LSM storage systems, showing that static memory partitioning causes up to 2-5x suboptimal I/O bytes per operation under shifting read/write workloads."},{"m":"agent draft (gemini)","l":"systems performance","tx":"Surveys LSM-tree write amplification, L0 SST accumulation, compaction bandwidth contention, and memory-limit write stalls, directly grounding Doctor v2.5's storage and promdump checks in promdump_metrics.py (_check_rocksdb_stalls, _check_write_amplification with a 1 GiB compaction volume floor, _check_memory_pressure using server_memory_soft_limit and…"}]},{"k":"ma2018queryforecast","a":"foundations-systems","t":"Query-based Workload Forecasting for Self-Driving Database Management Systems","y":2018,"s":"peer_reviewed","e":3,"d":"10.1145/3183713.3196908","v":"Proceedings of the 2018 International Conference on Management of Data","a_":["Ma et al."],"bl":"https://doi.org/10.1145/3183713.3196908"},{"k":"maalej-2016","a":"foundations-se","t":"Toward Data-Driven Requirements Engineering","y":2016,"s":"peer_reviewed","e":5,"d":"10.1109/MS.2015.153","v":"IEEE Software","a_":["Maalej et al."],"bl":"https://doi.org/10.1109/MS.2015.153","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Requirements should be driven by continuously mined usage and feedback data (implicit telemetry plus explicit feedback), combined, with attention to representativeness. Supports using bundle/ticket evidence as a discovery input, but alongside usage telemetry (call-home), not instead of it."}]},{"k":"machado-2009","a":"foundations-viz","t":"A Physiologically-based Model for Simulation of Color Vision Deficiency","y":2009,"s":"peer_reviewed","e":2,"d":"10.1109/TVCG.2009.113","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Machado et al."],"bl":"https://doi.org/10.1109/TVCG.2009.113","pc":1,"sm":[{"m":"agent draft (claude)","l":"limitations caveats","tx":"Physiologically-based stage-theory CVD model that handles anomalous trichromacy and dichromacy in one framework via a severity parameter; matrices for protan/deutan at severity 1.0 are the de-facto standard in colour tooling (colour-science, DaltonLens)."}]},{"k":"machanavajjhala2007ldiversity","a":"security-governance","t":"l-Diversity: Privacy Beyond k-Anonymity","y":2007,"s":"peer_reviewed","e":3,"d":"10.1145/1217299.1217302","v":"ACM Transactions on Knowledge Discovery from Data","a_":["Machanavajjhala et al."],"bl":"https://doi.org/10.1145/1217299.1217302","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Shows k-anonymity fails when a k-group shares one sensitive value (homogeneity) or the attacker has background knowledge; requires l well-represented sensitive values per equivalence class. For the fleet gallery: a cell of 5 orgs that all show the same failing posture (e.g. trust-any HBA) is k-anonymous but discloses the attribute;"}]},{"k":"mackinnon-2018","a":"econometrics","t":"The wild bootstrap for few (treated) clusters","y":2018,"s":"peer_reviewed","e":5,"d":"10.1111/ectj.12107","v":"The Econometrics Journal","a_":["MacKinnon & Webb"],"bl":"https://doi.org/10.1111/ectj.12107","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Cluster-robust t and pairs/wild bootstraps over-reject when few clusters are treated; wild restricted (WCR) and subcluster bootstraps fix much of it. Relevant to relrate (2025.2 rests on 10-17 orgs), fs_infra Azure (6 orgs) and uppath paths (8-9 orgs): percentile pairs-cluster bootstraps there undercover."}]},{"k":"mackinnon2017wild","a":"econometrics","t":"Wild Bootstrap Inference for Wildly Different Cluster Sizes","y":2016,"s":"peer_reviewed","e":3,"d":"10.1002/jae.2508","v":"Journal of Applied Econometrics","a_":["MacKinnon & Webb"],"bl":"https://doi.org/10.1002/jae.2508","pc":4,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Wild cluster bootstrap with wildly unequal cluster sizes over-rejects when one cluster dominates; relevant to slownode (Rakuten 46% of flagged bundles) and fsync (top 5 orgs 54% of stalls)."}]},{"k":"maclure1991casecrossover","a":"ai-operations","t":"The Case-Crossover Design: A Method for Studying Transient Effects on the Risk of Acute Events","y":1991,"s":"peer_reviewed","e":2,"d":"10.1093/oxfordjournals.aje.a115853","v":"American Journal of Epidemiology","a_":["Maclure"],"bl":"https://doi.org/10.1093/oxfordjournals.aje.a115853","pc":3,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Change-before-ticket rates need a referent window from the same unit. A case-crossover design compares the hazard window (7 days before a ticket) with control windows from the same universe, giving an odds ratio rather than a bare share. Precondition: the exposure must be observable in control windows."}]},{"k":"majumdar2018partition","a":"foundations-telemetry","t":"Why is random testing effective for partition tolerance bugs?","y":2017,"s":"peer_reviewed","e":1,"d":"10.1145/3158134","v":"Proceedings of the ACM on Programming Languages","a_":["Majumdar & Niksic"],"bl":"https://doi.org/10.1145/3158134","pc":1},{"k":"manski-2000","a":"econometrics","t":"Monotone Instrumental Variables: With an Application to the Returns to Schooling","y":2000,"s":"peer_reviewed","e":1,"d":"10.1111/1468-0262.00144","v":"Econometrica","a_":["Manski & Pepper"],"bl":"https://doi.org/10.1111/1468-0262.00144","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Monotone treatment selection (MTS) and monotone IV assumptions tighten worst-case Manski bounds. Applied to bundle-only rates (Bundle Insights bias/fs_prev): assuming bundled tickets are at least as severe as unbundled (MTS), the bundle-only rate is an upper bound on the queue rate, which the CE validation (26% vs 14%) supports;"}]},{"k":"mark2016emailduration","a":"foundations-se","t":"Email Duration, Batching and Self-interruption","y":2016,"s":"peer_reviewed","e":2,"d":"10.1145/2858036.2858262","v":"Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems","a_":["Mark et al."],"bl":"https://doi.org/10.1145/2858036.2858262","pc":1,"sm":[{"m":"agent draft (claude)","l":"effort estimation","tx":"Logged foreground email-client time for 40 information workers over ~12 workdays: nearly 1.5 h/day on email, 77 checks/day, about 4.5 h/day total computer time. Email duration was sessionized with a 5-minute keyboard/mouse inactivity cutoff. Gives a daily email-time anchor and an example of a short inactivity cutoff for in-application activity;"}]},{"k":"mcc2025imbalance","a":"ai-operations","t":"The Advantages of the Matthews Correlation Coefficient (MCC) over F1 Score and Accuracy in Binary Classification Evaluation","y":2025,"s":"peer_reviewed","e":1,"d":"10.1186/s12864-019-6413-7","v":"BMC Genomics","a_":["Chicco & Jurman"],"bl":"https://doi.org/10.1186/s12864-019-6413-7","pc":1},{"k":"mcduff2025ddxllm","a":"foundations-other","t":"Towards accurate differential diagnosis with large language models","y":2025,"s":"peer_reviewed","e":2,"d":"10.1038/s41586-025-08869-4","v":"Nature","a_":["McDuff et al."],"bl":"https://doi.org/10.1038/s41586-025-08869-4","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"LLM optimised for differential diagnosis evaluated on 302 NEJM cases with top-k accuracy and specialist ratings of list quality/appropriateness/comprehensiveness, plus an assisted-clinician arm. Relevant to the triage audit: separates standalone list accuracy from whether the list helps the human, which mirrors helpfulness vs directional correctness."}]},{"k":"mckeeman1998differential","a":"foundations-se","t":"Differential Testing for Software","y":1998,"s":"external","e":2,"v":"Digital Technical Journal","a_":["McKeeman"],"pc":1,"sm":[{"m":"agent draft (gemini)","l":"methodology","tx":"Formulates differential testing (Digital Technical Journal 1998): feeding identical randomized inputs to two independent implementations (such as a verified Lean 4 reference oracle and a SIMD-optimized Rust/C++ engine) so any divergence exposes a translation or implementation defect without manual test oracles."}]},{"k":"mcnutt-2020","a":"ai-operations","t":"Surfacing Visualization Mirages","y":2020,"s":"peer_reviewed","e":4,"d":"10.1145/3313831.3376420","v":"Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems","a_":["McNutt et al."],"bl":"https://doi.org/10.1145/3313831.3376420"},{"k":"meli2019git","a":"security-governance","t":"How Bad Can It Git? Characterizing Secret Leakage in Public GitHub Repositories","y":2019,"s":"peer_reviewed","e":1,"d":"10.14722/ndss.2019.23418","v":"Proceedings 2019 Network and Distributed System Security Symposium","a_":["Meli et al."],"bl":"https://doi.org/10.14722/ndss.2019.23418","pc":4,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Large-scale scan of public GitHub found over 100k leaked secrets; regex plus entropy plus context filters were needed."}]},{"k":"melnik-2010","a":"ai-operations","t":"Dremel","y":2010,"s":"peer_reviewed","e":8,"d":"10.14778/1920841.1920886","v":"Proceedings of the VLDB Endowment","a_":["Melnik et al."],"bl":"https://doi.org/10.14778/1920841.1920886","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Dremel introduced columnar striping of nested records via repetition and definition levels, the model Parquet adopted. For a profiler: nested/repeated columns carry level streams whose cost grows with nesting depth; SizeStatistics level histograms expose this."}]},{"k":"meng2018paradoxes","a":"foundations-stats","t":"Statistical paradises and paradoxes in big data (I): Law of large populations, big data paradox, and the 2016 US presidential election","y":2018,"s":"peer_reviewed","e":3,"d":"10.1214/18-AOAS1161SF","v":"The Annals of Applied Statistics","a_":["Meng"],"bl":"https://doi.org/10.1214/18-AOAS1161SF","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"PM seat (council 5): Data defect correlation: a huge non-random sample can be worth less than a small random one. The bundle fleet (about 1,185 universes) cannot stand in for the installed base without a selection model; call-home or a small random audit is worth more for adoption shares."}]},{"k":"meredith-1993","a":"foundations-other","t":"Measurement Invariance, Factor Analysis and Factorial Invariance","y":1993,"s":"peer_reviewed","e":3,"d":"10.1007/BF02294825","v":"Psychometrika","a_":["Meredith"],"bl":"https://doi.org/10.1007/BF02294825","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Formalises configural, metric (weak), scalar (strong) and strict factorial invariance. Mean comparisons across groups need at least scalar invariance. We test loadings and indicator thresholds by era."}]},{"k":"meyer2017worklife","a":"foundations-se","t":"The Work Life of Developers: Activities, Switches and Perceived Productivity","y":2017,"s":"peer_reviewed","e":4,"d":"10.1109/TSE.2017.2656886","v":"IEEE Transactions on Software Engineering","a_":["Meyer et al."],"bl":"https://doi.org/10.1109/TSE.2017.2656886","pc":1,"sm":[{"m":"agent draft (claude)","l":"measurement validity","tx":"Combined computer-interaction monitoring with self-reports for professional developers at several companies: work is highly fragmented (frequent activity switches within minutes), and a large share of the day goes to email, meetings and collaboration rather than coding."}]},{"k":"min2023factscore","a":"foundations-other","t":"FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generation","y":2023,"s":"preprint","e":1,"x":"2305.14251","v":"arXiv:2305.14251","a_":["Min et al."],"bl":"https://arxiv.org/abs/2305.14251","pc":2,"sm":[{"m":"agent draft (claude)","l":"llmops","tx":"Decomposes long-form generations into atomic facts and scores the share supported by a knowledge source."},{"m":"agent draft (claude)","l":"eval kpis","tx":"FActScore decomposes long-form generations into atomic facts and computes the fraction supported by a knowledge source, with an automated estimator validated against human annotation."}]},{"k":"miranda-2003","a":"foundations-other","t":"Nursing activities score","y":2003,"s":"peer_reviewed","e":2,"d":"10.1097/01.CCM.0000045567.78801.CC","v":"Critical Care Medicine","a_":["Miranda et al."],"bl":"https://doi.org/10.1097/01.CCM.0000045567.78801.CC","pc":3,"sm":[{"m":"agent draft (claude)","l":"support effort scoring","tx":"The Nursing Activities Score (NAS) replaced TISS-28 with 23 activity items whose weights were fitted to nursing time measured in a multinational time-and-motion study; it explains about 81% of nursing time versus 43% for TISS-28."},{"m":"agent draft (claude)","l":"workload measurement analogue for support tickets","tx":"The Nursing Activities Score (NAS) replaced TISS-28's intervention counts with 23 activity items whose weights were fitted to measured nursing time from a multinational time-and-motion study (99 ICUs). It explains 81% of nursing time, against 43% for TISS-28."},{"m":"agent draft (claude-sonnet-5-5)","l":"construct separation","tx":"Nursing Activities Score weights 23 activities by measured time, so the score is nursing workload (effort) and is explicitly not severity of illness; it explains about 81% of nursing time versus 43% for TISS-28, which was tied to severity. Clean example of an effort measure built from time with no severity items."}]},{"k":"mitchell2019modelcards","a":"security-governance","t":"Model Cards for Model Reporting","y":2019,"s":"peer_reviewed","e":6,"d":"10.1145/3287560.3287596","v":"Proceedings of the Conference on Fairness, Accountability, and Transparency","a_":["Mitchell et al."],"bl":"https://doi.org/10.1145/3287560.3287596","pc":1,"sm":[{"m":"agent draft (claude)","l":"iso42001 audit","tx":"Model cards: short structured documents stating intended use, out-of-scope uses, evaluation data, metrics disaggregated by relevant factors, and caveats. Under ISO 42001 a model/system card per AI system is natural evidence for A.6.2.7 technical documentation, A.9.4 intended use and A.8.2 user information."}]},{"k":"mockus-2008","a":"foundations-se","t":"Interval quality","y":2008,"s":"peer_reviewed","e":2,"d":"10.1145/1368088.1368190","v":"Proceedings of the 30th international conference on Software engineering","a_":["Mockus & Weiss"],"bl":"https://doi.org/10.1145/1368088.1368190","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Interval quality: probability a customer hits a field failure within a window after installing a release, joining support data to deployment counts; defect density alone mis-ranks releases. Template for a per-release field-quality metric with an installed-base denominator, which the relrate/upgrade cards approximate with YBA bystander universes."}]},{"k":"mockus2005customerquality","a":"foundations-se","t":"Predictors of customer perceived software quality","y":2005,"s":"peer_reviewed","e":6,"d":"10.1145/1062455.1062506","v":"Proceedings of the 27th international conference on Software engineering - ICSE '05","a_":["Mockus et al."],"bl":"https://doi.org/10.1145/1062455.1062506","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Customer-perceived quality (defects/incidents per installed system) is driven as much by deployment time, platform and install size as by code; comparing releases needs exposure normalisation and a release-age / time-since-install axis, otherwise early-adopter and calendar effects masquerade as release quality."}]},{"k":"mogul2019nines","a":"foundations-systems","t":"Nines are Not Enough","y":2019,"s":"peer_reviewed","e":3,"d":"10.1145/3317550.3321432","v":"Proceedings of the Workshop on Hot Topics in Operating Systems","a_":["Mogul & Wilkes"],"bl":"https://doi.org/10.1145/3317550.3321432","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Argues availability nines hide the distribution of outage durations; customers care how often and how long. Supports reporting fsync stall cost per length bucket rather than one stalled-seconds total: in the YB bundle fleet 80% of >=1 s stalls are 1-5 s but carry only 18% of write-unavailable seconds and 0% of >=5 s timeouts."}]},{"k":"mokander2023auditingllm","a":"security-governance","t":"Auditing large language models: a three-layered approach","y":2023,"s":"peer_reviewed","e":5,"d":"10.1007/s43681-023-00289-2","v":"AI and Ethics","a_":["Mökander et al."],"bl":"https://doi.org/10.1007/s43681-023-00289-2","pc":2,"sm":[{"m":"agent draft (claude)","l":"iso42001 audit","tx":"Argues LLM audits need three coordinated layers: governance audits of the provider (management system), model audits (pre-release capability/limitation evaluation), and application audits (use-case-specific, continuous). Key limitation the authors flag: no layer alone suffices and application-level risks are context-specific."},{"m":"agent draft (gemini)","l":"operational translation","tx":"Proposes a three-layered auditing architecture for foundation models and LLMs: (1) Governance audits of model providers (organizational processes, risk management, data provenance), (2) Model audits (pre-deployment capability, calibration, and safety evaluations), and (3) Application audits (continuous, context-specific validation of downstream…"}]},{"k":"montgomery2017escalation","a":"ai-operations","t":"What Do Support Analysts Know About Their Customers? On the Study and Prediction of Support Ticket Escalations in Large Software Organizations","y":2017,"s":"peer_reviewed","e":4,"d":"10.1109/RE.2017.61","v":"2017 IEEE 25th International Requirements Engineering Conference (RE)","a_":["Montgomery & Damian"],"bl":"https://doi.org/10.1109/RE.2017.61","pc":8,"sm":[{"m":"agent draft (claude)","l":"empirical findings","tx":"RE'17 study at IBM: support analysts rely on customer context (history, account temperature) to anticipate escalations; formalised into features that predict Critical Situations. The Bundle Insights gallery has no customer-history features for CE prediction (prior CEs per org, recurrence), only bundle signatures;"},{"m":"agent draft (claude)","l":"operational translation","tx":"IBM study: support analysts' knowledge about customers, encoded as features, predicts which support tickets will escalate. Relevant as the closest peer-reviewed precedent for combining customer/environment context with ticket data in a support org;"},{"m":"agent draft (claude)","l":"support ops review","tx":"Studies support-ticket escalation at IBM; customer-level attributes (history, open tickets, recent escalations) predict escalation better than ticket text. Supports building a customer-risk view (account-level open P1/P2, CE escalations, aging, recent breaches) before QBRs rather than ticket-level lists only."}]},{"k":"montgomery2018escalationfe","a":"foundations-telemetry","t":"Customer support ticket escalation prediction using feature engineering","y":2018,"s":"peer_reviewed","e":4,"d":"10.1007/s00766-018-0292-3","v":"Requirements Engineering","a_":["Montgomery et al."],"bl":"https://doi.org/10.1007/s00766-018-0292-3","pc":3,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Montgomery & Damian (IBM Canada, ~2.5M support tickets) engineer customer- and ticket-level features (support analysts' knowledge of customer: number of open tickets, recent escalations, days open, priority changes) and predict escalation (recall ~0.79)."},{"m":"agent draft (claude)","l":"operational translation","tx":"IBM design-science study: escalation risk is predicted from analyst-elicited customer and ticket-process features (customer history, ticket age, activity, recent escalations), not from technical log content. Recall 87.36% over 2.5M tickets and 10k escalations, with an 88% cut in analyst review load."}]},{"k":"morgan2006metrics","a":"foundations-se","t":"The Value of Different Customer Satisfaction and Loyalty Metrics in Predicting Business Performance","y":2006,"s":"peer_reviewed","e":2,"d":"10.1287/mksc.1050.0180","v":"Marketing Science","a_":["Morgan & Rego"],"bl":"https://doi.org/10.1287/mksc.1050.0180","pc":1,"sm":[{"m":"agent draft (claude)","l":"vp support dashboard","tx":"Compares satisfaction and loyalty metrics (average satisfaction, top-2-box, repurchase intent, recommend intent, NPS, complaints) as predictors of firm business performance; average and top-2-box satisfaction predict best while recommendation-based metrics such as NPS have little or no predictive value."}]},{"k":"moritz-2019","a":"foundations-viz","t":"Formalizing Visualization Design Knowledge as Constraints: Actionable and Extensible Models in Draco","y":2019,"s":"peer_reviewed","e":7,"d":"10.1109/TVCG.2018.2865240","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Moritz et al."],"bl":"https://doi.org/10.1109/TVCG.2018.2865240","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Draco formalizes visualization design knowledge as hard and soft constraints, enabling automated checking of chart specs. A gallery can carry a small hard-constraint set (values in domain, CI contains estimate, title numbers equal data numbers) checked at export time."}]},{"k":"morris-2019","a":"foundations-viz","t":"Proposals on Kaplan–Meier plots in medical research and a survey of stakeholder views: KMunicate","y":2019,"s":"peer_reviewed","e":1,"d":"10.1136/bmjopen-2019-030215","v":"BMJ Open","a_":["Morris et al."],"bl":"https://doi.org/10.1136/bmjopen-2019-030215","pc":1,"sm":[{"m":"agent draft (claude-opus-5.5)","l":"operational translation","tx":"KMunicate: survey of 1,100+ stakeholders on Kaplan-Meier presentation; recommends confidence bands and a numbers-at-risk table under the curve. Applied to the ticket-to-bundle and bundle-to-analysis curves for business readers."}]},{"k":"morrison-1968","a":"foundations-systems","t":"PATRICIA—Practical Algorithm To Retrieve Information Coded in Alphanumeric","y":1968,"s":"peer_reviewed","e":4,"d":"10.1145/321479.321481","v":"Journal of the ACM","a_":["Morrison"],"bl":"https://doi.org/10.1145/321479.321481","pc":1,"sm":[{"m":"agent draft (gemini)","l":"log template mining","tx":"Morrison (JACM 1968, PATRICIA) introduces path-compressed radix trees where single-child internal chains are collapsed into multi-byte edge labels (or skip offsets), reducing tree depth to the number of branching points and enabling fast prefix lookup and longest-prefix matching."}]},{"k":"moura2021lean4","a":"foundations-se","t":"The Lean 4 Theorem Prover and Programming Language","y":2021,"s":"peer_reviewed","e":1,"d":"10.1007/978-3-030-79876-5_37","v":"Lecture Notes in Computer Science","a_":["Moura & Ullrich"],"bl":"https://doi.org/10.1007/978-3-030-79876-5_37","pc":4,"sm":[{"m":"agent draft (gemini)","l":"methodology","tx":"Establishes the Lean 4 dependently typed theorem prover and functional programming language (CADE-28). In Report 08, Lean 4 v4.13.0 serves as the zero-axiom proof kernel across five production subsystems (157 theorems, 0 sorry, 0 axiom) and emits C code with @[export] C-ABI entry points for direct in-process differential testing against Rust."},{"m":"agent draft (gemini)","l":"formal verification","tx":"Lean 4 unifies a dependently typed interactive theorem prover in the Calculus of Inductive Constructions (CIC) with a compiled systems programming language emitting C IR with Perceus reference-counting reuse and @[export] C-ABI symbols."},{"m":"agent draft (gemini)","l":"methodology","tx":"Introduces Lean 4, combining interactive theorem proving in the Calculus of Inductive Constructions with a functional programming language using Perceus reference counting and native C code generation."}]},{"k":"muhlbauer2013instant","a":"foundations-systems","t":"Instant loading for main memory databases","y":2013,"s":"peer_reviewed","e":4,"d":"10.14778/2556549.2556555","v":"Proceedings of the VLDB Endowment","a_":["Mühlbauer et al."],"bl":"https://doi.org/10.14778/2556549.2556555","pc":1,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Instant Loading (VLDB '13) achieves wire-speed multi-core bulk loading of variable-length text/CSV files by partitioning raw byte ranges across threads (`pread`/mmap slices) and synchronizing each worker to the first valid record boundary in its slice."}]},{"k":"munzner2014vad","a":"foundations-viz","t":"Visualization Analysis and Design","y":2014,"s":"peer_reviewed","e":7,"d":"10.1201/b17511","v":"A K Peters/CRC Press","a_":["Munzner"],"bl":"https://doi.org/10.1201/b17511","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Munzner's what-why-how: match idiom to task (lookup, compare, rank, trend). Several gallery bar lists are ranking tasks left unsorted, or mix dimensions and units on one axis (fs_infra, pipeline, fleet, closure), which violates expressiveness."},{"m":"agent draft (claude)","l":"dashboard design review","tx":"The what-why-how framework: abstract the task first, then choose the idiom. Expressiveness and effectiveness principles: categorical hue only for categorical data, and position on a common aligned scale for the most important comparison. Also covers juxtaposition (small multiples) vs superposition and the cost of navigation."}]},{"k":"muralidhar-2025","a":"foundations-se","t":"Operationalizing selective transparency using progressive disclosure in artificial intelligence clinical diagnosis systems","y":2025,"s":"peer_reviewed","e":1,"d":"10.1016/j.ijhcs.2025.103591","v":"International Journal of Human-Computer Studies","a_":["Muralidhar et al."],"bl":"https://doi.org/10.1016/j.ijhcs.2025.103591","pc":3,"sm":[{"m":"agent draft (gemini)","l":"Diagnostic UI/UX & Dual Persona Progressive Disclosure","tx":"Demonstrates empirically that selective transparency via progressive disclosure in high-stakes diagnostic decision-support interfaces significantly improves practitioner comprehension and calibration over flat explanation dumps, while affective cues and decorative emojis degrade expert trust."},{"m":"agent draft (gemini)","l":"ux design engineering","tx":"Empirical HCI study in IJHCS (Oct 2025) evaluating progressive disclosure in high-stakes AI diagnostic systems. Demonstrates that layering explanations (high-level verdict -> interactive causal diagram -> on-demand evidence drilldown) significantly improves domain-expert comprehension and decision velocity compared to flat explanation dumps, while…"}]},{"k":"murphy1973new","a":"calibration-conformal","t":"A New Vector Partition of the Probability Score","y":1973,"s":"peer_reviewed","e":1,"d":"10.1175/1520-0450(1973)012<0595:ANVPOT>2.0.CO;2","v":"Journal of Applied Meteorology","a_":["Murphy"],"bl":"https://doi.org/10.1175/1520-0450(1973)012<0595:ANVPOT>2.0.CO;2","pc":2,"sm":[{"m":"agent draft (gemini)","l":"report 07 brier decomposition","tx":"Decomposes the quadratic Brier probability score into three orthogonal components: Reliability (calibration error), Resolution (sharpness/discrimination across outcome bins), and Uncertainty (marginal base-rate variance). Essential for diagnosing when a model has high resolution in logprobs vs collapsed resolution in verbalized confidence."}]},{"k":"musa1993opprofile","a":"foundations-stats","t":"Operational profiles in software-reliability engineering","y":1993,"s":"peer_reviewed","e":1,"d":"10.1109/52.199724","v":"IEEE Software","a_":["Musa"],"bl":"https://doi.org/10.1109/52.199724","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Musa's operational profile is the set of operations and their occurrence probabilities in the field; test effort is allocated in proportion to it so that reliability measured in test predicts field reliability."}]},{"k":"muse2020sqlsmells","a":"foundations-se","t":"On the Prevalence, Impact, and Evolution of SQL Code Smells in Data-Intensive Systems","y":2020,"s":"peer_reviewed","e":3,"d":"10.1145/3379597.3387467","v":"Proceedings of the 17th International Conference on Mining Software Repositories","a_":["Muse et al."],"bl":"https://doi.org/10.1145/3379597.3387467"},{"k":"muthen-1984","a":"foundations-stats","t":"A General Structural Equation Model with Dichotomous, Ordered Categorical, and Continuous Latent Variable Indicators","y":1984,"s":"peer_reviewed","e":2,"d":"10.1007/BF02294210","v":"Psychometrika","a_":["Muthén"],"bl":"https://doi.org/10.1007/BF02294210"},{"k":"muthen-1989","a":"foundations-other","t":"Latent Variable Modeling in Heterogeneous Populations","y":1989,"s":"peer_reviewed","e":2,"d":"10.1007/BF02296397","v":"Psychometrika","a_":["Muthén"],"bl":"https://doi.org/10.1007/BF02296397","pc":1,"sm":[{"m":"agent draft (claude-sonnet-5-5)","l":"complexity measurement without effort","tx":"Multiple-group / MIMIC-style latent variable modelling for heterogeneous populations, with covariate effects on the latent factor and direct effects on indicators (the MIMIC DIF test). Lets us regress the latent complexity on covariates (version, deployment type) and test direct effects on specific items."}]},{"k":"mytkowicz-2010","a":"ai-operations","t":"Evaluating the accuracy of Java profilers","y":2010,"s":"peer_reviewed","e":7,"d":"10.1145/1806596.1806618","v":"Proceedings of the 31st ACM SIGPLAN Conference on Programming Language Design and Implementation","a_":["Mytkowicz et al."],"bl":"https://doi.org/10.1145/1806596.1806618","pc":2,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Four Java profilers disagreed on the hottest method due to yieldpoint sampling bias; random-timer sampling and causality checks (perturb a method, observe runtime) validate profiles. Python analog: in-process signal profilers sample only between bytecodes on the main thread; out-of-process samplers (py-spy, Austin, Tachyon) avoid this."},{"m":"agent draft (claude)","l":"operational translation","tx":"Different sampling profilers disagree on hot methods because samples are biased to yield points; analogous bias exists for YB: in-process stack collection (/threadz, LongOperationTracker) returns empty stacks when a thread is inside tcmalloc alloc/dealloc (GH #34851), and perf with frame pointers omitted truncates stacks."}]},{"k":"mytkowicz2009wrong","a":"foundations-systems","t":"Producing wrong data without doing anything obviously wrong!","y":2009,"s":"peer_reviewed","e":9,"d":"10.1145/1508244.1508275","v":"Proceedings of the 14th international conference on Architectural support for programming languages and operating systems","a_":["Mytkowicz et al."],"bl":"https://doi.org/10.1145/1508244.1508275","pc":4,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Innocuous setup (env size, link order) biases measurements enough to flip conclusions; randomize setup and use causal analysis. For Python agents: multi-process pyperf runs, report distributions, not single before/after deltas."},{"m":"agent draft (claude)","l":"operational translation","tx":"Measurement bias: innocuous setup changes (UNIX env size, link order) shift measured speedups enough to flip conclusions. For profiler agents: one before/after run on one setup cannot validate an optimization; randomize setup or report setup sensitivity."},{"m":"agent draft (gemini)","l":"experimental design","tx":"Demonstrates how subtle environmental confounds (link order, environment size, OS layout) bias systems benchmarks by 10-40%. Grounding for WTL-P2's paired single-VM ABBA/ABCDDCBA protocol and A/A noise calibration."}]},{"k":"mytkowicz2014fsm","a":"foundations-systems","t":"Data-parallel finite-state machines","y":2014,"s":"peer_reviewed","e":4,"d":"10.1145/2541940.2541988","v":"Proceedings of the 19th international conference on Architectural support for programming languages and operating systems","a_":["Mytkowicz et al."],"bl":"https://doi.org/10.1145/2541940.2541988","pc":1,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Mytkowicz et al. (ASPLOS '14) formalize how to execute finite-state machines (such as multi-line log assemblers, tokenizers, and Aho-Corasick/DFA matchers) in data-parallel fashion across chunks of a single input stream by breaking state dependencies across chunk boundaries via small speculative state-transition functions or lookback/convergence properties."}]},{"k":"nagaraj2012distalyzer","a":"ai-operations","t":"Structured Comparative Analysis of Systems Logs to Diagnose Performance Problems","y":2012,"s":"external","e":5,"v":"Proceedings of the 9th USENIX Symposium on Networked Systems Design and Implementation (NSDI '12)","a_":["Nagaraj et al."],"pc":3,"sm":[{"m":"agent draft (gemini)","l":"methodology","tx":"Motivates yb-collector's deterministic node tiering and uniform per-node OS/HTTP collection so hagen-tools can run peer outlier queries (promdump outlier_node, sysstat disk_peers, sql_stats snapshot_coverage)."},{"m":"agent draft (gemini)","l":"operational translation","tx":"Nagaraj et al. (NSDI 2012) present Distalyzer, which diagnoses distributed systems problems by statistically comparing event logs and state metrics between well-behaving and poorly-behaving nodes. For `yb-collector` (`ybc_compare_nodes`), Distalyzer's comparative methodology requires a critical production safeguard: coverage-aware diffing."}]},{"k":"nakatumba2010resourcebehavior","a":"foundations-other","t":"Analyzing Resource Behavior Using Process Mining","y":2010,"s":"peer_reviewed","e":2,"d":"10.1007/978-3-642-12186-9_8","v":"Lecture Notes in Business Information Processing","a_":["Nakatumba & Aalst"],"bl":"https://doi.org/10.1007/978-3-642-12186-9_8","pc":1,"sm":[{"m":"agent draft (claude)","l":"effort estimation","tx":"Mines event logs to show resource service times depend on workload (Yerkes-Dodson-like effects: people work faster under higher load). Implication for ticket effort: a fixed minutes-per-comment prior ignores load; effort per event should be allowed to vary with concurrent queue load, or at least load should be a sensitivity covariate."}]},{"k":"nash2022void","a":"foundations-telemetry","t":"Tales from the VOID: The Scary Truth about Incident Metrics","y":2022,"s":"peer_reviewed","e":1,"a_":["Nash"],"bl":"https://www.usenix.org/conference/srecon22americas/presentation/nash","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"VOID report on thousands of public incident reports: MTTR and incident counts are heavy-tailed and unreliable as performance metrics; recommends learning-oriented measures. Warns against reading CE recurrence or RCA coverage as a team KPI."}]},{"k":"neumann-2011","a":"ai-operations","t":"Efficiently compiling efficient query plans for modern hardware","y":2011,"s":"peer_reviewed","e":7,"d":"10.14778/2002938.2002940","v":"Proceedings of the VLDB Endowment","a_":["Neumann"],"bl":"https://doi.org/10.14778/2002938.2002940"},{"k":"newey1987hac","a":"econometrics","t":"A Simple, Positive Semi-Definite, Heteroskedasticity and Autocorrelation Consistent Covariance Matrix","y":1987,"s":"peer_reviewed","e":2,"d":"10.2307/1913610","v":"Econometrica","a_":["Newey & West"],"bl":"https://doi.org/10.2307/1913610","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"HAC covariance used for the monthly ITS gap series (maxlags 2, small-sample t). With about 20 monthly points HAC SEs can be optimistic; show the CI and the minimum detectable effect rather than a significance verdict."}]},{"k":"ni2026genaialibaba","a":"foundations-stats","t":"Generative AI in Action: Field Experimental Evidence from Alibaba's Customer Service Operations","y":2026,"s":"preprint","e":4,"x":"2603.29888","v":"arXiv:2603.29888","a_":["Ni et al."],"bl":"https://arxiv.org/abs/2603.29888","pc":4,"sm":[{"m":"agent draft (claude)","l":"contrasting view","tx":"Common-good lens: randomized GenAI assistance at Alibaba sped work and raised ratings but did not improve objective quality, and top performers got worse."},{"m":"agent draft (claude)","l":"operational translation","tx":"Randomized access to a GenAI diagnosis/reply drafter at Alibaba: faster identification and chats, better customer ratings, but no effect on objective quality (customer retrials); top performers got worse. For Yugabyte: speed and satisfaction gains from AI notes need not show up in reopen/write-back;"},{"m":"agent draft (claude)","l":"support human ai","tx":"Randomized field experiment at Alibaba after-sales chat. Agents were given a GenAI assistant that drafts diagnoses and solutions, which they could adopt, modify or ignore. Speed and customer ratings improved on average, but there was no significant effect on objective quality (customer retrials). Lower-performing agents benefited most."}]},{"k":"nicol2004oncall","a":"foundations-other","t":"On-call work and health: a review","y":2004,"s":"peer_reviewed","e":2,"d":"10.1186/1476-069X-3-15","v":"Environmental Health","a_":["Nicol & Botterill"],"bl":"https://doi.org/10.1186/1476-069X-3-15","pc":2,"sm":[{"m":"agent draft (claude)","l":"support ops review","tx":"Review linking on-call work to sleep disruption, stress and health effects even when not called. Supports measuring on-call burden (night pages, pages per shift, interruptions per person-week at team level) as an agent-experience metric, especially for regions taking most out-of-hours pages."},{"m":"agent draft (claude)","l":"vp support dashboard","tx":"Review of health effects of on-call work: being on call, even without being called, is associated with disrupted sleep, stress and impaired wellbeing, though the evidence base was small and mostly cross-sectional."}]},{"k":"nistai100_1","a":"foundations-other","t":"Artificial Intelligence Risk Management Framework (AI RMF 1.0)","y":2023,"s":"external","e":2,"d":"10.6028/NIST.AI.100-1","v":"National Institute of Standards and Technology (U.S.)","a_":["Tabassi"],"bl":"https://doi.org/10.6028/NIST.AI.100-1","pc":1,"sm":[{"m":"agent draft (claude)","l":"iso42001 audit","tx":"NIST AI RMF 1.0 organizes AI risk management into GOVERN, MAP, MEASURE, MANAGE functions and defines trustworthiness characteristics (valid/reliable, safe, secure/resilient, accountable/transparent, explainable, privacy-enhanced, fair). Voluntary, not certifiable. Crosswalks cleanly to ISO 42001: GOVERN ~ clauses 5/7 and A.2-A.3; MAP ~ 6.1/8.2 and A.5;"}]},{"k":"noda-2023","a":"foundations-se","t":"DevEx: What Actually Drives Productivity","y":2023,"s":"peer_reviewed","e":1,"d":"10.1145/3595878","v":"Queue","a_":["Noda et al."],"bl":"https://doi.org/10.1145/3595878","pc":1,"sm":[{"m":"agent draft (claude)","l":"support human ai","tx":"DevEx proposes three core dimensions: feedback loops, cognitive load and flow state. Each is measured through perceptions (surveys) and workflows (system data), with KPIs at the top level."}]},{"k":"norman2010diagerror","a":"foundations-other","t":"Diagnostic error and clinical reasoning","y":2009,"s":"peer_reviewed","e":2,"d":"10.1111/j.1365-2923.2009.03507.x","v":"Medical Education","a_":["Norman & Eva"],"bl":"https://doi.org/10.1111/j.1365-2923.2009.03507.x","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Argues diagnostic errors stem mostly from knowledge deficits rather than from fast-thinking biases, and debiasing training has limited effect. Implication for LLM triage: retrieval of the right domain knowledge (known issues, mechanisms) matters more than adding generic 'consider alternatives' instructions."}]},{"k":"noufaily-2012","a":"foundations-stats","t":"An improved algorithm for outbreak detection in multiple surveillance systems","y":2012,"s":"peer_reviewed","e":2,"d":"10.1002/sim.5595","v":"Statistics in Medicine","a_":["Noufaily et al."],"bl":"https://doi.org/10.1002/sim.5595","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Improved Farrington: better dispersion handling, longer baselines, reweighting of past outbreaks; fewer false alarms across thousands of series. Relevant to one threshold per log template: multiplicity across about 4,000 ERROR/FATAL templates needs calibrated per-series baselines or an FDR step."}]},{"k":"nsvif2026instruction","a":"ai-operations","t":"Neuro-Symbolic Verification on Instruction Following of LLMs","y":2026,"s":"preprint","e":1,"x":"2601.17789","v":"arXiv:2601.17789","a_":["Su et al."],"bl":"https://arxiv.org/abs/2601.17789","pc":1,"sm":[{"m":"agent draft (gemini)","l":"verification","tx":"Introduces NSVIF (Su, Xu, Gao, Yang, Li, Yang, & Xu, 2026), a neuro-symbolic framework for verifying LLM instruction following by compiling constraints into deterministic logic checks paired with semantic predicates."}]},{"k":"nuñez-2018","a":"foundations-viz","t":"Optimizing colormaps with consideration for color vision deficiency to enable accurate interpretation of scientific data","y":2018,"s":"peer_reviewed","e":4,"d":"10.1371/journal.pone.0199239","v":"PLOS ONE","a_":["Nuñez et al."],"bl":"https://doi.org/10.1371/journal.pone.0199239"},{"k":"obermeyer-2019","a":"foundations-systems","t":"Dissecting racial bias in an algorithm used to manage the health of populations","y":2019,"s":"peer_reviewed","e":1,"d":"10.1126/science.aax2342","v":"Science","a_":["Obermeyer et al."],"bl":"https://doi.org/10.1126/science.aax2342","pc":1,"sm":[{"m":"agent draft (claude)","l":"limitations caveats","tx":"Cost-as-proxy-for-need algorithm under-referred Black patients. Warning for resource-based ticket-complexity scores: they encode past allocation (calls, escalations, senior engineers), not need."}]},{"k":"oliner2007supercomputers","a":"foundations-telemetry","t":"What Supercomputers Say: A Study of Five System Logs","y":2007,"s":"peer_reviewed","e":7,"d":"10.1109/DSN.2007.103","v":"37th Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN'07)","a_":["Oliner & Stearley"],"bl":"https://doi.org/10.1109/DSN.2007.103","pc":3,"sm":[{"m":"agent draft (claude)","l":"limitations caveats","tx":"Five supercomputer syslogs. Alerts dominated by a few categories; logs heavily redundant and must be filtered spatially and temporally; timestamps unreliable (missing years, inter-node drift, corrupted messages). Directly relevant to yearless glog/syslog parsing and to deduping repeated FATAL lines before counting crashes."},{"m":"agent draft (gemini)","l":"dataset benchmark","tx":"Establishes the classic five-supercomputer log corpus (BGL 4.75M lines / 708.8 MB, Thunderbird 211.2M lines / 29.60 GB, Spirit 272.3M lines / 37.8 GB, Liberty 265.6M lines / 31.5 GB, Red Storm) with domain-expert alert annotations."}]},{"k":"ong2025routellm","a":"ai-operations","t":"RouteLLM: Learning to Route LLMs with Preference Data","y":2025,"s":"external","e":1,"v":"Proceedings of the Thirteenth International Conference on Learning Representations (ICLR 2025)","a_":["Ong et al."],"pc":3,"sm":[{"m":"agent draft (gemini)","l":"local inference economics","tx":"Trains lightweight preference-calibrated routers (matrix factorization, BERT, causal LLM) between strong and weak models, achieving 85% cost reduction on MT-Bench at 95% GPT-4 quality. In our local Security-1/Security-2 pipeline, unclamped 2-option choice routing (tau=0.28) with symmetrized pairwise runoff serves as a domain-specific epistemic router."},{"m":"agent draft (qwen)","l":"operational translation","tx":"RouteLLM learns routers from preference data and reports order-of-magnitude cost savings routing between strong/weak pairs while retaining 95%+ quality — but on instruction-following mixes where the weak model is actually capable; hard-reasoning queries are where routers over-route."}]},{"k":"o’neil-1996","a":"distributed-systems","t":"The log-structured merge-tree (LSM-tree)","y":1996,"s":"peer_reviewed","e":15,"d":"10.1007/s002360050048","v":"Acta Informatica","a_":["O’Neil et al."],"bl":"https://doi.org/10.1007/s002360050048","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"LSM read cost = seeks across levels plus nexts over obsolete versions/tombstones. For YugabyteDB DocDB (RocksDB-derived) this explains why EXPLAIN (ANALYZE, DIST, DEBUG) rocksdb_number_db_seek/next and pg_stat_statements docdb_obsolete_rows_scanned diverge from rows returned after bulk deletes;"}]},{"k":"padilla-2021","a":"foundations-viz","t":"Uncertainty Visualization","y":2021,"s":"peer_reviewed","e":5,"d":"10.1002/9781118445112.stat08296","v":"Wiley StatsRef: Statistics Reference Online","a_":["Padilla et al."],"bl":"https://doi.org/10.1002/9781118445112.stat08296","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Padilla et al.: choose uncertainty displays for the decision; intervals and frequency framings beat point-only, and deterministic construal errors are common. In the gallery, evidence-grade chips and 'lower bound' wording help, but CI-free bars invite deterministic reading."}]},{"k":"pagano-2013","a":"foundations-se","t":"User feedback in the appstore: An empirical study","y":2013,"s":"peer_reviewed","e":5,"d":"10.1109/RE.2013.6636712","v":"2013 21st IEEE International Requirements Engineering Conference (RE)","a_":["Pagano & Maalej"],"bl":"https://doi.org/10.1109/RE.2013.6636712","pc":1,"sm":[{"m":"agent draft (claude)","l":"empirical findings","tx":"App-store feedback is frequent but skewed to few users and topics, mixes bug reports with feature requests, and needs classification before it can drive requirements. Analogy for tickets/bundles: dominated by a few heavy orgs and must be classified (bug vs gap vs how-to) before it feeds a roadmap."}]},{"k":"page-1954","a":"foundations-stats","t":"CONTINUOUS INSPECTION SCHEMES","y":1954,"s":"peer_reviewed","e":2,"d":"10.1093/biomet/41.1-2.100","v":"Biometrika","a_":["PAGE"],"bl":"https://doi.org/10.1093/biomet/41.1-2.100","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Introduces CUSUM: accumulate deviations from a target and alarm when the cumulative sum crosses a decision interval, trading average run length to false alarm against detection delay. Suits slow rises in per-release FATAL rates per 1,000 node-days (e.g. clock-skew loops), but only the glog plane has exposure denominators today."}]},{"k":"palkar2018weld","a":"foundations-systems","t":"Evaluating end-to-end optimization for data analytics applications in weld","y":2018,"s":"peer_reviewed","e":2,"d":"10.14778/3213880.3213890","v":"Proceedings of the VLDB Endowment","a_":["Palkar et al."],"bl":"https://doi.org/10.14778/3213880.3213890","pc":1,"sm":[{"m":"agent draft (gemini)","l":"PyO3 + Arrow C Data Interface FFI Architecture","tx":"Weld quantifies the 'data movement tax' when composing multi-library Python analytics pipelines (Pandas, NumPy, SciPy): even when individual C/C++/Rust kernels are fast, materializing intermediate objects and crossing FFI boundaries repeatedly degrades performance by up to 23x."}]},{"k":"panchenko-2019","a":"ai-operations","t":"BOLT: A Practical Binary Optimizer for Data Centers and Beyond","y":2019,"s":"peer_reviewed","e":2,"d":"10.1109/CGO.2019.8661201","v":"2019 IEEE/ACM International Symposium on Code Generation and Optimization (CGO)","a_":["Panchenko et al."],"bl":"https://doi.org/10.1109/CGO.2019.8661201","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"BOLT appends a second executable PT_LOAD segment and moves hot code, so symbolizers that assume one text segment (e.g. absl::Symbolize used by YB /pprof/heap_snapshot and yb_backend_heap_snapshot) can resolve frames ~4 KiB off on x86_64 BOLTed yb-server/postgres (GH yugabyte-db #35533)."},{"m":"agent draft (qwen)","l":"operational translation","tx":"In the yugabyte-db perf lab: our gate arms are built with yb_build.sh release at -O2 with no PGO/BOLT, while yb_build.sh exposes --pgo-instrument-type/--pgo-data-path/--enable-bolt and BOLT-class gains (up to ~20% on top of FDO+LTO) are the largest single untested lever in the program."}]},{"k":"panda2019iaso","a":"ai-operations","t":"IASO: A Fail-Slow Detection and Mitigation Framework for Distributed Storage Services","y":2019,"s":"peer_reviewed","e":4,"v":"2019 USENIX Annual Technical Conference (USENIX ATC 19)","a_":["Panda et al."],"bl":"https://www.usenix.org/conference/atc19/presentation/panda","pc":9,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"fsync T5-T7: IASO (Nutanix, ATC'19) detects fail-slow nodes from timeouts that PEERS observe: per peer pair, timeout/response ratio per 5 s epoch against a 10% threshold, additive score increase, decay only after minTTR = 2 min without timeouts (hysteresis), outlier detection over the N x (N-1) score table, one node quarantined at a time;"},{"m":"agent draft (claude)","l":"methodology","tx":"IASO scores peers' timeouts per 5 s epoch with decay and a 2-minute clean period before recovery: frequency-and-persistence-based. A good model for a chronic-slowness detector alongside a per-stall magnitude alert."},{"m":"agent draft (claude)","l":"operational translation","tx":"fsync POC: IASO's peer-timeout-based fail-slow detection maps to YugabyteDB peer signals (UpdateConsensus took Nms, request timed out, Fail or stepdown of leader) and motivates stall-aware graceful leader step-down with hysteresis over k8s/VM kill scripts."}]},{"k":"pandey-2015","a":"foundations-viz","t":"How Deceptive are Deceptive Visualizations?","y":2015,"s":"peer_reviewed","e":7,"d":"10.1145/2702123.2702608","v":"Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems","a_":["Pandey et al."],"bl":"https://doi.org/10.1145/2702123.2702608","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Controlled experiments showing truncated axes, area-as-quantity and aspect-ratio tricks measurably change reader judgments. Supports treating a silently out-of-range value (a 250% share, a point estimate outside its own interval) as a misleading-visualization defect, not a cosmetic one."}]},{"k":"pang2006devil","a":"ai-operations","t":"The devil and packet trace anonymization","y":2006,"s":"peer_reviewed","e":2,"d":"10.1145/1111322.1111330","v":"ACM SIGCOMM Computer Communication Review","a_":["Pang et al."],"bl":"https://doi.org/10.1145/1111322.1111330","pc":3,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Pang, Allman, Paxson, and Lee (SIGCOMM CCR 2006) detail the operational lessons of packet and log trace anonymization: multi-layer scrubbing must reconcile syntax-aware field parsing with defense-in-depth scrubbing, verify that truncation or malformed records do not leak unredacted suffixes, and emit explicit audit metadata describing what was transformed."}]},{"k":"panjer-2007","a":"foundations-telemetry","t":"Predicting Eclipse Bug Lifetimes","y":2007,"s":"peer_reviewed","e":8,"d":"10.1109/MSR.2007.25","v":"Fourth International Workshop on Mining Software Repositories (MSR'07:ICSE Workshops 2007)","a_":["Panjer"],"bl":"https://doi.org/10.1109/MSR.2007.25","pc":1,"sm":[{"m":"agent draft (claude-sonnet-5-5)","l":"complexity measurement without effort","tx":"Early MSR study predicting Eclipse bug lifetime with standard classifiers on Bugzilla fields; accuracy modest (about 34-31% for multi-class bins, as recalled). Shows lifetime is hard to predict from report attributes, i.e. elapsed time is only loosely coupled to anything observable at filing. Reinforces not using lifetime as the complexity label."}]},{"k":"pappalardo2019events","a":"sports-analytics","t":"A public data set of spatio-temporal match events in soccer competitions","y":2019,"s":"peer_reviewed","e":1,"d":"10.1038/s41597-019-0247-7","v":"Scientific Data","a_":["Pappalardo et al."],"bl":"https://doi.org/10.1038/s41597-019-0247-7","pc":3,"sm":[{"m":"vault note","l":"summary","tx":"Complete 2017-18 seasons of the five largest European leagues plus Euro 2016 and the 2018 World Cup, as spatio-temporal events. The files are CC BY 4.0, verified through the figshare API, which makes this the only openly licensed source of foul location, foul timing and player-level cards this project has found."}]},{"k":"parasuraman2010complacency","a":"foundations-se","t":"Complacency and Bias in Human Use of Automation: An Attentional Integration","y":2010,"s":"peer_reviewed","e":4,"d":"10.1177/0018720810376055","v":"Human Factors: The Journal of the Human Factors and Ergonomics Society","a_":["Parasuraman & Manzey"],"bl":"https://doi.org/10.1177/0018720810376055","pc":3,"sm":[{"m":"agent draft (claude)","l":"dashboard fix validation","tx":"Review integrating automation complacency and automation bias: humans over-rely on reliable automated aids, under-monitor them, and miss the aid's errors; expertise and training do not eliminate it. For an AI 'safe to close' list: a human click-through is not a sufficient guardrail;"},{"m":"agent draft (claude)","l":"support human ai","tx":"Integrative review: complacency and automation bias (omission and commission errors) stem from attention allocation. They occur in experts as well as novices, are not prevented by simple training or instructions, and grow under high workload and with highly reliable automation."},{"m":"agent draft (gemini)","l":"operational translation","tx":"Synthesizes decades of human factors research on automation complacency (attention allocation away from monitoring automated functions under multi-task load) and automation bias (omission errors when humans fail to notice automation misses, and commission errors when humans follow incorrect automated recommendations over contradicting raw evidence)."}]},{"k":"parnas1972criteria","a":"foundations-se","t":"On the criteria to be used in decomposing systems into modules","y":1972,"s":"peer_reviewed","e":7,"d":"10.1145/361598.361623","v":"Communications of the ACM","a_":["Parnas"],"bl":"https://doi.org/10.1145/361598.361623","pc":2,"sm":[{"m":"agent draft (gemini)","l":"software consolidation modularity","tx":"Parnas (1972) establishes information hiding as the foundational criterion for decomposing systems into modules: modules must encapsulate design decisions that are likely to change (such as representation formats, byte-scanning algorithms, or storage schemas) rather than steps in a processing flowchart."},{"m":"agent draft (gemini)","l":"systems architecture","tx":"Parnas (CACM 1972) establishes that modules must be decomposed around information hiding (encapsulating volatile design decisions and internal data representations behind stable interfaces) rather than pipeline execution steps."}]},{"k":"parnas1979extension","a":"foundations-se","t":"Designing Software for Ease of Extension and Contraction","y":1979,"s":"peer_reviewed","e":2,"d":"10.1109/TSE.1979.234169","v":"IEEE Transactions on Software Engineering","a_":["Parnas"],"bl":"https://doi.org/10.1109/TSE.1979.234169","pc":1,"sm":[{"m":"agent draft (gemini)","l":"software consolidation hierarchy","tx":"Parnas (1979) shows that designing software for ease of extension and contraction requires a strict acyclic 'uses' hierarchy where lower-level virtual machine layers have zero dependency on upper-level consumers."}]},{"k":"patel2015vibration","a":"foundations-stats","t":"Assessment of vibration of effects due to model specification can demonstrate the instability of observational associations","y":2015,"s":"peer_reviewed","e":2,"d":"10.1016/j.jclinepi.2015.05.029","v":"Journal of Clinical Epidemiology","a_":["Patel et al."],"bl":"https://doi.org/10.1016/j.jclinepi.2015.05.029","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Vibration of effects: re-estimate an association across all covariate subsets; a 'Janus effect' (sign flips across specs) marks instability. Used as the sign-stability criterion: tabletskew CE gap flips sign (-11 to +10 pp) across buckets and min-node choices."}]},{"k":"patel2024acorn","a":"ai-operations","t":"ACORN: Performant and Predicate-Agnostic Search Over Vector Embeddings and Structured Data","y":2024,"s":"peer_reviewed","e":1,"d":"10.1145/3654923","v":"Proceedings of the ACM on Management of Data (PACMMOD / SIGMOD)","a_":["Patel et al."],"bl":"https://doi.org/10.1145/3654923","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"ACORN: predicate-agnostic filtered HNSW traversing only predicate-passing nodes with expanded neighbor lists. pgvector instead post-filters plus iterative scans; low-selectivity filters remain a recall/latency risk to measure per selectivity."}]},{"k":"paul2022planencoders","a":"foundations-systems","t":"Database workload characterization with query plan encoders","y":2021,"s":"peer_reviewed","e":1,"d":"10.14778/3503585.3503600","v":"Proceedings of the VLDB Endowment","a_":["Paul et al."],"bl":"https://doi.org/10.14778/3503585.3503600"},{"k":"peerally-2016","a":"foundations-telemetry","t":"The problem with root cause analysis","y":2016,"s":"peer_reviewed","e":2,"d":"10.1136/bmjqs-2016-005511","v":"BMJ Quality &amp; Safety","a_":["Peerally et al."],"bl":"https://doi.org/10.1136/bmjqs-2016-005511","pc":1,"sm":[{"m":"agent draft (claude)","l":"limitations caveats","tx":"Peerally et al. list RCA failure modes: hindsight bias, the quest for a single root cause, weak recommendations (training, reminders) that are never implemented, and no tracking of whether actions reduced recurrence."}]},{"k":"pelkonen2015gorilla","a":"ai-operations","t":"Gorilla: A Fast, Scalable, In-Memory Time Series Database","y":2015,"s":"peer_reviewed","e":5,"d":"10.14778/2824032.2824078","v":"Proceedings of the VLDB Endowment","a_":["Pelkonen et al."],"bl":"https://doi.org/10.14778/2824032.2824078","pc":3,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Parquet dictionary encoding + DELTA_BINARY_PACKED timestamps + BYTE_STREAM_SPLIT doubles + ZSTD on the coordinator provides SQL-native rate queries without requiring a custom Gorilla decoder."}]},{"k":"pereira2017energy","a":"energy-grids","t":"Energy efficiency across programming languages: how do energy, time, and memory relate?","y":2017,"s":"peer_reviewed","e":1,"d":"10.1145/3136014.3136031","v":"Proceedings of the 10th ACM SIGPLAN International Conference on Software Language Engineering","a_":["Pereira et al."],"bl":"https://doi.org/10.1145/3136014.3136031","pc":2,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"Pereira et al. (SLE 2017) empirically evaluate 27 programming languages across the Computer Language Benchmarks Game using Intel RAPL energy sensors, execution time, and peak memory."}]},{"k":"pereira2021ranking","a":"energy-grids","t":"Ranking programming languages by energy efficiency","y":2021,"s":"peer_reviewed","e":2,"d":"10.1016/j.scico.2021.102609","v":"Science of Computer Programming","a_":["Pereira et al."],"bl":"https://doi.org/10.1016/j.scico.2021.102609","pc":2,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"Extended journal study (Science of Computer Programming 2021) analyzing the relationship between energy, execution time, and peak/DRAM power across 27 languages."}]},{"k":"pezeshkpour2024order","a":"foundations-other","t":"Large Language Models Sensitivity to The Order of Options in Multiple-Choice Questions","y":2024,"s":"peer_reviewed","e":2,"d":"10.18653/v1/2024.findings-naacl.130","v":"Findings of the Association for Computational Linguistics: NAACL 2024","a_":["Pezeshkpour & Hruschka"],"bl":"https://doi.org/10.18653/v1/2024.findings-naacl.130","pc":1,"sm":[{"m":"agent draft (gemini)","l":"report 07 option order sensitivity","tx":"Demonstrates that LLM accuracy in multiple-choice selection can swing by 13 to 75 percentage points simply from reordering candidate options, driven by positional and token-prior biases. Motivates position-debiased candidate scoring and cyclic/swap-order averaging in single-pass 4B decision gates."}]},{"k":"pham2025rcaeval","a":"ai-operations","t":"RCAEval: A Benchmark for Root Cause Analysis of Microservice Systems with Telemetry Data","y":2025,"s":"peer_reviewed","e":2,"d":"10.1145/3701716.3715290","v":"Companion Proceedings of the ACM on Web Conference 2025 (WWW Companion)","a_":["Pham et al."],"bl":"https://doi.org/10.1145/3701716.3715290","pc":4,"sm":[{"m":"agent draft (gemini)","l":"rca benchmark","tx":"Provides RCAEval (Zenodo 14590730, 5.16 GB compressed, CC-BY-4.0), comprising 735 fault-injection cases across Online Boutique, Sock Shop, and Train Ticket in three suites (RE1, RE2, RE3) with synchronized multi-service logs, metrics, and traces."}]},{"k":"phatak2021dirty","a":"sports-analytics","t":"The Dirty League: English Premier League Provides Higher Incentives for Fouling as Compared to Other European Soccer Leagues","y":2021,"s":"peer_reviewed","e":1,"d":"10.2478/hukin-2021-0095","v":"Journal of Human Kinetics","a_":["Phatak et al."],"bl":"https://doi.org/10.2478/hukin-2021-0095","pc":3,"sm":[{"m":"vault note","l":"summary","tx":"Uses fouls-to-yellow-cards across five leagues as an indicator of fouling incentive. Reports Simpson's paradox on this ratio: the pooled trend differs from within-league trends, so it must never be analysed pooled across leagues."}]},{"k":"pika2017resourceprofiles","a":"foundations-telemetry","t":"Mining Resource Profiles from Event Logs","y":2017,"s":"peer_reviewed","e":2,"d":"10.1145/3041218","v":"ACM Transactions on Management Information Systems","a_":["Pika et al."],"bl":"https://doi.org/10.1145/3041218","pc":1,"sm":[{"m":"agent draft (claude)","l":"effort estimation","tx":"Defines a set of resource-behaviour indicators (skills, utilisation, preferences, productivity, collaboration) computable from event logs and evaluates them on real logs."}]},{"k":"pimentel2019jupyterrepro","a":"foundations-se","t":"A Large-Scale Study About Quality and Reproducibility of Jupyter Notebooks","y":2019,"s":"peer_reviewed","e":1,"d":"10.1109/MSR.2019.00077","v":"2019 IEEE/ACM 16th International Conference on Mining Software Repositories (MSR)","a_":["Pimentel et al."],"bl":"https://doi.org/10.1109/MSR.2019.00077","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Of ~1.4M GitHub notebooks only ~24% ran without error and ~4% reproduced stored results; out-of-order execution, hidden state and unpinned dependencies dominate."}]},{"k":"ports2012ssipg","a":"distributed-systems","t":"Serializable Snapshot Isolation in PostgreSQL","y":2012,"s":"preprint","e":1,"x":"1208.4179","v":"arXiv:1208.4179","a_":["Ports & Grittner"],"bl":"https://arxiv.org/abs/1208.4179","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"PostgreSQL's SSI implementation (SIREAD locks, safe snapshots, read-only optimisations) and its measured false-positive abort rate. YSQL inherits PG error semantics (SQLSTATE 40001), so error surveillance must separate SSI/conflict aborts from read-restart errors, which YB raises under the same code."}]},{"k":"prentice1989surrogate","a":"foundations-stats","t":"Surrogate endpoints in clinical trials: Definition and operational criteria","y":1989,"s":"peer_reviewed","e":1,"d":"10.1002/sim.4780080407","v":"Statistics in Medicine","a_":["Prentice"],"bl":"https://doi.org/10.1002/sim.4780080407","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Operational translation: a surrogate (here the North Star 'evidence-ready in 24h') can stand in for the true outcome only if the treatment effect on the outcome is captured through the surrogate."}]},{"k":"psaraftis-2021","a":"maritime","t":"A comparative evaluation of market based measures for shipping decarbonization","y":2021,"s":"peer_reviewed","e":1,"d":"10.1016/j.martra.2021.100019","v":"Maritime Transport Research","a_":["Psaraftis et al."],"bl":"https://doi.org/10.1016/j.martra.2021.100019","pc":1,"sm":[{"m":"agent draft (gemini)","l":"maritime emissions telemetry","tx":"Psaraftis & Kontovas (2021) Decarbonization of maritime transport: Is there light at the end of the tunnel?"}]},{"k":"pullenayegum-2016","a":"foundations-stats","t":"Longitudinal data subject to irregular observation: A review of methods with a focus on visit processes, assumptions, and study design","y":2016,"s":"peer_reviewed","e":1,"d":"10.1177/0962280214536537","v":"Statistical Methods in Medical Research","a_":["Pullenayegum & Lim"],"bl":"https://doi.org/10.1177/0962280214536537","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Review of longitudinal data with irregular, possibly outcome-dependent observation times: ignorable, conditionally ignorable via inverse-intensity weighting, or informative needing joint models."}]},{"k":"putnick-2016","a":"foundations-other","t":"Measurement invariance conventions and reporting: The state of the art and future directions for psychological research","y":2016,"s":"peer_reviewed","e":2,"d":"10.1016/j.dr.2016.06.004","v":"Developmental Review","a_":["Putnick & Bornstein"],"bl":"https://doi.org/10.1016/j.dr.2016.06.004","pc":1,"sm":[{"m":"agent draft (claude-sonnet-5-5)","l":"complexity measurement without effort","tx":"Review of how measurement invariance is tested and reported in psychology: configural, metric, scalar steps, CFI change of 0.01 and RMSEA change of 0.015 as common conventions, and frequent under-reporting. Gives a concrete reporting checklist for checking that our complexity scale means the same thing across versions, eras, and customer segments."}]},{"k":"qiu2013schemacoevolution","a":"foundations-systems","t":"An empirical analysis of the co-evolution of schema and code in database applications","y":2013,"s":"peer_reviewed","e":10,"d":"10.1145/2491411.2491431","v":"Proceedings of the 2013 9th Joint Meeting on Foundations of Software Engineering","a_":["Qiu et al."],"bl":"https://doi.org/10.1145/2491411.2491431","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Large-scale empirical study of schema changes in ten popular open-source database applications and how application code co-changes with them. Builds a taxonomy of atomic schema change types (table/column add/drop/rename, type change, key/index/constraint changes) and measures frequency and the code impact."}]},{"k":"raasveldt2017hostage","a":"foundations-systems","t":"Don't hold my data hostage","y":2017,"s":"peer_reviewed","e":2,"d":"10.14778/3115404.3115408","v":"Proceedings of the VLDB Endowment","a_":["Raasveldt & Mühleisen"],"bl":"https://doi.org/10.14778/3115404.3115408","pc":1,"sm":[{"m":"agent draft (gemini)","l":"PyO3 + Arrow C Data Interface FFI Architecture","tx":"Raasveldt and Mühleisen prove empirically that traditional row/object-at-a-time serialization across client/language boundaries dominates analytical query latency by 10x–100x compared to columnar bulk vector transfer."}]},{"k":"raasveldt2019duckdb","a":"ai-operations","t":"DuckDB: An Embeddable Analytical Database","y":2019,"s":"peer_reviewed","e":9,"d":"10.1145/3299869.3320212","v":"Proceedings of the 2019 ACM SIGMOD International Conference on Management of Data","a_":["Raasveldt & Mühleisen"],"bl":"https://doi.org/10.1145/3299869.3320212","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Embedded in-process OLAP engine: vectorized execution over columnar vectors, MVCC, single-file storage. Per-operator timing reflects vector-at-a-time operators; host-language boundary (Python UDF, result conversion) is a first-class cost."},{"m":"agent draft (gemini)","l":"systems performance","tx":"By converting 896 MB Prometheus JSON exports into 2.1 MB ZSTD-compressed Parquet ordered by (metric_name, ts) with ROW_GROUP_SIZE=1,000,000, DuckDB exploits vectorized execution and row-group zone-map min/max pruning to execute all 14 deterministic health checks in 4-6 ms per query (~117 ms total)."}]},{"k":"rae2013f1","a":"ai-operations","t":"Online, Asynchronous Schema Change in F1","y":2013,"s":"peer_reviewed","e":3,"d":"10.14778/2536222.2536230","v":"Proceedings of the VLDB Endowment (PVLDB)","a_":["Rae et al."],"bl":"https://doi.org/10.14778/2536222.2536230"},{"k":"raemaekers2017semver","a":"foundations-se","t":"Semantic versioning and impact of breaking changes in the Maven repository","y":2017,"s":"peer_reviewed","e":1,"d":"10.1016/j.jss.2016.04.008","v":"Journal of Systems and Software","a_":["Raemaekers et al."],"bl":"https://doi.org/10.1016/j.jss.2016.04.008","pc":1,"sm":[{"m":"agent draft (gemini)","l":"api breaking change detection","tx":"Investigates SemVer adherence and binary/source breaking changes across >100,000 artifacts, motivating automated API/schema compatibility gates (cargo-semver-checks) and immutable version-tag enforcement in CI."}]},{"k":"raji2020closing","a":"security-governance","t":"Closing the AI accountability gap","y":2020,"s":"peer_reviewed","e":7,"d":"10.1145/3351095.3372873","v":"Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency","a_":["Raji et al."],"bl":"https://doi.org/10.1145/3351095.3372873","pc":4,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Proposes SMACTR, an end-to-end internal algorithmic audit framework (Scoping, Mapping, Artifact collection, Testing, Reflection) producing a documented audit trail with sign-off."},{"m":"agent draft (claude)","l":"ai platform review","tx":"SMACTR says accountability is a document trail from scoping to remediation. human_oversight is 'met' on the strength of a decision log entry, with no evidence that oversight operates (no override or edit records)."},{"m":"agent draft (claude)","l":"iso42001 audit","tx":"Proposes SMACTR, an end-to-end internal algorithmic audit framework (Scoping, Mapping, Artifact collection, Testing, Reflection) producing a document trail (ethical review, stakeholder map, model cards/datasheets, FMEA, risk analysis, remediation plan)."}]},{"k":"raju-2017","a":"foundations-systems","t":"PebblesDB","y":2017,"s":"peer_reviewed","e":8,"d":"10.1145/3132747.3132765","v":"Proceedings of the 26th Symposium on Operating Systems Principles","a_":["Raju et al."],"bl":"https://doi.org/10.1145/3132747.3132765","pc":1,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"PebblesDB introduces Fragmented Log-Structured Merge Trees (FLSM), using skip-list-inspired guards to partition SSTs across levels so data is never rewritten within the same level. Reduces write amplification by 2.4-3x and increases write throughput by 3-6.7x compared to RocksDB."}]},{"k":"rambachan2023honest","a":"foundations-stats","t":"A More Credible Approach to Parallel Trends","y":2023,"s":"peer_reviewed","e":4,"d":"10.1093/restud/rdad018","v":"Review of Economic Studies","a_":["Rambachan & Roth"],"bl":"https://doi.org/10.1093/restud/rdad018","sm":[{"m":"vault note","l":"summary","tx":"Sensitivity of DiD/event-study to bounded parallel-trend violations"}]},{"k":"ransohoff1978spectrum","a":"ai-operations","t":"Problems of Spectrum and Bias in Evaluating the Efficacy of Diagnostic Tests","y":1978,"s":"peer_reviewed","e":3,"d":"10.1056/NEJM197810262991705","v":"New England Journal of Medicine","a_":["Ransohoff & Feinstein"],"bl":"https://doi.org/10.1056/NEJM197810262991705","pc":2},{"k":"ravid-2019","a":"foundations-telemetry","t":"EPM 20/20: A Review, Framework, and Research Agenda for Electronic Performance Monitoring","y":2019,"s":"peer_reviewed","e":1,"d":"10.1177/0149206319869435","v":"Journal of Management","a_":["Ravid et al."],"bl":"https://doi.org/10.1177/0149206319869435"},{"k":"redbench2026","a":"foundations-telemetry","t":"Redbench: Workload Synthesis from Cloud Traces","y":2026,"s":"peer_reviewed","e":3,"d":"10.14778/3819518.3819538","v":"Proceedings of the VLDB Endowment","a_":["Wehrstein et al."],"bl":"https://doi.org/10.14778/3819518.3819538"},{"k":"redman1998dataquality","a":"foundations-other","t":"The impact of poor data quality on the typical enterprise","y":1998,"s":"peer_reviewed","e":2,"d":"10.1145/269012.269025","v":"Communications of the ACM","a_":["Redman"],"bl":"https://doi.org/10.1145/269012.269025","sm":[{"m":"vault note","l":"summary","tx":"Classic estimate of poor-data-quality cost; practitioner-level evidence (CACM viewpoint), not a controlled study"}]},{"k":"reis-1996","a":"foundations-other","t":"Simplified Therapeutic Intervention Scoring System","y":1996,"s":"peer_reviewed","e":1,"d":"10.1097/00003246-199601000-00012","v":"Critical Care Medicine","a_":["Miranda et al."],"bl":"https://doi.org/10.1097/00003246-199601000-00012"},{"k":"reise-2012","a":"foundations-stats","t":"The Rediscovery of Bifactor Measurement Models","y":2012,"s":"peer_reviewed","e":2,"d":"10.1080/00273171.2012.715555","v":"Multivariate Behavioral Research","a_":["Reise"],"bl":"https://doi.org/10.1080/00273171.2012.715555","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Reviews bifactor models: a general factor plus orthogonal group (source or method) factors. Omega-hierarchical tells how much of the score variance is general. For a multi-source effort latent, a general effort factor plus Slack-specific and Zendesk-specific method factors separates real work from source-specific traces."}]},{"k":"reitsma-2009","a":"foundations-stats","t":"A review of solutions for diagnostic accuracy studies with an imperfect or missing reference standard","y":2009,"s":"peer_reviewed","e":3,"d":"10.1016/j.jclinepi.2009.02.005","v":"Journal of Clinical Epidemiology","a_":["Reitsma et al."],"bl":"https://doi.org/10.1016/j.jclinepi.2009.02.005","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Reviews strategies for diagnostic-accuracy studies when the reference standard is imperfect or missing: correcting with known reference error rates, latent-class models, composite reference standards, panel or consensus diagnosis, and validating the reference itself."}]},{"k":"remil2021subgroupsql","a":"foundations-systems","t":"\"What makes my queries slow?\": Subgroup Discovery for SQL Workload Analysis","y":2021,"s":"peer_reviewed","e":4,"d":"10.1109/ase51524.2021.9678915","v":"2021 36th IEEE/ACM International Conference on Automated Software Engineering (ASE)","a_":["Remil et al."],"bl":"https://doi.org/10.1109/ase51524.2021.9678915"},{"k":"ren-2010","a":"ai-operations","t":"Google-Wide Profiling: A Continuous Profiling Infrastructure for Data Centers","y":2010,"s":"peer_reviewed","e":6,"d":"10.1109/MM.2010.68","v":"IEEE Micro","a_":["Ren et al."],"bl":"https://doi.org/10.1109/MM.2010.68","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"GWP establishes low-overhead, always-on sampling (perf/pprof style) as the default for fleet servers. For YB tserver/master, the analogue is short, rate-limited perf record -F 99 windows plus tcmalloc sampled heap profiles (YB sets 1 MiB sampling period), aggregated across nodes rather than one-off long captures on a single node."}]},{"k":"ridder1994down","a":"sports-analytics","t":"Down to Ten: Estimating the Effect of a Red Card in Soccer","y":1994,"s":"peer_reviewed","e":1,"d":"10.1080/01621459.1994.10476850","v":"Journal of the American Statistical Association","a_":["Ridder et al."],"bl":"https://doi.org/10.1080/01621459.1994.10476850"},{"k":"rizun-2021","a":"foundations-telemetry","t":"Assessing business process complexity based on textual data: Evidence from ITIL IT ticket processing","y":2021,"s":"peer_reviewed","e":1,"d":"10.1108/BPMJ-04-2021-0217","v":"Business Process Management Journal","a_":["Rizun et al."],"bl":"https://doi.org/10.1108/BPMJ-04-2021-0217"},{"k":"roberts1959ewma","a":"ai-operations","t":"Control Chart Tests Based on Geometric Moving Averages","y":1959,"s":"peer_reviewed","e":2,"d":"10.1080/00401706.1959.10489860","v":"Technometrics","a_":["Roberts"],"bl":"https://doi.org/10.1080/00401706.1959.10489860"},{"k":"rocher2019reidentification","a":"foundations-stats","t":"Estimating the success of re-identifications in incomplete datasets using generative models","y":2019,"s":"peer_reviewed","e":1,"d":"10.1038/s41467-019-10933-3","v":"Nature Communications","a_":["Rocher et al."],"bl":"https://doi.org/10.1038/s41467-019-10933-3","pc":1,"sm":[{"m":"agent draft (claude)","l":"limitations caveats","tx":"Shows with generative copula models that 15 demographic attributes re-identify 99.98% of Americans and that sampling/incompleteness gives weak protection."}]},{"k":"rodrigues2021clp","a":"ai-operations","t":"CLP: Efficient and Scalable Search on Compressed Text Logs","y":2021,"s":"peer_reviewed","e":4,"v":"15th USENIX Symposium on Operating Systems Design and Implementation (OSDI 21)","a_":["Rodrigues et al."],"bl":"https://www.usenix.org/conference/osdi21/presentation/rodrigues","pc":10,"sm":[{"m":"agent draft (gemini)","l":"columnar log storage","tx":"CLP (OSDI 2021) parses text logs into schema-less dictionary/non-dictionary variables for compressed grep search."},{"m":"agent draft (gemini)","l":"log template mining","tx":"Rodrigues et al. (OSDI 2021, CLP) introduce lossless semantic log compression and search by decomposing unstructured log messages on-the-fly into a static 'logtype' (template skeleton) and three variable streams: dictionary-encoded repetitive/compound strings (UUIDs, IPs, paths), compactly encoded pure integers, and custom-encoded floating-point…"},{"m":"agent draft (gemini)","l":"operational translation","tx":"CLP decomposes unstructured log lines into a three-way tuple: (1) timestamp delta stream, (2) static 'logtype' template ID, and (3) variable values split between a deduplicated dictionary (for multi-character categorical tokens like UUIDs, IPs, paths) and a raw encoded numeric stream."}]},{"k":"rodrigues2022tracesim","a":"foundations-telemetry","t":"TraceSim: An Alignment Method for Computing Stack Trace Similarity","y":2022,"s":"peer_reviewed","e":6,"d":"10.1007/s10664-021-10070-w","v":"Empirical Software Engineering","a_":["Rodrigues et al."],"bl":"https://doi.org/10.1007/s10664-021-10070-w","pc":2,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Rodrigues et al. (EMSE 2022) unify position-discounted stack depth weighting (from ReBucket) with inverse-document-frequency frame specificity (from Brodie et al. and CrashLocator) into a single global alignment algorithm (TraceSim) for stack trace deduplication."}]},{"k":"roth2023trending","a":"econometrics","t":"What’s trending in difference-in-differences? A synthesis of the recent econometrics literature","y":2023,"s":"peer_reviewed","e":6,"d":"10.1016/j.jeconom.2023.03.008","v":"Journal of Econometrics","a_":["Roth et al."],"bl":"https://doi.org/10.1016/j.jeconom.2023.03.008","sm":[{"m":"vault note","l":"summary","tx":"Review of modern DiD: staggered, heterogeneity, parallel-trends violations"}]},{"k":"rousseeuw1987silhouettes","a":"foundations-systems","t":"Silhouettes: A graphical aid to the interpretation and validation of cluster analysis","y":1987,"s":"peer_reviewed","e":2,"d":"10.1016/0377-0427(87)90125-7","v":"Journal of Computational and Applied Mathematics","a_":["Rousseeuw"],"bl":"https://doi.org/10.1016/0377-0427(87)90125-7"},{"k":"rubin1976missing","a":"ai-operations","t":"Inference and Missing Data","y":1976,"s":"peer_reviewed","e":7,"d":"10.1093/biomet/63.3.581","v":"Biometrika","a_":["Rubin"],"bl":"https://doi.org/10.1093/biomet/63.3.581"},{"k":"rule2020auditlogreview","a":"foundations-telemetry","t":"Using electronic health record audit logs to study clinical activity: a systematic review of aims, measures, and methods","y":2019,"s":"peer_reviewed","e":1,"d":"10.1093/jamia/ocz196","v":"Journal of the American Medical Informatics Association","a_":["Rule et al."],"bl":"https://doi.org/10.1093/jamia/ocz196","pc":1,"sm":[{"m":"agent draft (claude)","l":"measurement validity","tx":"Systematic review of studies that turn EHR audit logs into time-on-task. Its central methodological point for us: inactivity timeouts vary widely across studies and are often unreported, few studies validate against observation, and results are sensitive to the timeout choice."}]},{"k":"rust-1995","a":"foundations-other","t":"Return on Quality (ROQ): Making Service Quality Financially Accountable","y":1995,"s":"peer_reviewed","e":2,"d":"10.1177/002224299505900205","v":"Journal of Marketing","a_":["Rust et al."],"bl":"https://doi.org/10.1177/002224299505900205","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Return on Quality: quality spending must be justified through satisfaction and retention to revenue; many programmes fail that test. The gallery measures process quality (coverage, latency) but not the retention or revenue link ROQ requires."}]},{"k":"saadfalcon2024ares","a":"ai-operations","t":"ARES: An Automated Evaluation Framework for Retrieval-Augmented Generation Systems","y":2023,"s":"preprint","e":2,"x":"2311.09476","v":"arXiv:2311.09476","a_":["Saad-Falcon et al."],"bl":"https://arxiv.org/abs/2311.09476","pc":2,"sm":[{"m":"agent draft (claude)","l":"llmops","tx":"ARES trains lightweight judges for context relevance, faithfulness and answer relevance on synthetic data, then uses a few hundred human annotations with prediction-powered inference to produce confidence intervals on system scores."},{"m":"agent draft (claude)","l":"eval kpis","tx":"ARES trains lightweight judges (context relevance, answer faithfulness, answer relevance) on synthetic data and uses prediction-powered inference with a small human-annotated set (on the order of a few hundred points) to give confidence intervals on system-level scores."}]},{"k":"saerens2002prior","a":"ai-operations","t":"Adjusting the Outputs of a Classifier to New a Priori Probabilities: A Simple Procedure","y":2002,"s":"peer_reviewed","e":4,"d":"10.1162/089976602753284446","v":"Neural Computation","a_":["Saerens et al."],"bl":"https://doi.org/10.1162/089976602753284446","pc":2,"sm":[{"m":"agent draft (gemini)","l":"report 07 log odds prior shift","tx":"Derives exact Bayes' rule adjustment of classifier posterior probabilities under shifted class priors in log-odds space. Provides the mathematical foundation for keeping base-rate priors out of the LLM prompt and applying prior-shift correction additively in log-odds space."}]},{"k":"safford-2007","a":"foundations-other","t":"Patient Complexity: More Than Comorbidity. The Vector Model of Complexity","y":2007,"s":"peer_reviewed","e":2,"d":"10.1007/s11606-007-0307-0","v":"Journal of General Internal Medicine","a_":["Safford et al."],"bl":"https://doi.org/10.1007/s11606-007-0307-0","pc":1,"sm":[{"m":"agent draft (claude-sonnet-5-5)","l":"measurement","tx":"Vector Model of Complexity: patient complexity spans several axes (biologic, socioeconomic, cultural, behavioural, health-system) and is more than comorbidity count. Each axis has magnitude and direction. A later critique says the maths is only metaphorical and a single score may not help."}]},{"k":"saha-2022","a":"foundations-telemetry","t":"Mining root cause knowledge from cloud service incident investigations for AIOps","y":2022,"s":"peer_reviewed","e":1,"d":"10.1145/3510457.3513030","v":"Proceedings of the 44th International Conference on Software Engineering: Software Engineering in Practice","a_":["Saha & Hoi"],"bl":"https://doi.org/10.1145/3510457.3513030","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Salesforce ICSE-SEIP 2022 (ICA): mines past incident investigations into a root-cause knowledge graph and retrieves similar past incidents for a new one. Relevance: the similar-past-cases panel should match on structured symptoms (error templates, subsystem, cluster/customer) and resolution, not trigram subject text;"}]},{"k":"saisana-2005","a":"foundations-stats","t":"Uncertainty and Sensitivity Analysis Techniques as Tools for the Quality Assessment of Composite Indicators","y":2005,"s":"peer_reviewed","e":2,"d":"10.1111/j.1467-985X.2005.00350.x","v":"Journal of the Royal Statistical Society Series A: Statistics in Society","a_":["Saisana et al."],"bl":"https://doi.org/10.1111/j.1467-985X.2005.00350.x","pc":1,"sm":[{"m":"agent draft (claude)","l":"sensitivity analysis for scoring formulas","tx":"Propagates every construction choice (weights, normalisation, aggregation, imputation) through Monte Carlo, then uses variance-based sensitivity indices to attribute the resulting rank uncertainty to each choice. A composite is robust only if the ranks it reports stay stable across plausible choices."}]},{"k":"sambasivan2021cascades","a":"foundations-se","t":"“Everyone wants to do the model work, not the data work”: Data Cascades in High-Stakes AI","y":2021,"s":"peer_reviewed","e":3,"d":"10.1145/3411764.3445518","v":"Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems","a_":["Sambasivan et al."],"bl":"https://doi.org/10.1145/3411764.3445518","pc":3,"sm":[{"m":"agent draft (claude)","l":"ai platform review","tx":"Data cascades are delayed and opaque."},{"m":"agent draft (claude)","l":"data quality","tx":"Interview study (53 practitioners) showing data cascades: compounding, delayed, opaque downstream effects of upstream data issues, prevalent (92%) and triggered by undervalued data work."},{"m":"agent draft (gemini)","l":"operational translation","tx":"Empirical study of 53 AI practitioners demonstrating that 92% of AI projects suffer at least one Data Cascade (and 45.3% suffer two or more)—compounding, opaque, delayed downstream failures triggered by conventional AI/ML culture that valorizes 'model work' over 'data work'."}]},{"k":"sanders2019leastprivabac","a":"security-governance","t":"Mining least privilege attribute based access control policies","y":2019,"s":"peer_reviewed","e":1,"d":"10.1145/3359789.3359805","v":"Proceedings of the 35th Annual Computer Security Applications Conference","a_":["Sanders & Yue"],"bl":"https://doi.org/10.1145/3359789.3359805"},{"k":"sandve2013tenrules","a":"foundations-se","t":"Ten Simple Rules for Reproducible Computational Research","y":2013,"s":"peer_reviewed","e":1,"d":"10.1371/journal.pcbi.1003285","v":"PLoS Computational Biology","a_":["Sandve et al."],"bl":"https://doi.org/10.1371/journal.pcbi.1003285","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Track provenance of every result, no manual data steps, archive tool versions, record seeds, store data behind plots, link claims to results. Bundle Insights meets seed and package-version rules (wavec RNG seed, manifest) but not input pinning (row counts, not content hashes) nor the no-manual-step rule (fsync JSON)."}]},{"k":"sanyam-2024-arxiv","a":"ai-operations","t":"Large Language Models Must Be Taught to Know What They Don't Know","y":2024,"s":"preprint","e":1,"x":"2406.08391","v":"arXiv:2406.08391","a_":["Kapoor et al."],"bl":"https://arxiv.org/abs/2406.08391","pc":1,"sm":[{"m":"agent draft (claude)","l":"system one calibration","tx":"Kapoor et al. (NeurIPS 2024): prompting alone (verbalized or zero-shot logit) gives poor uncertainty estimates; fine-tuning a small LoRA/probe head on ~1,000 graded examples yields well-calibrated, generalizing uncertainty, cheaply."}]},{"k":"sarikaya2019dashboards","a":"foundations-viz","t":"What Do We Talk About When We Talk About Dashboards?","y":2019,"s":"peer_reviewed","e":1,"d":"10.1109/TVCG.2018.2864903","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Sarikaya et al."],"bl":"https://doi.org/10.1109/TVCG.2018.2864903","pc":1,"sm":[{"m":"agent draft (claude)","l":"dashboard design review","tx":"Surveys 83 dashboards and builds a typology by purpose (strategic, operational, organizational, learning), audience, update cadence and visual features. Main point: \"dashboard\" spans very different genres."}]},{"k":"satyanarayan-2017","a":"foundations-viz","t":"Vega-Lite: A Grammar of Interactive Graphics","y":2017,"s":"peer_reviewed","e":17,"d":"10.1109/TVCG.2016.2599030","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Satyanarayan et al."],"bl":"https://doi.org/10.1109/TVCG.2016.2599030","pc":1,"sm":[{"m":"agent draft (gemini)","l":"systems architecture","tx":"Defines a declarative JSON grammar of graphics that separates data series, encodings, scales, and threshold annotations from the rendering backend."}]},{"k":"schad-2010","a":"foundations-systems","t":"Runtime measurements in the cloud","y":2010,"s":"peer_reviewed","e":2,"d":"10.14778/1920841.1920902","v":"Proceedings of the VLDB Endowment","a_":["Schad et al."],"bl":"https://doi.org/10.14778/1920841.1920902","pc":1,"sm":[{"m":"agent draft (claude)","l":"empirical findings","tx":"Early EC2 measurement: CPU, memory, disk and network performance vary widely across instances and over time (coefficients of variation far above local clusters), and instance type/zone matter. Baseline evidence that cloud disk performance is not a fixed property of the instance type."}]},{"k":"schelter2018deequ","a":"foundations-systems","t":"Automating large-scale data quality verification","y":2018,"s":"peer_reviewed","e":6,"d":"10.14778/3229863.3229867","v":"Proceedings of the VLDB Endowment","a_":["Schelter et al."],"bl":"https://doi.org/10.14778/3229863.3229867","pc":4,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Deequ: declarative unit tests for data (completeness, uniqueness, consistency) computed incrementally at scale with anomaly detection on metric histories."},{"m":"agent draft (claude)","l":"ai platform review","tx":"Declarative data unit tests whose metrics are stored per batch and checked for anomalies. A Deequ-style dq_check_result table is the missing primitive."},{"m":"agent draft (claude)","l":"data quality","tx":"Deequ: declarative unit tests for data (completeness, uniqueness, size, ranges, referential constraints) computed as incremental metrics over batches, plus anomaly detection on the metric history (e.g. row-count deltas)."}]},{"k":"schipper-2019","a":"foundations-telemetry","t":"Tracing Back Log Data to its Log Statement: From Research to Practice","y":2019,"s":"peer_reviewed","e":6,"d":"10.1109/msr.2019.00081","v":"2019 IEEE/ACM 16th International Conference on Mining Software Repositories (MSR)","a_":["Schipper et al."],"bl":"https://doi.org/10.1109/msr.2019.00081","pc":1,"sm":[{"m":"agent draft (gemini)","l":"log template mining","tx":"Schipper, Aniche, and van Deursen (MSR 2019) evaluate static source-to-log tracing in enterprise Java systems (using Eclipse JDT AST analysis to convert SLF4J/Log4j statements into regular expressions) and document five real-world failure modes that break naive AST extractors: (1) wrapper logging methods and custom facades, (2) string concatenation…"}]},{"k":"schloss-2019","a":"foundations-viz","t":"Mapping Color to Meaning in Colormap Data Visualizations","y":2019,"s":"peer_reviewed","e":4,"d":"10.1109/TVCG.2018.2865147","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Schloss et al."],"bl":"https://doi.org/10.1109/TVCG.2018.2865147","pc":1,"sm":[{"m":"agent draft (claude)","l":"empirical findings","tx":"Background colour changes inferred colour-quantity mappings only when a colormap appears to vary in opacity: without apparent opacity variation people infer dark-is-more; with it, opaque-is-more. On light backgrounds the biases agree; on dark backgrounds they conflict, so the 'right' direction of a sequential ramp in dark mode is ambiguous."}]},{"k":"schroeder-2006","a":"foundations-systems","t":"A large-scale study of failures in high-performance computing systems","y":2006,"s":"peer_reviewed","e":2,"d":"10.1109/DSN.2006.5","v":"International Conference on Dependable Systems and Networks (DSN'06)","a_":["Schroeder & Gibson"],"bl":"https://doi.org/10.1109/DSN.2006.5","pc":1,"sm":[{"m":"agent draft (claude)","l":"empirical findings","tx":"DSN 2006 (trunk year field wrongly says 2026). Nine years of LANL HPC failure records; root causes split into hardware, software, network, environment, human, unknown. Failure rate scales with system size and is roughly per-node constant; time between failures is Weibull with decreasing hazard."}]},{"k":"sculley2015hiddendebt","a":"foundations-other","t":"Hidden Technical Debt in Machine Learning Systems","y":2015,"s":"external","e":3,"v":"Advances in Neural Information Processing Systems 28 (NeurIPS 2015)","a_":["Sculley et al."],"bl":"https://papers.nips.cc/paper/5656-hidden-technical-debt-in-machine-learning-systems","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Hidden debt in ML systems: entanglement (CACE), undeclared consumers, unstable data dependencies, glue code, configuration debt."},{"m":"agent draft (gemini)","l":"operational translation","tx":"Establishes the foundational taxonomy of ML system-level technical debt: boundary erosion, entanglement ('Changing Anything Changes Everything' / CACE principle), hidden feedback loops, undeclared consumers, unstable data dependencies, configuration debt, and pipeline jungles."}]},{"k":"segel2010narrative","a":"foundations-viz","t":"Narrative Visualization: Telling Stories with Data","y":2010,"s":"peer_reviewed","e":3,"d":"10.1109/TVCG.2010.179","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Segel & Heer"],"bl":"https://doi.org/10.1109/TVCG.2010.179","sm":[{"m":"vault note","l":"summary","tx":"gallery council 9 (dataviz/communication), 2026-10-28"}]},{"k":"segura2016metamorphicsurvey","a":"foundations-se","t":"A Survey on Metamorphic Testing","y":2016,"s":"peer_reviewed","e":3,"d":"10.1109/TSE.2016.2532875","v":"IEEE Transactions on Software Engineering","a_":["Segura et al."],"bl":"https://doi.org/10.1109/TSE.2016.2532875","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Survey of metamorphic testing: metamorphic relations replace a missing oracle by checking relations between outputs of related inputs."}]},{"k":"selinger-1979","a":"distributed-systems","t":"Access path selection in a relational database management system","y":1979,"s":"peer_reviewed","e":2,"d":"10.1145/582095.582099","v":"Proceedings of the 1979 ACM SIGMOD international conference on Management of data - SIGMOD '79","a_":["Selinger et al."],"bl":"https://doi.org/10.1145/582095.582099","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Origin of cost-based access path selection; YugabyteDB's CBO (yb_enable_cbo=on) extends it with network round-trip and LSM seek/next costs, so plan choice is still only as good as statistics, now with RPC count as a first-class cost."}]},{"k":"shah2026logjack","a":"ai-operations","t":"LogJack: Indirect Prompt Injection Through Cloud Logs Against LLM Debugging Agents","y":2026,"s":"preprint","e":2,"x":"2604.15368","v":"arXiv:2604.15368","a_":["Shah"],"bl":"https://arxiv.org/abs/2604.15368","pc":5,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"LogJack demonstrates that attacker-controlled strings in application/database logs can hijack LLM debugging agents via indirect prompt injection."},{"m":"agent draft (qwen)","l":"operational translation","tx":"LogJack's indirect-prompt-injection-through-logs attack class is the direct threat model for any log engine whose Parquet output is consumed by LLM diagnostic agents. Operational translation: (1) stamp an untrusted-source flag column into Parquet so downstream agent harnesses can wrap/sanitize log text;"},{"m":"agent draft (qwen)","l":"contrasting view","tx":"LogJack demonstrates indirect prompt injection through log lines themselves - the exact data plane ybc ingests from customer bundles. ybc's counter-design is architectural, not model-side: untrusted bundle text stays fenced inside an output jail (epistemic tags, plane_coverage/filter_semantics metadata, no tool-call affordance from log payloads)."}]},{"k":"shankar2022operationalizing","a":"foundations-other","t":"Operationalizing Machine Learning: An Interview Study","y":2022,"s":"preprint","e":1,"x":"2209.09125","v":"arXiv:2209.09125","a_":["Shankar et al."],"bl":"https://arxiv.org/abs/2209.09125","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"MLOps practice centres on velocity, validation and versioning; ad-hoc deploys are common."},{"m":"agent draft (claude)","l":"llmops","tx":"Interview study of 18 MLEs: MLOps success hinges on Velocity, Validation and Versioning (the three Vs). Practitioners favour simple, actionable monitors over many dashboards, suffer alert fatigue, and rely on staged rollouts plus human review of live data."}]},{"k":"shankar2023brokendata","a":"foundations-systems","t":"Moving Fast With Broken Data","y":2023,"s":"preprint","e":2,"x":"2303.06094","v":"arXiv:2303.06094","a_":["Shankar et al."],"bl":"https://arxiv.org/abs/2303.06094","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Shankar et al. argue that production data pipelines fail silently through stale, partial or rewritten data rather than crashes, and that teams need cheap, always-on checks on freshness, completeness and schema drift."},{"m":"agent draft (claude)","l":"sre platform","tx":"Studies data validation for ML pipelines that ingest data continuously. Static threshold checks produce too many false positives on noisy partitions. The paper proposes GATE, which compares partition summary statistics with history to flag corrupted batches."}]},{"k":"shankar2024validators","a":"ai-operations","t":"Who Validates the Validators? Aligning LLM-Assisted Evaluation of LLM Outputs with Human Preferences","y":2024,"s":"preprint","e":2,"x":"2404.12272","v":"arXiv:2404.12272","a_":["Shankar et al."],"bl":"https://arxiv.org/abs/2404.12272","pc":5,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"LLM-as-judge criteria drift once graders see outputs; evaluators must be aligned to a small human-labelled set and re-checked as criteria evolve."},{"m":"agent draft (claude)","l":"ai platform review","tx":"Evaluators must be selected by alignment with human grades, and criteria drift while grading."},{"m":"agent draft (claude)","l":"llmops","tx":"EvalGen: humans grade a small sample, candidate LLM/code assertions are selected by alignment with those grades. Key finding is 'criteria drift': graders refine their criteria while grading, so eval criteria cannot be fixed fully upfront."}]},{"k":"shao2020dbaccessantipatterns","a":"foundations-systems","t":"Database-Access Performance Antipatterns in Database-Backed Web Applications","y":2020,"s":"peer_reviewed","e":6,"d":"10.1109/ICSME46990.2020.00016","v":"2020 IEEE International Conference on Software Maintenance and Evolution (ICSME)","a_":["Shao et al."],"bl":"https://doi.org/10.1109/ICSME46990.2020.00016","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Taxonomy of database-access performance anti-patterns mined from real issues; distinguishes query-level from access-pattern-level problems. Useful as the target vocabulary for query_event.anti_patterns, which today covers only 6 text-regex query-level patterns."}]},{"k":"shen-2023","a":"ai-operations","t":"Propeller: A Profile Guided, Relinking Optimizer for Warehouse-Scale Applications","y":2023,"s":"peer_reviewed","e":5,"d":"10.1145/3575693.3575727","v":"Proceedings of the 28th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 2","a_":["Shen et al."],"bl":"https://doi.org/10.1145/3575693.3575727"},{"k":"shepperd-2012","a":"foundations-stats","t":"Evaluating prediction systems in software project estimation","y":2012,"s":"peer_reviewed","e":2,"d":"10.1016/j.infsof.2011.12.008","v":"Information and Software Technology","a_":["Shepperd & MacDonell"],"bl":"https://doi.org/10.1016/j.infsof.2011.12.008"},{"k":"shippee-2012","a":"foundations-stats","t":"Cumulative complexity: a functional, patient-centered model of patient complexity can improve research and practice","y":2012,"s":"peer_reviewed","e":2,"d":"10.1016/j.jclinepi.2012.05.005","v":"Journal of Clinical Epidemiology","a_":["Shippee et al."],"bl":"https://doi.org/10.1016/j.jclinepi.2012.05.005","pc":1,"sm":[{"m":"agent draft (claude-sonnet-5-5)","l":"construct separation","tx":"Cumulative Complexity Model: patient complexity is the balance between workload of demands (illness, treatment burden, life) and capacity (function, money, social support, literacy), with feedback loops. Complexity is defined from the demand/capacity imbalance, separate from the volume of care delivered."}]},{"k":"shneiderman1996eyes","a":"foundations-viz","t":"The eyes have it: a task by data type taxonomy for information visualizations","y":1996,"s":"peer_reviewed","e":7,"d":"10.1109/VL.1996.545307","v":"Proceedings 1996 IEEE Symposium on Visual Languages","a_":["Shneiderman"],"bl":"https://doi.org/10.1109/VL.1996.545307","sm":[{"m":"vault note","l":"summary","tx":"Visual Information-Seeking Mantra: Overview first, zoom and filter, then details-on-demand; foundational for dual-persona diagnostic interfaces."}]},{"k":"shypula2024pie","a":"foundations-se","t":"Learning Performance-Improving Code Edits","y":2023,"s":"preprint","e":1,"x":"2302.07867","v":"arXiv:2302.07867","a_":["Shypula et al."],"bl":"https://arxiv.org/abs/2302.07867","pc":2,"sm":[{"m":"agent draft (gemini)","l":"local inference economics","tx":"Introduces PIE (Performance-Improving Code Edits), showing that conditioning LLMs on execution/profiling feedback and sampling multiple candidate rewrites yields substantial speedups over human reference code, whereas single-shot unguided prompting often degrades performance or breaks correctness."}]},{"k":"siddiqui2022isum","a":"foundations-systems","t":"ISUM: Efficiently Compressing Large and Complex Workloads for Scalable Index Tuning","y":2022,"s":"peer_reviewed","e":2,"d":"10.1145/3514221.3526152","v":"Proceedings of the 2022 International Conference on Management of Data","a_":["Siddiqui et al."],"bl":"https://doi.org/10.1145/3514221.3526152"},{"k":"siegel-2022","a":"foundations-stats","t":"The impact of electronic monitoring on employees' job satisfaction, stress, performance, and counterproductive work behavior: A meta-analysis","y":2022,"s":"peer_reviewed","e":1,"d":"10.1016/j.chbr.2022.100227","v":"Computers in Human Behavior Reports","a_":["Siegel et al."],"bl":"https://doi.org/10.1016/j.chbr.2022.100227","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Meta-analysis (70 samples, 233 effects): electronic monitoring slightly lowers job satisfaction (r=-.10), raises stress (r=.11) and counterproductive behaviour (r=.09), with no performance gain (r=-.01); targets and feedback attached to monitoring worsen the effects."}]},{"k":"sigelman2010dapper","a":"ai-operations","t":"Dapper, a Large-Scale Distributed Systems Tracing Infrastructure","y":2010,"s":"preprint","e":1,"v":"Google, Inc.","a_":["Sigelman et al."],"bl":"https://research.google/pubs/dapper-a-large-scale-distributed-systems-tracing-infrastructure/","pc":5,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Dapper's sampled request tracing maps to YB's per-RPC traces (rpc_slow_query_threshold_ms default 10000 ms logs traces for slow calls) and ASH's root_request_id/rpc_request_id linkage across YSQL and DocDB; use them to tie a CPU profile hotspot to specific request paths instead of whole-process aggregates."},{"m":"agent draft (gemini)","l":"operational translation","tx":"Establishes the industry gold standard for low-overhead, self-monitoring production telemetry (<1% CPU overhead, adaptive sampling); operationalized in yb-collector's single-core GOMAXPROCS=1 / RAYON_NUM_THREADS=1, nice 19, ionice idle class, and built-in self_telemetry.json validation."}]},{"k":"silberzahn2018manyanalysts","a":"foundations-stats","t":"Many Analysts, One Data Set: Making Transparent How Variations in Analytic Choices Affect Results","y":2018,"s":"peer_reviewed","e":3,"d":"10.1177/2515245917747646","v":"Advances in Methods and Practices in Psychological Science","a_":["Silberzahn et al."],"bl":"https://doi.org/10.1177/2515245917747646","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"29 teams, one dataset, effect estimates from OR 0.89 to 2.93: analyst choices alone generate wide dispersion. Motivates reporting card headlines with their perturbation range, not one number."}]},{"k":"sim-2005","a":"foundations-stats","t":"The Kappa Statistic in Reliability Studies: Use, Interpretation, and Sample Size Requirements","y":2005,"s":"peer_reviewed","e":4,"d":"10.1093/ptj/85.3.257","v":"Physical Therapy","a_":["Sim & Wright"],"bl":"https://doi.org/10.1093/ptj/85.3.257","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Explains kappa's dependence on prevalence and bias, and gives sample-size tables for testing kappa against a null or estimating it to a given precision. Read across: with 61 doubly-labelled items and a binary label, a true kappa of 0.6 has a 95% interval of roughly 0.4-0.8 (our simulation)."}]},{"k":"simmons2011false","a":"foundations-stats","t":"False-Positive Psychology: Undisclosed Flexibility in Data Collection and Analysis Allows Presenting Anything as Significant","y":2011,"s":"peer_reviewed","e":6,"d":"10.1177/0956797611417632","v":"Psychological Science","a_":["Simmons et al."],"bl":"https://doi.org/10.1177/0956797611417632","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Researcher degrees of freedom: undisclosed flexibility in thresholds, exclusions and windows produces false positives. Threshold perturbations of +-25% show which gallery numbers are threshold-sensitive (crash-loop rate about -38% at >=5 starts per 60 min)."}]},{"k":"simonsohn-2020","a":"foundations-stats","t":"Specification curve analysis","y":2020,"s":"peer_reviewed","e":3,"d":"10.1038/s41562-020-0912-z","v":"Nature Human Behaviour","a_":["Simonsohn et al."],"bl":"https://doi.org/10.1038/s41562-020-0912-z","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Specification curve: rank estimates over all defensible specifications and test jointly against the null."}]},{"k":"sirin-2016","a":"foundations-systems","t":"Micro-architectural Analysis of In-memory OLTP","y":2016,"s":"peer_reviewed","e":4,"d":"10.1145/2882903.2882916","v":"Proceedings of the 2016 International Conference on Management of Data","a_":["Sirin et al."],"bl":"https://doi.org/10.1145/2882903.2882916","pc":1,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"Analyzes in-memory and disk-based OLTP engines using Intel Top-Down Microarchitecture Analysis Method (TMAM) and hardware PMU counters, showing that even when 100% of data fits in memory, OLTP engines spend >50-70% of execution cycles stalled (primarily L1-I/ITLB Front-End Bound and L3/DRAM Back-End Memory Bound pointer-chasing stalls) and retire only…"}]},{"k":"sirin-2017","a":"foundations-systems","t":"A methodology for OLTP micro-architectural analysis","y":2017,"s":"peer_reviewed","e":2,"d":"10.1145/3076113.3076116","v":"Proceedings of the 13th International Workshop on Data Management on New Hardware","a_":["Sirin et al."],"bl":"https://doi.org/10.1145/3076113.3076116","pc":1,"sm":[{"m":"agent draft (gemini)","l":"methodology","tx":"Demonstrates that standard Top-Down Microarchitecture Analysis (TMAM) underestimates memory stalls in pointer-chasing OLTP workloads by up to 50% due to out-of-order overlap accounting, and provides a corrected OLTP PMU stall decomposition."}]},{"k":"sirin-2021","a":"foundations-systems","t":"Micro-architectural analysis of in-memory OLTP: Revisited","y":2021,"s":"peer_reviewed","e":4,"d":"10.1007/s00778-021-00663-8","v":"The VLDB Journal","a_":["Sirin et al."],"bl":"https://doi.org/10.1007/s00778-021-00663-8","pc":1,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"Extended VLDBJ study revisiting OLTP microarchitectural behavior across modern multi-socket Xeon processors, demonstrating how L3/DRAM bandwidth saturation and cross-core coherency stalls govern multi-tenant scaling even when single-thread instruction count barely changes."}]},{"k":"skau-2016","a":"foundations-viz","t":"Arcs, Angles, or Areas: Individual Data Encodings in Pie and Donut Charts","y":2016,"s":"peer_reviewed","e":5,"d":"10.1111/cgf.12888","v":"Computer Graphics Forum","a_":["Skau & Kosara"],"bl":"https://doi.org/10.1111/cgf.12888"},{"k":"skoulis2015growingup","a":"foundations-systems","t":"Growing up with stability: How open-source relational databases evolve","y":2015,"s":"peer_reviewed","e":4,"d":"10.1016/j.is.2015.03.009","v":"Information Systems","a_":["Skoulis et al."],"bl":"https://doi.org/10.1016/j.is.2015.03.009"},{"k":"slagell2005sharing","a":"security-governance","t":"Sharing computer network logs for security and privacy: a motivation for new methodologies of anonymization","y":2005,"s":"peer_reviewed","e":2,"d":"10.1109/seccmw.2005.1588299","v":"Workshop of the 1st International Conference on Security and Privacy for Emerging Areas in Communication Networks, 2005.","a_":["Slagell & Yurcik"],"bl":"https://doi.org/10.1109/seccmw.2005.1588299","pc":2,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Slagell and Yurcik (SecureComm 2005) analyze multi-field log anonymization requirements for forensic sharing, highlighting the discrepancy between rule-firing counts (how many regex/parser rules matched across a file) and artifact-level redaction accounting (how many files and total replacements occurred), as well as the risk of unverified partial…"}]},{"k":"sliwerski2005szz","a":"foundations-se","t":"When do changes induce fixes?","y":2005,"s":"peer_reviewed","e":2,"d":"10.1145/1083142.1083147","v":"Proceedings of the 2005 international workshop on Mining software repositories - MSR '05","a_":["Śliwerski et al."],"bl":"https://doi.org/10.1145/1083142.1083147","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"SZZ: link fix commits to issues via ids in messages, trace fix-inducing changes. Basis for joining field-linked Jira keys to fix commits (fixtest); linkage only finds commits that cite the key."}]},{"k":"snell2024scaling","a":"ai-operations","t":"Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters","y":2024,"s":"preprint","e":1,"x":"2408.03314","v":"arXiv:2408.03314","a_":["Snell et al."],"bl":"https://arxiv.org/abs/2408.03314","pc":2,"sm":[{"m":"agent draft (gemini)","l":"local inference economics","tx":"Proves that scaling test-time compute (via verifier-guided beam/lookahead search or iterative revision) can outperform a 14x larger model at matched FLOPs on hard tasks."}]},{"k":"sors2021silence","a":"sports-analytics","t":"The Sound of Silence in Association Football: Home Advantage and Referee Bias Decrease in Matches Played without Spectators","y":2021,"s":"peer_reviewed","e":1,"d":"10.1080/17461391.2020.1845814","v":"European Journal of Sport Science","a_":["Sors et al."],"bl":"https://doi.org/10.1080/17461391.2020.1845814"},{"k":"souihli2021bullfrog","a":"foundations-systems","t":"BullFrog: Online Schema Evolution via Lazy Evaluation","y":2021,"s":"peer_reviewed","e":3,"d":"10.1145/3448016.3452842","v":"Proceedings of the 2021 International Conference on Management of Data","a_":["Bhattacherjee et al."],"bl":"https://doi.org/10.1145/3448016.3452842"},{"k":"sovrano2026mitigating","a":"ai-operations","t":"Mitigating Prompt-Induced Cognitive Biases in General-Purpose AI for Software Engineering","y":2026,"s":"peer_reviewed","e":1,"d":"10.1145/3808115","v":"Proceedings of the ACM on Software Engineering","a_":["Sovrano et al."],"bl":"https://doi.org/10.1145/3808115","pc":4,"sm":[{"m":"agent draft (gemini)","l":"empirical positioning","tx":"Introduces PROBE-SWE (FSE 2026) to test 8 prompt-induced cognitive biases (including anchoring and overconfidence) in SE decision support. Shows standard CoT and self-debiasing prompts fail to significantly reduce bias sensitivity, whereas axiomatic reasoning cues reduce bias sensitivity by 51%."},{"m":"agent draft (gemini)","l":"Preventing prompt induced cognitive bias & leading questions in customer side MCP","tx":"Demonstrates in FSE 2026 that suggestive or directional prompts induce severe confirmation bias, framing effects, and confabulation in software engineering AI assistants, and proves that neutral, open-ended, evidence-separated prompting eliminates prompt-induced bias."}]},{"k":"spiegelhalter2005funnel","a":"foundations-viz","t":"Funnel plots for comparing institutional performance","y":2004,"s":"peer_reviewed","e":6,"d":"10.1002/sim.1970","v":"Statistics in Medicine","a_":["Spiegelhalter"],"bl":"https://doi.org/10.1002/sim.1970","pc":5,"sm":[{"m":"agent draft (claude)","l":"limitations caveats","tx":"For error signatures vs escalation three cautions apply: use exact binomial limits at small n; adjust for overdispersion (multiplicative or random-effects) when many units fall outside limits, which signals systematic confounding (incident size, customer escalation propensity) rather than many true outliers;"},{"m":"agent draft (claude)","l":"operational translation","tx":"Funnel plots compare units with different n against a target with limits that narrow as n grows, avoiding league-table errors. Direct fit for per-customer P1 attainment with n 1-100+."},{"m":"agent draft (claude)","l":"dashboard design review","tx":"Funnel plots chart each unit's rate against its volume, with control limits that narrow as n grows. This avoids league tables that flag small units by chance. It is the right idiom for the \"P1 by customer\" and \"Response by shift\" tables, which now colour percentages for customers with one or two pages."}]},{"k":"spiegelhalter2005overdispersion","a":"foundations-stats","t":"Handling over-dispersion of performance indicators","y":2005,"s":"peer_reviewed","e":3,"d":"10.1136/qshc.2005.013755","v":"Quality and Safety in Health Care","a_":["Spiegelhalter"],"bl":"https://doi.org/10.1136/qshc.2005.013755","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Funnel-plot control limits built from the binomial alone assume the only between-unit variation is sampling noise. Real institutional (here: error-signature) data are over-dispersed, so too many units fall outside the limits."}]},{"k":"spiegelhalter2012regulation","a":"foundations-stats","t":"Statistical Methods for Healthcare Regulation: Rating, Screening and Surveillance","y":2011,"s":"peer_reviewed","e":6,"d":"10.1111/j.1467-985X.2011.01010.x","v":"Journal of the Royal Statistical Society Series A: Statistics in Society","a_":["Spiegelhalter et al."],"bl":"https://doi.org/10.1111/j.1467-985X.2011.01010.x","sm":[{"m":"vault note","l":"summary","tx":"gallery council 9 (dataviz/communication), 2026-10-28"}]},{"k":"srinivasan-2023","a":"foundations-viz","t":"Azimuth: Designing Accessible Dashboards for Screen Reader Users","y":2023,"s":"peer_reviewed","e":6,"d":"10.1145/3597638.3608405","v":"The 25th International ACM SIGACCESS Conference on Computers and Accessibility","a_":["Srinivasan et al."],"bl":"https://doi.org/10.1145/3597638.3608405","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Azimuth: dashboard-level screen-reader access needs an overview, consistent per-chart summaries and navigation between views. Supports short L1/L2 chart descriptions, named disclosures and a role filter that announces what changed."}]},{"k":"starmer-2014","a":"foundations-se","t":"Changes in Medical Errors after Implementation of a Handoff Program","y":2014,"s":"peer_reviewed","e":2,"d":"10.1056/NEJMsa1405556","v":"New England Journal of Medicine","a_":["Starmer et al."],"bl":"https://doi.org/10.1056/NEJMsa1405556","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Nine-hospital prospective study of the I-PASS structured handoff bundle (mnemonic, training, verbal + written template): medical errors fell 23% and preventable adverse events 30% with no measured increase in handoff time."}]},{"k":"steegen-2016","a":"foundations-stats","t":"Increasing Transparency Through a Multiverse Analysis","y":2016,"s":"peer_reviewed","e":4,"d":"10.1177/1745691616658637","v":"Perspectives on Psychological Science","a_":["Steegen et al."],"bl":"https://doi.org/10.1177/1745691616658637","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Multiverse analysis: enumerate every reasonable processing choice and report the whole distribution of results. Applied in the Bundle Insights chaos council by re-running 15 gallery headlines under case exclusions (year-bug, stale _derived, shared bundles, top-3 orgs, LOO-org, last 6 months), threshold and window shifts of +-25% and bootstrap seeds."}]},{"k":"stokes-2022","a":"foundations-viz","t":"Striking a Balance: Reader Takeaways and Preferences when Integrating Text and Charts","y":2022,"s":"peer_reviewed","e":3,"d":"10.1109/TVCG.2022.3209383","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Stokes et al."],"bl":"https://doi.org/10.1109/TVCG.2022.3209383","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Readers prefer charts with substantial integrated text; annotation placement drives takeaways. Supports short on-chart captions over long hover-only panel descriptions."}]},{"k":"strathern-1997","a":"foundations-other","t":"‘Improving ratings’: audit in the British University system","y":1997,"s":"peer_reviewed","e":2,"d":"10.1002/(sici)1234-981x(199707)5:3<305::aid-euro184>3.0.co;2-4","v":"European Review","a_":["Strathern"],"bl":"https://doi.org/10.1002/(sici)1234-981x(199707)5:3<305::aid-euro184>3.0.co;2-4","pc":1,"sm":[{"m":"agent draft (claude)","l":"support ops dashboard","tx":"Source of the popular formulation of Goodhart's law (\"when a measure becomes a target, it ceases to be a good measure\") in the context of audit culture. Cited for the risk that dashboard targets (backlog 100, CSAT 90, FRT 30 min) reshape behavior rather than reflect service quality."}]},{"k":"suissa2008immortal","a":"ai-operations","t":"Immortal Time Bias in Pharmacoepidemiology","y":2008,"s":"peer_reviewed","e":1,"d":"10.1093/aje/kwm324","v":"American Journal of Epidemiology","a_":["Suissa"],"bl":"https://doi.org/10.1093/aje/kwm324","pc":2},{"k":"sun2021event","a":"econometrics","t":"Estimating Dynamic Treatment Effects in Event Studies with Heterogeneous Treatment Effects","y":2021,"s":"peer_reviewed","e":6,"d":"10.1016/j.jeconom.2020.09.006","v":"Journal of Econometrics","a_":["Sun & Abraham"],"bl":"https://doi.org/10.1016/j.jeconom.2020.09.006","pc":2,"sm":[{"m":"agent draft (gemini)","l":"methodology","tx":"Proves that two-way fixed effects (TWFE) event-study regressions under staggered adoption contaminate lead and lag coefficients with heterogeneous treatment effects from other cohorts, and introduces an interaction-weighted estimator."}]},{"k":"swaminathan-rogers-1990","a":"foundations-stats","t":"Detecting Differential Item Functioning Using Logistic Regression Procedures","y":1990,"s":"peer_reviewed","e":2,"d":"10.1111/j.1745-3984.1990.tb00754.x","v":"Journal of Educational Measurement","a_":["Swaminathan & Rogers"],"bl":"https://doi.org/10.1111/j.1745-3984.1990.tb00754.x","pc":1,"sm":[{"m":"agent draft (claude-sonnet-5-5)","l":"complexity measurement without effort","tx":"Logistic regression DIF: tests whether an item response depends on group membership after conditioning on total score, separating uniform from non-uniform DIF. Apply to each rubric item with group = product/version/era/customer tier to catch items that signal complexity differently across groups."}]},{"k":"sweeney2002kanonymity","a":"security-governance","t":"k-ANONYMITY: A MODEL FOR PROTECTING PRIVACY","y":2002,"s":"peer_reviewed","e":5,"d":"10.1142/S0218488502001648","v":"International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems","a_":["SWEENEY"],"bl":"https://doi.org/10.1142/S0218488502001648","pc":1,"sm":[{"m":"agent draft (gemini)","l":"security and privacy","tx":"Defines k-anonymity for microdata release: every quasi-identifier equivalence class must contain at least k indistinguishable records to prevent record linkage re-identification."}]},{"k":"szafir-2018","a":"foundations-viz","t":"Modeling Color Difference for Visualization Design","y":2018,"s":"peer_reviewed","e":8,"d":"10.1109/TVCG.2017.2744359","v":"IEEE Transactions on Visualization and Computer Graphics","a_":["Szafir"],"bl":"https://doi.org/10.1109/TVCG.2017.2744359","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Crowdsourced color-difference thresholds measured for points, bars and lines show that perceptible difference depends strongly on mark type and size: thin lines and small points need much larger color distances than large fields. A palette validator that applies one fixed deltaE threshold regardless of mark (e.g."}]},{"k":"szlang2025snowflakeworkload","a":"foundations-systems","t":"Workload Insights from the Snowflake Data Cloud: What Do Production Analytic Queries Really Look Like?","y":2025,"s":"peer_reviewed","e":6,"d":"10.14778/3750601.3750632","v":"Proceedings of the VLDB Endowment","a_":["Szlang et al."],"bl":"https://doi.org/10.14778/3750601.3750632","pc":1,"sm":[{"m":"agent draft (claude)","l":"empirical findings","tx":"Snowflake analysis of 667M production BI-tool queries over one week (Oct 2024, one region). Statement mix: SELECT 47%, SHOW 31%, SESSION (parameter-setting) 16%, DDL 4%, DML 2%, transaction control 0.5%. Joins in 46% of queries, CTEs 25%, window functions 11%; 21% of statements exceed 10KB."}]},{"k":"taft2020cockroachdb","a":"distributed-systems","t":"CockroachDB: The Resilient Geo-Distributed SQL Database","y":2020,"s":"peer_reviewed","e":6,"d":"10.1145/3318464.3386134","v":"Proceedings of the 2020 ACM SIGMOD International Conference on Management of Data (SIGMOD '20)","a_":["Taft et al."],"bl":"https://doi.org/10.1145/3318464.3386134","pc":5,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"CockroachDB is the closest public comparator to YSQL; useful contrast: CRDB range-partitions by default whereas YSQL hash-shards the first PK column by default (yb_use_hash_splitting_by_default), which changes which predicates become multi-tablet fan-out scans."},{"m":"agent draft (gemini)","l":"operational translation","tx":"Describes CockroachDB's geo-distributed SQL architecture on Raft and LSM storage (RocksDB/Pebble), highlighting how parallel commits, read-refresh / transaction restart avoidance, and RPC batching minimize cross-node round-trips per SQL statement."}]},{"k":"taipalus2021rdbmserrors","a":"foundations-systems","t":"Error messages in relational database management systems: A comparison of effectiveness, usefulness, and user confidence","y":2021,"s":"peer_reviewed","e":2,"d":"10.1016/j.jss.2021.111034","v":"Journal of Systems and Software","a_":["Taipalus et al."],"bl":"https://doi.org/10.1016/j.jss.2021.111034"},{"k":"taipalus2023sqlerrorframework","a":"distributed-systems","t":"Framework for SQL Error Message Design: A Data-Driven Approach","y":2023,"s":"peer_reviewed","e":3,"d":"10.1145/3607180","v":"ACM Transactions on Software Engineering and Methodology","a_":["Taipalus & Grahn"],"bl":"https://doi.org/10.1145/3607180","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Data-driven framework of nine SQL error-message design guidelines, derived from 128 messages across eight DBMSs, 4,796 queries from 311 participants and content analysis of 1,505 user suggestions."}]},{"k":"taipalus2025enhancedsql","a":"foundations-se","t":"Enhanced SQL error messages facilitate faster error fixing","y":2025,"s":"peer_reviewed","e":2,"d":"10.1007/s10664-025-10695-1","v":"Empirical Software Engineering","a_":["Taipalus et al."],"bl":"https://doi.org/10.1007/s10664-025-10695-1","pc":1,"sm":[{"m":"agent draft (claude)","l":"empirical findings","tx":"Controlled experiment (EMSE 2025): SQL error messages rewritten per the 2023 design framework led to faster error fixing than stock DBMS messages. Strongest available evidence that rewriting messages (not only adding docs) reduces time-to-fix. Limits: novice population, single-query tasks;"}]},{"k":"takane-1987","a":"foundations-stats","t":"On the Relationship between iTem Response Theory and Factor Analysis of Discretized Variables","y":1987,"s":"peer_reviewed","e":2,"d":"10.1007/BF02294363","v":"Psychometrika","a_":["Takane & Leeuw"],"bl":"https://doi.org/10.1007/BF02294363","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Proves that the normal-ogive two-parameter IRT model (and the graded response model) is equivalent to a factor analysis of discretised variables with tetrachoric or polychoric correlations."}]},{"k":"tanabe-2004","a":"foundations-stats","t":"Reliability and Validity of Scores on the Emergency Severity Index Version 3","y":2004,"s":"peer_reviewed","e":2,"d":"10.1197/j.aem.2003.06.013","v":"Academic Emergency Medicine","a_":["Tanabe et al."],"bl":"https://doi.org/10.1197/j.aem.2003.06.013","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"ESI v3 reliability and validity; ESI lower-acuity levels are defined by the number of predicted resource types. Template for an ESI-style ex-post ticket-complexity grade (count of distinct resource types consumed) validated with weighted kappa."},{"m":"agent draft (claude-sonnet-5-5)","l":"construct separation","tx":"ESI v3 reliability and validity. ESI is two-step: acuity first (levels 1-2, threat to life or organ), then predicted resource types for levels 3-5. So one tool keeps acuity (impact-like) apart from expected resource need, and the resource part is a prediction at triage, not realised use. Our ex-ante effort predictor is the analogue;"}]},{"k":"tang2015holistic","a":"ai-operations","t":"Holistic Configuration Management at Facebook","y":2015,"s":"peer_reviewed","e":2,"d":"10.1145/2815400.2815401","v":"Proceedings of the 25th ACM Symposium on Operating Systems Principles (SOSP '15)","a_":["Tang et al."],"bl":"https://doi.org/10.1145/2815400.2815401","pc":5,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Facebook's Configerator/Gatekeeper treat config as code: review, automated validation, canaried staged rollout and fast rollback. Yardstick for YBA gflag edits, which apply universe-wide without a validator or canary; the gallery sees a gflag change in the observed week before 4.8% of bundled tickets but cannot say which changes were risky."},{"m":"agent draft (gemini)","l":"operational translation","tx":"Supports fleet-wide config drift and effective-default analysis in hagen-tools and Doctor over yb-collector bundles."}]},{"k":"tang2024minicheck","a":"ai-operations","t":"MiniCheck: Efficient Fact-Checking of LLMs on Grounding Documents","y":2024,"s":"peer_reviewed","e":2,"d":"10.18653/v1/2024.emnlp-main.499","v":"Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing","a_":["Tang et al."],"bl":"https://doi.org/10.18653/v1/2024.emnlp-main.499","pc":2,"sm":[{"m":"agent draft (gemini)","l":"systems performance","tx":"Demonstrates that compact specialized fact-checking models verifying atomic claims against grounding documents match or exceed frontier LLMs at a fraction of the cost."},{"m":"agent draft (gemini)","l":"methodology","tx":"Introduces MiniCheck (770M-7B), demonstrating that small specialized entailment verifiers reach GPT-4 factual grounding accuracy at orders-of-magnitude lower cost when checking sentence-level LLM claims against grounding documents (LLM-AggreFact)."}]},{"k":"tang2025cola","a":"foundations-telemetry","t":"CoLA: Model Collaboration for Log-Based Anomaly Detection","y":2025,"s":"peer_reviewed","e":3,"d":"10.14778/3749646.3749668","v":"Proceedings of the VLDB Endowment","a_":["Zhu et al."],"bl":"https://doi.org/10.14778/3749646.3749668","pc":1,"sm":[{"m":"agent draft (qwen)","l":"operational translation","tx":"CoLA (small local model + frontier collaboration for log anomaly detection): the small model handles routine events locally and routes uncertain/anomalous events to the LRM, preserving detection while cutting cost/latency."}]},{"k":"tawosi-2023","a":"foundations-stats","t":"Agile Effort Estimation: Have We Solved the Problem Yet? Insights From a Replication Study","y":2023,"s":"peer_reviewed","e":3,"d":"10.1109/TSE.2022.3228739","v":"IEEE Transactions on Software Engineering","a_":["Tawosi et al."],"bl":"https://doi.org/10.1109/TSE.2022.3228739","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Replicates Deep-SE story-point estimation on the original and new data with proper baselines and finds it is not better than simple baselines (median or mean of past points) on most projects; earlier gains came from weak baselines and evaluation choices."}]},{"k":"tene2015latency","a":"foundations-systems","t":"How NOT to Measure Latency","y":2015,"s":"external","e":1,"a_":["Tene"],"bl":"https://www.youtube.com/watch?v=lJ8ydIuPFeU","pc":1,"sm":[{"m":"agent draft (claude)","l":"limitations caveats","tx":"Coordinated omission: monitors that wait on a stalled system under-record the stall, turning one long stall into one slow sample. For fsync stalls this means per-request latency percentiles and 5-minute means hide long stalls; a stall must be measured as an interval (duration of the in-flight sync), not as sampled latencies."}]},{"k":"tezcan2014chat","a":"foundations-se","t":"Routing and Staffing in Customer Service Chat Systems with Impatient Customers","y":2014,"s":"peer_reviewed","e":3,"d":"10.1287/opre.2014.1284","v":"Operations Research","a_":["Tezcan & Zhang"],"bl":"https://doi.org/10.1287/opre.2014.1284","sm":[{"m":"vault note","l":"summary","tx":"Concurrent-session (chat) staffing and routing; closest queueing analogue to engineers working several async tickets at once"}]},{"k":"thompson1968nfa","a":"ai-operations","t":"Programming Techniques: Regular expression search algorithm","y":1968,"s":"peer_reviewed","e":3,"d":"10.1145/363347.363387","v":"Communications of the ACM","a_":["Thompson"],"bl":"https://doi.org/10.1145/363347.363387","pc":2,"sm":[{"m":"agent draft (gemini)","l":"methodology","tx":"Ken Thompson's classic 1968 CACM paper introducing linear-time O(m*n) regular expression matching by compiling a regular expression of length m into a non-deterministic finite automaton (NFA) and stepping the active state set lockstep across each input character in O(m) work per byte, avoiding exponential backtracking."}]},{"k":"thomson-2012","a":"distributed-systems","t":"Calvin","y":2012,"s":"peer_reviewed","e":3,"d":"10.1145/2213836.2213838","v":"Proceedings of the 2012 ACM SIGMOD International Conference on Management of Data","a_":["Thomson et al."],"bl":"https://doi.org/10.1145/2213836.2213838","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Calvin avoids distributed commit cost by deterministic pre-ordering; contrasts with YugabyteDB wait-queue concurrency where contention surfaces as 40001/40P01 retries and ConflictResolution_WaitOnConflictingTxns ASH samples. Background for treating hot-row contention as a schema problem, not a tuning knob."}]},{"k":"tian-2016","a":"foundations-stats","t":"On the unreliability of bug severity data","y":2015,"s":"peer_reviewed","e":2,"d":"10.1007/s10664-015-9409-1","v":"Empirical Software Engineering","a_":["Tian et al."],"bl":"https://doi.org/10.1007/s10664-015-9409-1","pc":1,"sm":[{"m":"agent draft (claude-sonnet-5-5)","l":"reliability","tx":"About half of duplicate bug reports carry inconsistent human severity labels across OpenOffice, Mozilla and Eclipse, so severity labels are noisy ground truth. Implication: our impact score should use few, rule-based, checkable fields (priority, production, page flag) and report label noise, not rely on free human grading."}]},{"k":"tian2023justask","a":"calibration-conformal","t":"Just Ask for Calibration: Strategies for Eliciting Calibrated Confidence Scores from Language Models Fine-Tuned with Human Feedback","y":2023,"s":"peer_reviewed","e":2,"d":"10.18653/v1/2023.emnlp-main.330","v":"Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing","a_":["Tian et al."],"bl":"https://doi.org/10.18653/v1/2023.emnlp-main.330","pc":2},{"k":"tibshirani2019conformal","a":"calibration-conformal","t":"Conformal Prediction Under Covariate Shift","y":2019,"s":"external","e":1,"v":"Advances in Neural Information Processing Systems (NeurIPS 2019)","a_":["Tibshirani et al."],"pc":1},{"k":"tikv2026slowstore","a":"foundations-se","t":"Slow store detection and evict-slow-store-scheduler (TiKV/PD documentation)","y":2026,"s":"external","e":2,"a_":["PingCAP"],"bl":"https://docs.pingcap.com/tidb/stable/pd-control","pc":1,"sm":[{"m":"agent draft (claude)","l":"contrasting view","tx":"TiKV/PD slow-store detection uses a time-integrated slow score (1-100, slow at >=80) and evicts leaders from one slow store, recovering after 600-900 s. This is a persistence/volume-style trigger, the counterpart to CockroachDB's per-stall 20 s crash: it catches chronic slowness that never crosses a per-stall threshold."}]},{"k":"tory2023datavoice","a":"foundations-viz","t":"Finding Their Data Voice: Practices and Challenges of Dashboard Users","y":2023,"s":"peer_reviewed","e":2,"d":"10.1109/MCG.2021.3136545","v":"IEEE Computer Graphics and Applications","a_":["Tory et al."],"bl":"https://doi.org/10.1109/MCG.2021.3136545","pc":1,"sm":[{"m":"agent draft (claude)","l":"dashboard design review","tx":"Interview study of dashboard users: they rarely stop at the dashboard. They export, annotate and add context and narrative to communicate upward (\"data voice\"). Dashboards that lack explanation, context and story push users into side spreadsheets and slides. Supports adding narrative and annotation panels, and \"so-what\" text, on executive views."}]},{"k":"traini-2022","a":"ai-operations","t":"Towards effective assessment of steady state performance in Java software: are we there yet?","y":2022,"s":"peer_reviewed","e":3,"d":"10.1007/s10664-022-10247-x","v":"Empirical Software Engineering","a_":["Traini et al."],"bl":"https://doi.org/10.1007/s10664-022-10247-x","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Java microbenchmarks often never reach steady state; developer warmup settings are frequently wrong. For profiler agents: 'discard the first run' is not a warmup criterion; check iteration stability or report non-steady behavior."}]},{"k":"treynor-2017","a":"foundations-systems","t":"The calculus of service availability","y":2017,"s":"peer_reviewed","e":4,"d":"10.1145/3080202","v":"Communications of the ACM","a_":["Treynor et al."],"bl":"https://doi.org/10.1145/3080202","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Availability = f(MTBF, MTTR, impact fraction); for a fixed target, frequent short outages and rare long ones are interchangeable in the budget but not in user perception or dependency chains."}]},{"k":"tufte2001vdqi","a":"foundations-viz","t":"The Visual Display of Quantitative Information","y":2001,"s":"external","e":4,"v":"Graphics Press","a_":["Tufte"],"pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Tufte: maximise data-ink, use small multiples on shared scales, keep graphical integrity (proportional encodings, labelled units). Applied: gallery footers (30-35% of card height) and repeated Era/Evidence text are non-data ink at page level; shared-axis small multiples (bias, coverage, audit) are exemplary."}]},{"k":"turnbull1976censored","a":"foundations-stats","t":"The Empirical Distribution Function with Arbitrarily Grouped, Censored and Truncated Data","y":1976,"s":"peer_reviewed","e":1,"d":"10.1111/j.2517-6161.1976.tb01597.x","v":"Journal of the Royal Statistical Society Series B: Statistical Methodology","a_":["Turnbull"],"bl":"https://doi.org/10.1111/j.2517-6161.1976.tb01597.x","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"PM seat (council 5): NPMLE for interval-censored event times. Bundle snapshots bracket upgrades to a median 21 days, so time-to-adopt a release series is interval-censored; Turnbull is the right estimator for adoption curves on the fleet frame."}]},{"k":"turner-2021","a":"foundations-stats","t":"Comparison of six statistical methods for interrupted time series studies: empirical evaluation of 190 published series","y":2021,"s":"peer_reviewed","e":4,"d":"10.1186/s12874-021-01306-w","v":"BMC Medical Research Methodology","a_":["Turner et al."],"bl":"https://doi.org/10.1186/s12874-021-01306-w","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Re-analysed 190 published interrupted time series with six estimation methods: OLS, OLS with Newey-West, Prais-Winsten, REML, ARIMA and others. The choice of method changed the point estimates, the standard errors and often the statistical conclusion, especially in short series."}]},{"k":"tversky1974judgment","a":"foundations-stats","t":"Judgment under Uncertainty: Heuristics and Biases","y":1974,"s":"peer_reviewed","e":1,"d":"10.1126/science.185.4157.1124","v":"Science","a_":["Tversky & Kahneman"],"bl":"https://doi.org/10.1126/science.185.4157.1124","pc":1,"sm":[{"m":"agent draft (gemini)","l":"report 07 prior neglect and anchoring","tx":"Establishes representativeness (insensitivity to prior probabilities / base-rate neglect) and anchoring-and-adjustment in judgment under uncertainty."}]},{"k":"unkel-2011","a":"foundations-stats","t":"Statistical Methods for the Prospective Detection of Infectious Disease Outbreaks: A Review","y":2011,"s":"peer_reviewed","e":5,"d":"10.1111/j.1467-985X.2011.00714.x","v":"Journal of the Royal Statistical Society Series A: Statistics in Society","a_":["Unkel et al."],"bl":"https://doi.org/10.1111/j.1467-985X.2011.00714.x","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Review of prospective outbreak detection (regression, time series, SPC, multiple streams) with emphasis on false-alarm control and evaluation against known outbreaks. Supplies the evaluation frame used in the YB common-good backtest: known events, lead time, false-alarm rate, minimum signal size."}]},{"k":"ustun-2016","a":"foundations-other","t":"Supersparse linear integer models for optimized medical scoring systems","y":2015,"s":"peer_reviewed","e":1,"d":"10.1007/s10994-015-5528-6","v":"Machine Learning","a_":["Ustun & Rudin"],"bl":"https://doi.org/10.1007/s10994-015-5528-6","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"SLIM learns sparse linear models with small integer coefficients by integer programming, directly optimising accuracy and sparsity rather than rounding a fitted logistic model. Shows that such scoring systems can match black-box accuracy on many medical tasks while staying usable by hand."}]},{"k":"vaidya2007rolemining","a":"security-governance","t":"The role mining problem","y":2007,"s":"peer_reviewed","e":2,"d":"10.1145/1266840.1266870","v":"Proceedings of the 12th ACM symposium on Access control models and technologies","a_":["Vaidya et al."],"bl":"https://doi.org/10.1145/1266840.1266870"},{"k":"vakkari-1999-task","a":"foundations-other","t":"Task complexity, problem structure and information actions","y":1999,"s":"peer_reviewed","e":2,"d":"10.1016/s0306-4573(99)00028-x","v":"Information Processing &amp; Management","a_":["Vakkari"],"bl":"https://doi.org/10.1016/s0306-4573(99)00028-x","pc":1,"sm":[{"m":"agent draft (claude-sonnet-5-5)","l":"complexity measurement without effort","tx":"Links task complexity (a priori determinability of the task) to problem structure and to the information actions people take: more complex tasks need more, and more varied, information. Complexity is defined on the task side and effort-like actions are a consequence, not a definition."}]},{"k":"van-houwelingen-2007","a":"foundations-stats","t":"Dynamic Prediction by Landmarking in Event History Analysis","y":2006,"s":"peer_reviewed","e":1,"d":"10.1111/j.1467-9469.2006.00529.x","v":"Scandinavian Journal of Statistics","a_":["HOUWELINGEN"],"bl":"https://doi.org/10.1111/j.1467-9469.2006.00529.x","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Proposes landmarking for dynamic prediction: at each landmark time, fit a model on subjects still at risk using information available up to that time. Basis for our re-scoring design: at 24 h and 72 h the expected-effort estimate is refitted on still-open tickets with activity so far, raising AUC from about 0.72 to 0.82 and 0.85."}]},{"k":"vandenberg-2000","a":"foundations-se","t":"A Review and Synthesis of the Measurement Invariance Literature: Suggestions, Practices, and Recommendations for Organizational Research","y":2000,"s":"peer_reviewed","e":3,"d":"10.1177/109442810031002","v":"Organizational Research Methods","a_":["Vandenberg & Lance"],"bl":"https://doi.org/10.1177/109442810031002"},{"k":"vanderaalst2016processmining","a":"foundations-other","t":"Process Mining","y":2016,"s":"peer_reviewed","e":2,"d":"10.1007/978-3-662-49851-4","v":"Springer Berlin Heidelberg","a_":["Aalst"],"bl":"https://doi.org/10.1007/978-3-662-49851-4","pc":1,"sm":[{"m":"agent draft (claude)","l":"effort estimation","tx":"Standard reference for process mining. Its organizational perspective (who did what, handover-of-work and working-together social networks) and the distinction between waiting time and service time are the formal basis for reconstructing case-level resource involvement from an event log."}]},{"k":"vankempen2024green","a":"energy-grids","t":"It's Not Easy Being Green: On the Energy Efficiency of Programming Languages","y":2024,"s":"preprint","e":1,"x":"2410.05460","v":"arXiv:2410.05460","a_":["Kempen et al."],"bl":"https://arxiv.org/abs/2410.05460","pc":1,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"van Kempen, Kwon, Nguyen & Berger (arXiv:2410.05460, 2024) re-examine Pereira et al.'s (2017, 2021) language energy rankings, demonstrating that memory-safe compiled languages (Rust, C++) achieve 10x-70x wall-time and energy advantages over pure CPython bytecode loops, but when Python delegates hot loops to native C/C++/Rust kernels (such as DuckDB,…"}]},{"k":"vanrenen2024redshiftfleet","a":"foundations-systems","t":"Why TPC is Not Enough: An Analysis of the Amazon Redshift Fleet","y":2024,"s":"peer_reviewed","e":4,"d":"10.14778/3681954.3682031","v":"Proceedings of the VLDB Endowment","a_":["Renen et al."],"bl":"https://doi.org/10.14778/3681954.3682031","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"PM seat (council 5): Vendor fleet telemetry (Redshift) shows real workloads differ from benchmarks: precedent for installed-base usage telemetry (call-home) as the broad lens, with bundles as the deep, biased complement."}]},{"k":"vasconcelos-2023","a":"foundations-se","t":"Explanations Can Reduce Overreliance on AI Systems During Decision-Making","y":2023,"s":"peer_reviewed","e":3,"d":"10.1145/3579605","v":"Proceedings of the ACM on Human-Computer Interaction","a_":["Vasconcelos et al."],"bl":"https://doi.org/10.1145/3579605","pc":1,"sm":[{"m":"agent draft (claude)","l":"support human ai","tx":"Models over-reliance as a strategic cost-benefit choice. People verify AI output only when verification is cheap relative to the task cost and the stakes. Explanations that make verification easier, along with higher task difficulty or rewards, reduce over-reliance. 'Evidence-link clicked before accept' is a useful verification proxy."}]},{"k":"vasey-2022","a":"foundations-other","t":"Reporting guideline for the early-stage clinical evaluation of decision support systems driven by artificial intelligence: DECIDE-AI","y":2022,"s":"peer_reviewed","e":1,"d":"10.1038/s41591-022-01772-9","v":"Nature Medicine","a_":["Vasey et al."],"bl":"https://doi.org/10.1038/s41591-022-01772-9","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"DECIDE-AI: reporting guideline for early-stage live clinical evaluation of AI decision support; AI-specific items on user characteristics, AI system version, how output is presented, user exposure/modification of AI output, safety/errors and human factors."}]},{"k":"vassiliadis2017rigidity","a":"foundations-systems","t":"Gravitating to rigidity: Patterns of schema evolution – and its absence – in the lives of tables","y":2017,"s":"peer_reviewed","e":2,"d":"10.1016/j.is.2016.06.010","v":"Information Systems","a_":["Vassiliadis et al."],"bl":"https://doi.org/10.1016/j.is.2016.06.010","pc":1,"sm":[{"m":"agent draft (claude)","l":"empirical findings","tx":"Studies the lives of tables across open-source schema histories. Most tables change rarely or never after birth (rigidity); change concentrates in few tables and early lifetimes, and wide tables tend to survive longer."}]},{"k":"vidigal-2026-q1","a":"ai-operations","t":"Putting an Agent on the Incident Queue: What I Measured Before I Trusted It.","y":2026,"s":"retracted_claim","e":7,"au":1,"v":"Technical Research Report (`n=102` production incidents, 5 arms, `510` runs)","a_":["Vidigal"],"sp":"/ai-operations/01-agent-on-the-incident-queue/","bl":"/ai-operations/01-agent-on-the-incident-queue/","pc":2,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"Audits autonomous LLM incident triage across 368 support cases and 787 runs under ISO/IEC 42001 and 5259 controls, proving that self-reported verbalized confidence carries no discriminative signal (AUROC 0.460 / 0.518) and retracting an earlier n=30 confidence-inversion claim."}],"retractions":["Self-reported confidence was reported as inverted at n=30."]},{"k":"vidigal-2026-q2","a":"ai-operations","t":"Four Frontier Models on the Same Support Queue: What Separates Them, and What Does Not.","y":2026,"s":"preprint","e":2,"au":1,"v":"Technical Research Report (`n=89` stratified incidents, 4 arms, `544` judged runs)","a_":["Vidigal"],"sp":"/ai-operations/02-four-models-on-the-queue/","bl":"/ai-operations/02-four-models-on-the-queue/","pc":2,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"Compares four frontier language models on the same production B2B support queue, measuring where reasoning models separate from fast models and enforcing a local 29-ms monotonic privacy redaction firewall before egress."}]},{"k":"vidigal2026acl","a":"ai-operations","t":"From TextBlob to LLM Agents: Sentiment Model Selection for B2B Technical Support with CSAT Ground Truth","y":2026,"s":"peer_reviewed","e":5,"au":1,"d":"10.18653/v1/2026.acl-industry.121","v":"Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 6: Industry Track)","a_":["Vidigal"],"bl":"https://aclanthology.org/2026.acl-industry.121/","br":1,"pc":8,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"Agent 1 perspective: proves multi-task sentiment collapses to MCC -0.018 under 20:1 imbalance."},{"m":"agent draft (claude-opus-4-6)","l":"methodology","tx":"Agent 2 perspective: demonstrates the RLHF Alignment Tax where Claude Opus 4.6 loses to Haiku 4.5 (p=0.003)."}]},{"k":"vidigal2026gorust","a":"security-governance","t":"Systems Language Trade-Offs in Production Telemetry Collectors: Controlled Evaluation of Go and Rust Across Regex Redaction, Long-Distance Zstandard, and Parquet Generation","y":2026,"s":"draft","e":1,"au":1,"v":"Yugabyte Applied AI & Systems Research","a_":["Vidigal"],"pc":1,"sm":[{"m":"agent draft (gemini)","l":"serverless finops sre","tx":"Vidigal (2026) evaluates Go vs. Rust across regex redaction, long-distance Zstandard compression, and columnar Parquet generation for YugabyteDB telemetry collectors (`yb-collector`)."}]},{"k":"vidigal2026mcpfirewall","a":"security-governance","t":"Defense-in-Depth for Forensic MCP Agents: DuckDB AST Sandboxing and Context-Aware Runtime Egress Redaction","y":2026,"s":"draft","e":1,"au":1,"v":"Yugabyte Applied AI & Systems Research","a_":["Vidigal"],"pc":6,"sm":[{"m":"agent draft (gemini)","l":"security and privacy","tx":"Combines DuckDB SQL AST sandboxing with context-aware runtime egress redaction so diagnostic MCP agents cannot execute mutating queries or exfiltrate customer secrets."}]},{"k":"vidigal2026rulepacks","a":"distributed-systems","t":"Most Health Rules Do Not Know Their Fleet: Population-Calibrated Thresholds, Three-Layer Effective Defaults, and Stratified Epidemiology of 212 Database Diagnostic Rules","y":2026,"s":"draft","e":2,"au":1,"v":"Yugabyte, Inc.","a_":["Vidigal"],"pc":2,"sm":[{"m":"agent draft (gemini)","l":"methodology","tx":"Evaluates 212 database diagnostic health rules across production telemetry, replacing static vendor thresholds with population-calibrated quantiles and stratified prevalence priors."}]},{"k":"vidigal2026wtlformal","a":"distributed-systems","t":"Machine-Checked Log Compression: Zero-Axiom Lean 4 Verification and C-ABI Differential Testing for Streaming Log Parsers and Temporal RLE","y":2026,"s":"draft","e":4,"au":1,"v":"Yugabyte, Inc.","a_":["Vidigal"],"pc":3,"sm":[{"m":"agent draft (gemini)","l":"formal verification","tx":"Proves round-trip multiset conservation and temporal run-length encoding invariants for streaming log parsers in Lean 4 with zero axioms, paired with C-ABI differential fuzzing."}]},{"k":"vidigal2026wtllakehouse","a":"distributed-systems","t":"What The Log: Compile-Time AST Template Mining and Two-Tier Statistical Temporal Compression for Agent-Queryable Parquet Log Lakehouses","y":2026,"s":"draft","e":10,"au":1,"v":"Yugabyte, Inc.","a_":["Vidigal"],"pc":3,"sm":[{"m":"agent draft (gemini)","l":"compile time ast mining","tx":"Extracts 53,848 log templates directly from polyglot source-code ASTs at compile time and compresses consecutive log bursts via two-tier statistical temporal run-length encoding into agent-queryable Apache Parquet."}]},{"k":"vidigal2026wtlspeedup","a":"distributed-systems","t":"Pre-Registering the Speedup: Cost-Bounded, Multiset-Equivalent Performance Engineering of a Terabyte-Scale Log-to-Parquet Engine","y":2026,"s":"draft","e":1,"au":1,"v":"Yugabyte, Inc.","a_":["Vidigal"],"pc":3,"sm":[{"m":"agent draft (gemini)","l":"serverless finops sre","tx":"Shows how streaming gzip/tar ingestion, right-sized Parquet bloom filters (v0.4.20 cutting silver Parquet bytes by ~0.5x), and bounded categorical ArgStats (v0.4.21 reducing peak RSS from 16-22 GB to 2-6 GB on YBA application logs) prevent OOM kills and quadratic IntervalSet stalls on 72-96.73 GiB support bundles."}]},{"k":"vidigal2027pv2","a":"ai-operations","t":"When Account Concentration Masquerades as a Software Regression: De-Confounding Observational Cloud Fleet Telemetry and Prior-Weighted e-BH Release Canarying","y":null,"s":"draft","e":1,"au":1,"v":"Companion Technical Report / Submission to ACM SIGKDD 2027","a_":["Vidigal"],"pc":2,"sm":[{"m":"agent draft (gemini)","l":"methodology","tx":"De-confounds observational cloud database release telemetry by separating account-concentration workload shifts from software regressions using coarsened stratification and prior-weighted e-BH control."}]},{"k":"vikram2023pbt","a":"foundations-other","t":"Can Large Language Models Write Good Property-Based Tests?","y":2023,"s":"preprint","e":1,"x":"2307.04346","v":"arXiv:2307.04346","a_":["Vikram et al."],"bl":"https://arxiv.org/abs/2307.04346","pc":2,"sm":[{"m":"agent draft (gemini)","l":"local inference economics","tx":"Evaluates LLM synthesis of QuickCheck/Hypothesis property-based tests (custom input generators + invariant properties), showing that compiler/execution feedback loops dramatically increase generator validity and edge-case bug discovery."}]},{"k":"voynov-2026","a":"foundations-other","t":"Typst: A Modern Typesetting Engine for Science","y":2026,"s":"peer_reviewed","e":1,"d":"10.9781/ijimai.2026.2269","v":"International Journal of Interactive Multimedia and Artificial Intelligence","a_":["Voynov et al."],"bl":"https://doi.org/10.9781/ijimai.2026.2269","pc":1,"sm":[{"m":"agent draft (gemini)","l":"systems architecture","tx":"Evaluates Typst (written in Rust by Laurenz Mädje and Martin Haug) as a programmable, incremental typesetting engine. Unlike LaTeX (Turing-complete macro expansion intertwined with single-pass layout) or AsciiDoctor-PDF (Ruby Prawn interpreter), Typst separates pure functional scripting from constraint-based layout memoization (`comemo`), achieving…"}]},{"k":"w3c2024wcag22","a":"foundations-se","t":"Web Content Accessibility Guidelines (WCAG) 2.2","y":2024,"s":"external","e":2,"a_":["World Wide Web Consortium"],"bl":"https://www.w3.org/TR/WCAG22/","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"WCAG 2.2 AA audit of the gallery: text contrast passes in both themes (muted 4.8:1 on the light page); 1.4.11 graphic contrast passes for blue, vermilion and green, with deemph-on-sheet borderline at 3.3:1; 2.4.1 bypass is missing; 2.5.3 Label in Name fails (visible 'Theme: light' vs name 'Colour theme');"}]},{"k":"wagner2002segmented","a":"foundations-stats","t":"Segmented regression analysis of interrupted time series studies in medication use research","y":2002,"s":"peer_reviewed","e":3,"d":"10.1046/j.1365-2710.2002.00430.x","v":"Journal of Clinical Pharmacy and Therapeutics","a_":["Wagner et al."],"bl":"https://doi.org/10.1046/j.1365-2710.2002.00430.x","pc":2,"sm":[{"m":"agent draft (claude)","l":"limitations caveats","tx":"Segmented regression needs enough points per segment (commonly about 8-12) and autocorrelation handling; with very short post periods slope changes are unstable and level changes dominate. Justifies a level-only model labelled early and underpowered when only 3 post months exist."},{"m":"agent draft (claude)","l":"operational translation","tx":"Segmented regression ITS with level and slope changes at known dates and autocorrelation handling."}]},{"k":"wallace-2005","a":"foundations-stats","t":"A Staffing Algorithm for Call Centers with Skill-Based Routing","y":2005,"s":"peer_reviewed","e":4,"d":"10.1287/msom.1050.0086","v":"Manufacturing &amp; Service Operations Management","a_":["Wallace & Whitt"],"bl":"https://doi.org/10.1287/msom.1050.0086"},{"k":"wang1996dataquality","a":"foundations-other","t":"Beyond Accuracy: What Data Quality Means to Data Consumers","y":1996,"s":"peer_reviewed","e":4,"d":"10.1080/07421222.1996.11518099","v":"Journal of Management Information Systems","a_":["Wang & Strong"],"bl":"https://doi.org/10.1080/07421222.1996.11518099","pc":3,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Consumer-centred data-quality framework (intrinsic, contextual, representational, accessibility). In the lake-chaos tests the failures were contextual (timeliness: stale _derived; completeness: unpinned frames, dropped CE cases) more than intrinsic accuracy, which mostly held."},{"m":"agent draft (claude)","l":"data quality","tx":"Consumer-centric DQ framework: 15 dimensions in 4 categories (intrinsic, contextual, representational, accessibility); timeliness and completeness are contextual, i.e. judged against the task."},{"m":"agent draft (claude)","l":"vp support dashboard","tx":"Empirically derived data-quality framework from data consumers: intrinsic (accuracy, objectivity, believability, reputation), contextual (relevance, timeliness, completeness, appropriate amount), representational (interpretability, consistency) and accessibility."}]},{"k":"wang2019hyperscan","a":"ai-operations","t":"Hyperscan: A Fast Multi-pattern Regex Matcher for Modern CPUs","y":2019,"s":"peer_reviewed","e":2,"v":"16th USENIX Symposium on Networked Systems Design and Implementation (NSDI 19)","a_":["Wang et al."],"bl":"https://www.usenix.org/conference/nsdi19/presentation/wang-xiang","pc":3,"sm":[{"m":"agent draft (gemini)","l":"systems performance","tx":"Establishes the principle of literal-substring prefiltering (extracting mandatory literal tokens from regular expressions and scanning buffers via SIMD/fast substring search before invoking NFA/DFA automata)."},{"m":"agent draft (gemini)","l":"methodology","tx":"USENIX NSDI 2019 paper presenting Intel Hyperscan, introducing graph decomposition (splitting regexes into literal strings and smaller automata) and the Teddy SIMD multi-literal shift-or/pshufb prefilter algorithm."}]},{"k":"wang2021outage","a":"foundations-telemetry","t":"Fast Outage Analysis of Large-Scale Production Clouds with Service Correlation Mining","y":2021,"s":"peer_reviewed","e":2,"d":"10.1109/ICSE43902.2021.00085","v":"2021 IEEE/ACM 43rd International Conference on Software Engineering (ICSE)","a_":["Wang et al."],"bl":"https://doi.org/10.1109/ICSE43902.2021.00085","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Mines service correlations from historical incidents to group alerts into outages quickly in Azure. Linking/aggregation is a first-class step before RCA: without it one systemic fault is counted as many independent escalations, inflating both recurrence and per-org counts."}]},{"k":"wang2023selfconsistency","a":"ai-operations","t":"Self-Consistency Improves Chain of Thought Reasoning in Language Models","y":2023,"s":"external","e":2,"v":"Proceedings of the Eleventh International Conference on Learning Representations (ICLR 2023)","a_":["Wang et al."],"pc":1,"sm":[{"m":"agent draft (qwen)","l":"operational translation","tx":"Self-consistency gains come from *diverse* sampled reasoning paths; its benefit shrinks when samples share biases."}]},{"k":"wang2023trustcodegen","a":"foundations-systems","t":"Investigating and Designing for Trust in AI-powered Code Generation Tools","y":2023,"s":"preprint","e":1,"x":"2305.11248","v":"arXiv:2305.11248","a_":["Wang et al."],"bl":"https://arxiv.org/abs/2305.11248","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Microsoft interview/design study: developers calibrate trust in AI code tools through community evidence (peers' experiences) and in-tool cues of reliability."}]},{"k":"wang2024humanllmannot","a":"ai-operations","t":"Human-LLM Collaborative Annotation Through Effective Verification of LLM Labels","y":2024,"s":"peer_reviewed","e":6,"d":"10.1145/3613904.3641960","v":"Proceedings of the CHI Conference on Human Factors in Computing Systems","a_":["Wang et al."],"bl":"https://doi.org/10.1145/3613904.3641960","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"CHI 2024 human-LLM collaborative annotation: LLM labels plus self-explanations shown to verifiers raise accuracy when the LLM is right but induce over-reliance when it is wrong; recommends verification only where the LLM is uncertain. For audit-grade reference labels, humans should label blind to LLM output."}]},{"k":"wei-1992","a":"foundations-stats","t":"The accelerated failure time model: A useful alternative to the cox regression model in survival analysis","y":1992,"s":"peer_reviewed","e":2,"d":"10.1002/sim.4780111409","v":"Statistics in Medicine","a_":["Wei"],"bl":"https://doi.org/10.1002/sim.4780111409"},{"k":"wei2021logreducer","a":"ai-operations","t":"On the Feasibility of Parser-based Log Compression in Large-Scale Cloud Systems","y":2021,"s":"peer_reviewed","e":3,"v":"19th USENIX Conference on File and Storage Technologies (FAST 21)","a_":["Wei et al."],"bl":"https://www.usenix.org/conference/fast21/presentation/wei","pc":7,"sm":[{"m":"agent draft (gemini)","l":"r6 stakeholder s02 cr05 wtl","tx":"Grounds what_the_log (wtl2 Wave-6e, 8d3b028) Schema v8 statistical RLE conservation law: because consecutive log events within 60s windows are folded into bursts with repetition_count >= 1 and 12-field args_summary moments/quantiles, downstream consumers (hagen-tools workload_signatures.py and doctor) must aggregate SUM(COALESCE(repetition_count, 1))…"},{"m":"agent draft (gemini)","l":"operational translation","tx":"LogReducer (FAST '21) demonstrates that after template extraction, timestamps account for >20% of compressed bytes in 8 of 18 log types and numeric variables carry strong intra-line and inter-line arithmetic correlations (constant deltas, strides, and correlated pairs)."},{"m":"agent draft (gemini)","l":"empirical findings","tx":"LogReducer shows that separating log templates from delta-encoded timestamps and correlated numeric variables beats general-purpose compressors, motivating ybc's coordinator-side header-parsed Parquet layer while retaining raw logs on the node."}]},{"k":"wei2023loggrep","a":"ai-operations","t":"LogGrep: Fast and Cheap Cloud Log Storage by Exploiting both Static and Runtime Patterns","y":2023,"s":"peer_reviewed","e":10,"d":"10.1145/3552326.3567484","v":"Proceedings of the Eighteenth European Conference on Computer Systems","a_":["Wei et al."],"bl":"https://doi.org/10.1145/3552326.3567484","pc":7,"sm":[{"m":"agent draft (gemini)","l":"systems performance","tx":"Wei et al. (EuroSys 2023) introduce LogGrep, exploiting static template decomposition and runtime variable patterns for fast log storage and querying without full decompression."},{"m":"agent draft (gemini)","l":"columnar log storage","tx":"LogGrep (EuroSys 2023) decomposes log lines into static templates and runtime variable capsules for fine-grained search over compressed cloud logs."},{"m":"agent draft (gemini)","l":"log template mining","tx":"Wei et al. (EuroSys 2023, LogGrep) observe that static log templates alone leave coarse wildcard slots ('<*>') whose runtime values still contain structured sub-patterns (e.g., composite 'IP:port', 'tablet_id/peer_id', or formatted status strings), and introduce fine-grained runtime sub-pattern decomposition ('capsules') paired with per-capsule min/max…"}]},{"k":"weiss-2007","a":"foundations-telemetry","t":"How Long Will It Take to Fix This Bug?","y":2007,"s":"peer_reviewed","e":3,"d":"10.1109/MSR.2007.13","v":"Fourth International Workshop on Mining Software Repositories (MSR'07:ICSE Workshops 2007)","a_":["Weiss et al."],"bl":"https://doi.org/10.1109/MSR.2007.13","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Predicts bug-fix effort from the most similar earlier reports; weak on novel reports. Reused in the triage audit as a TF-IDF nearest-closed-neighbour grade feature, with low maximum similarity as a novelty flag."}]},{"k":"weston-2025-arxiv","a":"ai-operations","t":"Lance: Efficient Random Access in Columnar Storage through Adaptive Structural Encodings","y":2025,"s":"preprint","e":2,"x":"2504.15247","v":"arXiv:2504.15247","a_":["Pace et al."],"bl":"https://arxiv.org/abs/2504.15247","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Lance v2 paper (arXiv 2504.15247): structural encoding (repetition/validity) dominates random-access cost on NVMe; reports correctly configured Parquet (small pages, page index, no/fast compression) can reach >60x better random access than defaults at the cost of scan speed and RAM."},{"m":"agent draft (gemini)","l":"operational translation","tx":"Lance provides superior random-access reads over Parquet via adaptive structural encodings, but lacks a mature pure-Go writer and universal zero-install reader support compared to Parquet in DuckDB/pyarrow/polars; ybc v1 therefore uses Parquet while keeping the catalog abstraction format-agnostic."}]},{"k":"won2018barrier","a":"ai-operations","t":"Barrier-Enabled IO Stack for Flash Storage","y":2018,"s":"peer_reviewed","e":1,"v":"16th USENIX Conference on File and Storage Technologies (FAST 18)","a_":["Won et al."],"bl":"https://www.usenix.org/conference/fast18/presentation/won","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Shows the Linux block layer enforces ordering and durability with Transfer-and-Flush (wait for DMA completion, then flush the device cache), which dominates fsync cost on flash; a barrier command removes it."}]},{"k":"wood-1986","a":"foundations-se","t":"Task complexity: Definition of the construct","y":1986,"s":"peer_reviewed","e":5,"d":"10.1016/0749-5978(86)90044-0","v":"Organizational Behavior and Human Decision Processes","a_":["Wood"],"bl":"https://doi.org/10.1016/0749-5978(86)90044-0","pc":1,"sm":[{"m":"agent draft (claude)","l":"methodology","tx":"Task complexity = component (distinct acts and information cues), coordinative (relations, timing, sequencing between acts) and dynamic (change in those relations over time). Triage-audit mapping: planes and artifacts (component); engineers, hand-offs, CE escalation (coordinative); novelty and new bugs (dynamic)."}]},{"k":"woodall-2006","a":"foundations-viz","t":"The Use of Control Charts in Health-Care and Public-Health Surveillance","y":2006,"s":"peer_reviewed","e":4,"d":"10.1080/00224065.2006.11918593","v":"Journal of Quality Technology","a_":["Woodall"],"bl":"https://doi.org/10.1080/00224065.2006.11918593","pc":1,"sm":[{"m":"agent draft (claude)","l":"metric statistics review","tx":"Review of control charts in health surveillance: distinguishes Phase I (establishing a stable baseline) from Phase II monitoring, warns about overdispersion, autocorrelation and the need for risk adjustment, and favours CUSUM/EWMA for small persistent shifts."}]},{"k":"wooldridge2023nonlineardid","a":"econometrics","t":"Simple approaches to nonlinear difference-in-differences with panel data","y":2023,"s":"peer_reviewed","e":4,"d":"10.1093/ectj/utad016","v":"The Econometrics Journal","a_":["Wooldridge"],"bl":"https://doi.org/10.1093/ectj/utad016","sm":[{"m":"vault note","l":"summary","tx":"Extended TWFE / Poisson for nonlinear DiD with staggered adoption"}]},{"k":"wright2003professional","a":"sports-analytics","t":"The Professional Foul in Football: Tactics and Deterrents","y":2003,"s":"peer_reviewed","e":2,"d":"10.1057/palgrave.jors.2601506","v":"Journal of the Operational Research Society","a_":["Wright & Hirotsu"],"bl":"https://doi.org/10.1057/palgrave.jors.2601506","pc":4,"sm":[{"m":"agent draft (gemini)","l":"empirical findings","tx":"Analyzes the tactical professional foul in association football, showing that when booking deterrence is weak, committing deliberate transition fouls is a rational expected-value strategy."}]},{"k":"wu2014crashlocator","a":"foundations-telemetry","t":"CrashLocator: locating crashing faults based on crash stacks","y":2014,"s":"peer_reviewed","e":5,"d":"10.1145/2610384.2610386","v":"Proceedings of the 2014 International Symposium on Software Testing and Analysis","a_":["Wu et al."],"bl":"https://doi.org/10.1145/2610384.2610386","pc":2,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Wu et al. (ISSTA 2014) propose CrashLocator, which ranks candidate faulty functions from crash stack traces using an inverse-frequency weighting scheme (analogous to TF-IDF: functions that appear frequently in a specific crash bucket's stacks but rarely across unrelated buckets receive high blame scores, weighted by inverse distance from the top of the…"}]},{"k":"wu2023changeinduced","a":"ai-operations","t":"An Empirical Study on Change-induced Incidents of Online Service Systems","y":2023,"s":"peer_reviewed","e":3,"d":"10.1109/ICSE-SEIP58684.2023.00027","v":"2023 IEEE/ACM 45th International Conference on Software Engineering: Software Engineering in Practice (ICSE-SEIP)","a_":["Wu et al."],"bl":"https://doi.org/10.1109/ICSE-SEIP58684.2023.00027","pc":3,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Change-induced incidents at a large online service provider: changes (code, config, deployment) are a leading incident trigger and linking incident to change under concurrent changes is the hard part."}]},{"k":"xia2018comprehension","a":"foundations-se","t":"Measuring Program Comprehension: A Large-Scale Field Study with Professionals","y":2018,"s":"peer_reviewed","e":2,"d":"10.1109/TSE.2017.2734091","v":"IEEE Transactions on Software Engineering","a_":["Xia et al."],"bl":"https://doi.org/10.1109/TSE.2017.2734091","pc":1,"sm":[{"m":"agent draft (claude)","l":"effort estimation","tx":"Interaction-logged field study of 78 professional developers over 3,148 working hours: 57.6% comprehension, 24.0% navigation, 13.4% other, only 5.0% editing. Sessions split at 1-hour idle; within-session gaps over 5 minutes excluded."}]},{"k":"xiong2024uncertainty","a":"calibration-conformal","t":"Can LLMs Express Their Uncertainty? An Empirical Evaluation of Confidence Elicitation in LLMs","y":2024,"s":"preprint","e":2,"x":"2306.13063","v":"Proceedings of the Twelfth International Conference on Learning Representations (ICLR 2024)","a_":["Xiong et al."],"bl":"https://arxiv.org/abs/2306.13063","pc":6,"sm":[{"m":"agent draft (gemini)","l":"report 07 verbalized overconfidence","tx":"Demonstrates empirically that verbalized confidence elicited in text from instruction-tuned LLMs clusters heavily at overconfident round numbers (80-100%) and fails to discriminate errors compared to white-box token log-probabilities and consistency aggregation."},{"m":"agent draft (qwen)","l":"operational translation","tx":"Verbalized confidence from instruct LLMs is systematically overconfident and poorly calibrated, especially without tuning; calibration method choice matters."}]},{"k":"xu-2025","a":"ai-operations","t":"Evolving the Cloud Block Store with Performance, Elasticity, Availability, and Hardware Offloading","y":2025,"s":"peer_reviewed","e":1,"d":"10.1145/3705925","v":"ACM Transactions on Storage","a_":["Xu et al."],"bl":"https://doi.org/10.1145/3705925","pc":1,"sm":[{"m":"agent draft (claude)","l":"empirical findings","tx":"Alibaba EBS evolution (TOS extension of the FAST'24 EBS paper): average latency is dominated by hardware/network, tail by software processing and background-task contention on block servers; availability events are localised by correlating many VMs. Guest-side the only symptom is I/O that does not complete."}]},{"k":"xu2002prefix","a":"security-governance","t":"Prefix-preserving IP address anonymization: measurement-based security evaluation and a new cryptography-based scheme","y":2002,"s":"peer_reviewed","e":5,"d":"10.1109/icnp.2002.1181415","v":"10th IEEE International Conference on Network Protocols, 2002. Proceedings.","a_":["Xu et al."],"bl":"https://doi.org/10.1109/icnp.2002.1181415","pc":3,"sm":[{"m":"agent draft (gemini)","l":"operational translation","tx":"Xu, Fan, Ammar, and Moon (ICNP 2002) introduce Crypto-PAn, proving that keyed cryptographic prefix-preserving pseudonymization preserves subnet/topology grouping (if two IPs share a $k$-bit prefix, their anonymized IPs share a $k$-bit prefix) while resisting frequency and known-plaintext attacks across traces."}]},{"k":"xu2015csur","a":"ai-operations","t":"Systems Approaches to Tackling Configuration Errors","y":2015,"s":"peer_reviewed","e":2,"d":"10.1145/2791577","v":"ACM Computing Surveys","a_":["Xu & Zhou"],"bl":"https://doi.org/10.1145/2791577","pc":2,"sm":[{"m":"agent draft (gemini)","l":"methodology","tx":"Grounds yb-collector's P0 priority placement of server.conf, varz.json, cmdline.txt, certs.json, and OS sysctl/ulimit/cgroup configuration."}]},{"k":"xu2015knobs","a":"ai-operations","t":"Hey, You Have Given Me Too Many Knobs!: Understanding and Dealing with Over-Designed Configuration in System Software","y":2015,"s":"peer_reviewed","e":1,"d":"10.1145/2786805.2786852","v":"Proceedings of the 2015 10th Joint Meeting on Foundations of Software Engineering (ESEC/FSE)","a_":["Xu et al."],"bl":"https://doi.org/10.1145/2786805.2786852","pc":4,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Large share of config parameters in mature systems are rarely or never set by users, and a minority of knobs carry most of the misconfiguration risk. For the Bundle Insights gflag cards this argues for ranking flags by hazard and deviation from default, not by how many orgs set them; the gfrisk card is the right direction but 'any set' conditions (e.g."},{"m":"agent draft (gemini)","l":"empirical findings","tx":"Motivates collecting both non-Default synthesized server.conf and complete effective varz.json + /varz HTML (for AutoFlags PROMOTED status) in yb-collector."}]},{"k":"xu2016pcheck","a":"ai-operations","t":"Early Detection of Configuration Errors to Reduce Failure Damage","y":2016,"s":"external","e":2,"v":"Proceedings of the 12th USENIX Symposium on Operating Systems Design and Implementation (OSDI '16)","a_":["Xu et al."],"pc":4,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Check config at startup so latent errors fail early. YBA analogue: warn at setup/health check when no default alert destination exists and when disk throughput exceeds the instance EBS baseline."},{"m":"agent draft (gemini)","l":"operational translation","tx":"Grounds yb-collector's collection of DMI instance identity (/sys/class/dmi/id/*), /proc/<pid>/limits, open FD/thread counts, and cgroup v1/v2 memory.max so Doctor can validate YB flags against actual container/VM capacity."}]},{"k":"xu2024divlog","a":"foundations-telemetry","t":"DivLog: Log Parsing with Prompt Enhanced In-Context Learning","y":2024,"s":"peer_reviewed","e":4,"d":"10.1145/3597503.3639155","v":"Proceedings of the IEEE/ACM 46th International Conference on Software Engineering","a_":["Xu et al."],"bl":"https://doi.org/10.1145/3597503.3639155","pc":1,"sm":[{"m":"agent draft (gemini)","l":"log template mining","tx":"Xu et al. (ICSE 2024, DivLog) apply in-context learning (ICL) with LLMs for log parsing by sampling a diverse candidate set via Determinantal Point Processes (DPP) and retrieving k-nearest neighbor demonstration examples per query log line."}]},{"k":"xu2024ragkgcs","a":"foundations-se","t":"Retrieval-Augmented Generation with Knowledge Graphs for Customer Service Question Answering","y":2024,"s":"preprint","e":2,"x":"2404.17723","v":"arXiv:2404.17723","a_":["Xu et al."],"bl":"https://arxiv.org/abs/2404.17723","sm":[{"m":"vault note","l":"summary","tx":"LinkedIn industry paper (SIGIR 2024 industry track); self-reported A/B, vendor-adjacent"}]},{"k":"yang2018webappperf","a":"foundations-systems","t":"How not to structure your database-backed web applications","y":2018,"s":"peer_reviewed","e":3,"d":"10.1145/3180155.3180194","v":"Proceedings of the 40th International Conference on Software Engineering","a_":["Yang et al."],"bl":"https://doi.org/10.1145/3180155.3180194","pc":2,"sm":[{"m":"agent draft (claude)","l":"empirical findings","tx":"Studies performance anti-patterns in 12 popular open-source Rails (ORM) applications, classifying causes such as inefficient queries, N+1 / redundant per-row queries, missing indexes, unnecessary data retrieval and view-layer issues, and shows that simple fixes yield large speedups."},{"m":"agent draft (claude)","l":"operational translation","tx":"Study of 12 Rails apps finds ORM-induced performance anti-patterns (unbounded queries, N+1, missing indexes, redundant data). Most detectors need the application or schema; from fingerprinted logs only a subset (unbounded result, select-star, N+1 via repetition bursts per transaction) is observable."}]},{"k":"yang2024sweagent","a":"ai-operations","t":"SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineering","y":2024,"s":"external","e":1,"v":"Advances in Neural Information Processing Systems (NeurIPS 2024)","a_":["Yang et al."],"pc":1,"sm":[{"m":"agent draft (qwen)","l":"operational translation","tx":"Agent-computer interface quality dominates raw model choice — a good interface lets a mid-size model outperform larger ones on the same task."}]},{"k":"yao2024taubench","a":"ai-operations","t":"τ-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains","y":2024,"s":"preprint","e":1,"x":"2406.12045","v":"arXiv:2406.12045","a_":["Yao et al."],"bl":"https://arxiv.org/abs/2406.12045","pc":2,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"tau-bench: simulated user plus tool-using agent under domain policy; success judged on final state, reliability by pass^k (all k trials succeed). GPT-4o under 50% pass^1 and under 25% pass^8 in retail."},{"m":"agent draft (claude)","l":"eval kpis","tx":"tau-bench simulates user-agent-tool conversations with domain policies and evaluates by comparing final database state to the goal state; introduces pass^k (probability all k i.i.d. trials succeed) to measure reliability, showing strong agents degrade sharply as k grows."}]},{"k":"yasin-2014","a":"ai-operations","t":"A Top-Down method for performance analysis and counters architecture","y":2014,"s":"peer_reviewed","e":6,"d":"10.1109/ISPASS.2014.6844459","v":"2014 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS)","a_":["Yasin"],"bl":"https://doi.org/10.1109/ISPASS.2014.6844459","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Profiling-methodology: Top-Down hierarchy (Frontend/Bad Spec/Backend[Mem,Core]/Retiring); drill into dominant node; exposed as perf stat --topdown / -M TopdownL1 and toplev; better than raw IPC thresholds."}]},{"k":"yin2011configerrors","a":"ai-operations","t":"An Empirical Study on Configuration Errors in Commercial and Open Source Systems","y":2011,"s":"peer_reviewed","e":7,"d":"10.1145/2043556.2043572","v":"Proceedings of the Twenty-Third ACM Symposium on Operating Systems Principles (SOSP)","a_":["Yin et al."],"bl":"https://doi.org/10.1145/2043556.2043572","pc":5,"sm":[{"m":"agent draft (claude)","l":"empirical findings","tx":"Empirical study of configuration errors in a commercial storage vendor's customer support cases and open-source systems: misconfigurations are a large share (about a quarter) of customer-reported issues and over-represented among high-severity cases."},{"m":"agent draft (claude)","l":"operational translation","tx":"Configuration errors are a large share of customer-reported issues, dominated by parameter value mistakes. For YB: platform_universe snapshots give named-gflag values per universe over time, but 75% of observed changes of a common flag are presence toggles (set/unset) rather than value edits, so drift must separate YBA-written defaults from operator…"},{"m":"agent draft (gemini)","l":"empirical findings","tx":"Directly motivates yb-collector's collection of both explicit server.conf and full /api/v1/varz + sysctl -a + process-specific cgroup limits so hagen-tools support_bundle.py (_gflag_drift_findings) and Doctor SysctlConfigExtractor detect cross-node drift."}]},{"k":"yiwei-2026-arxiv","a":"ai-operations","t":"Agentic AI and Human-in-the-Loop Interventions: Field Experimental Evidence from Alibaba's Customer Service Operations","y":2026,"s":"preprint","e":2,"x":"2605.14830","v":"arXiv:2605.14830","a_":["Wang et al."],"bl":"https://arxiv.org/abs/2605.14830","pc":2,"sm":[{"m":"agent draft (claude)","l":"Applied AI leader","tx":"Randomized field experiment at Alibaba Taobao: 647 service workers, about 680k chats, August 2024. In treatment, agentic AI handled eligible chats under human supervision. Chat duration fell; retrial rates moved little; ratings on AI-eligible chats fell substantially."},{"m":"agent draft (claude)","l":"operational translation","tx":"Alibaba RCT of agentic AI with human escalation: shorter chats, small retrial effect, lower ratings on AI-handled chats; human takeover works for technical escalations if early. Supports recording engineer override/use of AI notes and the timing of human review."}]},{"k":"yoon2016dbsherlock","a":"ai-operations","t":"DBSherlock: A Performance Diagnostic Tool for Transactional Databases","y":2016,"s":"peer_reviewed","e":2,"d":"10.1145/2882903.2915218","v":"Proceedings of the 2016 International Conference on Management of Data (SIGMOD '16)","a_":["Yoon et al."],"bl":"https://doi.org/10.1145/2882903.2915218","pc":4,"sm":[{"m":"agent draft (gemini)","l":"methodology","tx":"Supports yb-collector's multi-sample metrics, on-node sadf -x XML conversion, and vmstat UTC offset file so hagen-tools can align OS sar timelines with DB metrics in UTC."}]},{"k":"young2017modeluncertainty","a":"foundations-other","t":"Model Uncertainty and Robustness","y":2016,"s":"peer_reviewed","e":3,"d":"10.1177/0049124115610347","v":"Sociological Methods &amp; Research","a_":["Young & Holsteen"],"bl":"https://doi.org/10.1177/0049124115610347","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Model-robustness: report the share of models where the estimate is significant and the sign holds, plus the influence of each modelling choice."}]},{"k":"yu-2023","a":"foundations-telemetry","t":"Brain: Log Parsing With Bidirectional Parallel Tree","y":2023,"s":"peer_reviewed","e":9,"d":"10.1109/tsc.2023.3270566","v":"IEEE Transactions on Services Computing","a_":["Yu et al."],"bl":"https://doi.org/10.1109/tsc.2023.3270566","pc":2,"sm":[{"m":"agent draft (claude)","l":"log template mining","tx":"fsync T5: frequency-based longest-common-pattern parsing would also mask the S1 op phrase when it varies; catalog plus downstream typing is more stable across vintages."},{"m":"agent draft (gemini)","l":"log template mining","tx":"Yu et al. (IEEE TSC 2023, Brain) observe that log messages sharing a template often diverge at the prefix (e.g., when a variable or dynamic context appears early) while sharing a longest common word combination across the message, and construct a bidirectional parallel tree rooted at the most frequent co-occurring word combination that expands both…"}]},{"k":"yu2023nezha_fse","a":"foundations-telemetry","t":"Nezha: Interpretable Fine-Grained Root Causes Analysis for Microservices on Multi-modal Observability Data","y":2023,"s":"peer_reviewed","e":4,"d":"10.1145/3611643.3616249","v":"Proceedings of the 31st ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering","a_":["Yu et al."],"bl":"https://doi.org/10.1145/3611643.3616249","pc":1,"sm":[{"m":"agent draft (claude)","l":"applicability to support bundle fleet","tx":"Nezha converts heterogeneous observability data (log templates, metric alarms, trace spans) into a common event representation, mines event patterns, and localizes root causes by contrasting patterns between a fault-free and a faulty phase, giving interpretable code-region/resource-level answers without supervision."}]},{"k":"yu2024denum","a":"ai-operations","t":"Unlocking the Power of Numbers: Log Compression via Numeric Token Parsing","y":2024,"s":"peer_reviewed","e":2,"d":"10.1145/3691620.3695474","v":"Journal of the American Statistical Association","a_":["Yu et al."],"bl":"https://doi.org/10.1145/3691620.3695474","pc":4,"sm":[{"m":"agent draft (gemini)","l":"compression comparison","tx":"Demonstrates that numeric variables dominate the post-parsing entropy of system logs and introduces numeric token parsing and delta/elastic coding for lossless log compression (26.2 MB/s)."},{"m":"agent draft (gemini)","l":"empirical findings","tx":"Denum shows that numeric tokens (timestamps, offsets, IDs, latencies) dominate the entropy of operational logs after template extraction."}]},{"k":"yuan2010sherlog","a":"foundations-telemetry","t":"SherLog","y":2010,"s":"peer_reviewed","e":5,"d":"10.1145/1736020.1736038","v":"Proceedings of the fifteenth International Conference on Architectural support for programming languages and operating systems","a_":["Yuan et al."],"bl":"https://doi.org/10.1145/1736020.1736038","pc":3,"sm":[{"m":"agent draft (gemini)","l":"log template mining","tx":"Yuan et al. (ASPLOS 2010, SherLog) combine static source-code log template extraction with backward symbolic execution and constraint solving (SAT/SMT) to infer the exact control-flow paths and variable values that must have executed to produce a post-mortem runtime log sequence."},{"m":"agent draft (gemini)","l":"operational translation","tx":"Yuan et al. (ASPLOS 2010) present SherLog, which reconstructs post-mortem execution paths from runtime logs and source-level callsites without reproducing the failure."}]},{"k":"yuan2011logenhancer","a":"foundations-telemetry","t":"Improving software diagnosability via log enhancement","y":2011,"s":"peer_reviewed","e":5,"d":"10.1145/1950365.1950369","v":"Proceedings of the sixteenth international conference on Architectural support for programming languages and operating systems","a_":["Yuan et al."],"bl":"https://doi.org/10.1145/1950365.1950369","pc":3,"sm":[{"m":"agent draft (gemini)","l":"log template mining","tx":"Yuan et al. (ASPLOS 2011/2012, LogEnhancer) demonstrate that production log statements often omit critical causal variables along failure paths and use static data-flow slicing to automatically identify and append causally relevant in-scope variables to each log call site."},{"m":"agent draft (gemini)","l":"operational translation","tx":"Yuan et al. (ASPLOS 2011) demonstrate in LogEnhancer that post-mortem diagnosability depends on whether critical causal disambiguators (error codes, branch predicates, entity identifiers, timestamps) are captured at the point of failure."}]},{"k":"yuan2012characterizing","a":"ai-operations","t":"Characterizing logging practices in open-source software","y":2012,"s":"peer_reviewed","e":4,"d":"10.1109/icse.2012.6227202","v":"2012 34th International Conference on Software Engineering (ICSE)","a_":["Yuan et al."],"bl":"https://doi.org/10.1109/icse.2012.6227202","pc":6,"sm":[{"m":"agent draft (gemini)","l":"log template mining","tx":"Yuan et al. (ICSE 2012) empirically study logging practices and evolution across four large open-source C/C++ systems (Apache, OpenSSH, PostgreSQL, Squid), finding that logging is pervasive (~1 log line per 30 LOC), undergoes high churn (18% of all committed patches modify log statements, with 39% of modifications altering static text or format…"}]},{"k":"yuan2014simple","a":"ai-operations","t":"Simple Testing Can Prevent Most Critical Failures: An Analysis of Production Failures in Distributed Data-Intensive Systems","y":2014,"s":"peer_reviewed","e":3,"v":"11th USENIX Symposium on Operating Systems Design and Implementation (OSDI 14)","a_":["Yuan et al."],"bl":"https://www.usenix.org/conference/osdi14/technical-sessions/presentation/yuan","pc":3,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Of 198 user-reported failures in Cassandra/HBase/HDFS/MapReduce/Redis, most catastrophic ones came from incorrect handling of explicitly signalled non-fatal errors, and many reproduce on 3 nodes."}]},{"k":"zaharia2013dstreams","a":"foundations-systems","t":"Discretized streams","y":2013,"s":"peer_reviewed","e":2,"d":"10.1145/2517349.2522737","v":"Proceedings of the Twenty-Fourth ACM Symposium on Operating Systems Principles","a_":["Zaharia et al."],"bl":"https://doi.org/10.1145/2517349.2522737","pc":1,"sm":[{"m":"agent draft (gemini)","l":"micro batch vs columnar row group","tx":"Introduces Discretized Streams (D-Streams), modeling continuous computation as a series of deterministic, stateless micro-batch tasks over immutable RDDs. While operator-at-a-time streaming achieves sub-100ms latency, Parquet's PAX columnar format is physically a micro-batch format requiring 64k-256k row groups;"}]},{"k":"zeng2023columnar","a":"ai-operations","t":"An Empirical Evaluation of Columnar Storage Formats","y":2023,"s":"peer_reviewed","e":18,"d":"10.14778/3626292.3626298","v":"Proceedings of the VLDB Endowment","a_":["Zeng et al."],"bl":"https://doi.org/10.14778/3626292.3626298","pc":6,"sm":[{"m":"agent draft (gemini)","l":"systems performance","tx":"Zeng, Hui, Shen, Pavlo, McKinney & Zhang (PVLDB 2023) empirically evaluate Parquet and ORC columnar internals (row-group sizing, dictionary/RLE encoding, zone maps, and predicate pushdown)."},{"m":"agent draft (claude)","l":"operational translation","tx":"Benchmarks Parquet vs ORC across workloads; recommends dictionary encoding by default, favouring integer-encoding decode speed over ratio, optional block compression on fast storage, and finer-grained auxiliary structures (page-level indexes). Flags format inefficiencies for ML (wide tables) and GPU decoding."},{"m":"agent draft (gemini)","l":"operational translation","tx":"Zeng et al. (PVLDB '23) empirically evaluate Parquet and ORC across modern NVMe and cloud storage, showing that: (1) dictionary + RLE + general-purpose block compression (ZSTD/LZ4) shifts the bottleneck from I/O to CPU decoding, (2) LZ4 decompresses 3x-5x faster than ZSTD at the cost of ~1.4x-1.8x larger byte footprint, and (3) row-group sizing and…"}]},{"k":"zeng2023graphicalperception","a":"foundations-viz","t":"A Review and Collation of Graphical Perception Knowledge for Visualization Recommendation","y":2023,"s":"peer_reviewed","e":11,"d":"10.1145/3544548.3581349","v":"Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems","a_":["Zeng & Battle"],"bl":"https://doi.org/10.1145/3544548.3581349"},{"k":"zerouali2019techlagframework","a":"foundations-se","t":"A formal framework for measuring technical lag in component repositories — and its application to npm","y":2019,"s":"peer_reviewed","e":1,"d":"10.1002/smr.2157","v":"Journal of Software: Evolution and Process","a_":["Zerouali et al."],"bl":"https://doi.org/10.1002/smr.2157","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Formalises technical lag as a function of (used version, ideal version available at reference time t) with pluggable lag measures (version, time, number of releases)."}]},{"k":"zerrow2025arrow","a":"foundations-systems","t":"Zerrow: True Zero-Copy Arrow Pipelines in Bauplan","y":2025,"s":"preprint","e":2,"x":"2504.06151","v":"arXiv:2504.06151","a_":["Dai et al."],"bl":"https://arxiv.org/abs/2504.06151","pc":2,"sm":[{"m":"agent draft (gemini)","l":"columnar architecture","tx":"Analyzes why Python/process-boundary data pipelines pay heavy hidden serialization and memory-copy taxes even when claiming 'Arrow compatibility', and shows how true zero-copy Arrow RecordBatch pipelines eliminate double-buffering and materialization overhead."},{"m":"agent draft (gemini)","l":"systems architecture","tx":"Dai, Tagliabue, Arpaci-Dusseau, Arpaci-Dusseau & Caraza-Harter (arXiv:2504.06151, 2025) present Zerrow, proving that Apache Arrow's pointer-free columnar layout enables true zero-copy buffer handoff across language runtimes (Rust <-> Python/DuckDB) and process boundaries without serialization/deserialization copies."}]},{"k":"zhang-2013","a":"foundations-telemetry","t":"Predicting bug-fixing time: An empirical study of commercial software projects","y":2013,"s":"peer_reviewed","e":5,"d":"10.1109/ICSE.2013.6606654","v":"2013 35th International Conference on Software Engineering (ICSE)","a_":["Zhang et al."],"bl":"https://doi.org/10.1109/ICSE.2013.6606654","pc":1,"sm":[{"m":"agent draft (claude)","l":"empirical findings","tx":"Commercial (CA Technologies) projects: bug-fixing time is long-tailed; early report attributes carry signal for slow vs fast fixes (Markov-based and kNN models). Industrial evidence that initial-report attributes relate to resolution time, the hypothesis the intake seat tests with survival models."}]},{"k":"zhang2020confidenceexplanation","a":"calibration-conformal","t":"Effect of confidence and explanation on accuracy and trust calibration in AI-assisted decision making","y":2020,"s":"peer_reviewed","e":4,"d":"10.1145/3351095.3372852","v":"Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency","a_":["Zhang et al."],"bl":"https://doi.org/10.1145/3351095.3372852","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Showing per-case AI confidence helped people calibrate trust (rely more when confidence is high) but did not raise team accuracy on its own; local explanations did not help calibration."}]},{"k":"zhang2021upgrade","a":"ai-operations","t":"Understanding and Detecting Software Upgrade Failures in Distributed Systems","y":2021,"s":"peer_reviewed","e":4,"d":"10.1145/3477132.3483577","v":"Proceedings of the ACM SIGOPS 28th Symposium on Operating Systems Principles (SOSP '21)","a_":["Zhang et al."],"bl":"https://doi.org/10.1145/3477132.3483577","pc":4,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Upgrade failures in distributed systems are common and data-format driven. Council-5 DocDB lens: two 2025.2 field crash classes are upgrade/feature-promotion shaped (DB-24840 PITR sequences_data compaction after upgrade; DB-23719 AutoFlag promotion with in-flight transactions), so field crash rates should be split by days-since-upgrade."},{"m":"agent draft (gemini)","l":"s11 yba commissioner upgrade lifecycle","tx":"Operationalizes Zhang et al. Requires subtask-tree traversal of task_info.json (details.error.originMessage and pre-2.23 details.errorString) and logs-first per-node version skew detection."}]},{"k":"zhang2024ebsglory","a":"ai-operations","t":"What's the Story in EBS Glory: Evolutions and Lessons in Building Cloud Block Store","y":2024,"s":"peer_reviewed","e":3,"v":"22nd USENIX Conference on File and Storage Technologies (FAST 24)","a_":["Zhang et al."],"bl":"https://www.usenix.org/conference/fast24/presentation/zhang-weidong","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"fsync T6: in a cloud block store the guest sees only I/O that does not complete; with AWS's recommended nvme_core.io_timeout maximum this explains silent kernel logs during EBS-class stalls. Host-log processors are low-yield for the acute class; CSP-side evidence (VolumeStalledIOCheck) and in-process YB logs carry the signal."}]},{"k":"zhang2024genverifiers","a":"ai-operations","t":"Generative Verifiers: Reward Modeling as Next-Token Prediction","y":2024,"s":"preprint","e":2,"x":"2408.15240","v":"arXiv:2408.15240","a_":["Zhang et al."],"bl":"https://arxiv.org/abs/2408.15240","pc":2,"sm":[{"m":"agent draft (qwen)","l":"operational translation","tx":"A 7B/8B generative verifier beating GPT-4 as a critic of a larger generator's candidate solutions is the direct template for h2o-lightning-4b scoring qwen38-flash-next's candidate vulnerability reports: small-model-as-scorer is a supported deployment, small-model-as-generator of exploit chains is not."}]},{"k":"zhang2026piitrace","a":"security-governance","t":"PII-TRACE: A Benchmark for Context-Aware PII Detection in Multi-Turn LLM Conversations","y":2026,"s":"preprint","e":1,"x":"2609.22200","v":"arXiv preprint arXiv:2609.22200","a_":["Zhang et al."],"bl":"https://arxiv.org/abs/2609.22200","pc":1,"sm":[{"m":"agent draft (gemini)","l":"security and privacy","tx":"Introduces PII-TRACE, a multi-turn conversational benchmark evaluating context-dependent personally identifiable information and secret leakage across agent interactions."}]},{"k":"zhao-2020","a":"foundations-telemetry","t":"Automatically and Adaptively Identifying Severe Alerts for Online Service Systems","y":2020,"s":"peer_reviewed","e":5,"d":"10.1109/INFOCOM41043.2020.9155219","v":"IEEE INFOCOM 2020 - IEEE Conference on Computer Communications","a_":["Zhao et al."],"bl":"https://doi.org/10.1109/INFOCOM41043.2020.9155219","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"AlertRank ranks alerts by severity from textual, temporal and metric-anomaly features (XGBoost ranker; F1 about 0.89 on a bank's data). Framing for YBA alert review: measure precision and recall of default YBA alert policies against ticketed incidents rather than counting alerts fired."}]},{"k":"zhao-2021","a":"foundations-telemetry","t":"Identifying bad software changes via multimodal anomaly detection for online service systems","y":2021,"s":"peer_reviewed","e":11,"d":"10.1145/3468264.3468543","v":"Proceedings of the 29th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering","a_":["Zhao et al."],"bl":"https://doi.org/10.1145/3468264.3468543","pc":1,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"SCWarn flags bad software changes via multimodal anomaly detection (metrics + logs) around each change. A YB analogue anchors on declared YBA tasks (GFlagsUpgrade, SoftwareUpgrade, RestartUniverse) and compares minute signals before/after; feasible only on the ~100 cases whose task log reaches at least 1 day before the ticket."}]},{"k":"zhao2020alertstorm","a":"ai-operations","t":"Understanding and Handling Alert Storm for Online Service Systems","y":2020,"s":"peer_reviewed","e":3,"d":"10.1145/3377813.3381363","v":"Proceedings of the ACM/IEEE 42nd International Conference on Software Engineering: Software Engineering in Practice (ICSE-SEIP)","a_":["Zhao et al."],"bl":"https://doi.org/10.1145/3377813.3381363","pc":4,"sm":[{"m":"agent draft (claude)","l":"operational translation","tx":"Alert storms: many redundant low-value alerts at once; proposes clustering and representative-alert selection. For YBA the opposite failure also matters: on Comply 10-13 every alert raise failed to deliver (failedRaise) and no disk-latency check existed, so alert quality must include delivery and coverage, not only noise."},{"m":"agent draft (claude)","l":"sre platform","tx":"An empirical study of alert storms at a large service provider, where many correlated alerts fire during one incident. It proposes storm detection and alert aggregation and representative selection."}]},{"k":"zhao2021calibrate","a":"ai-operations","t":"Calibrate Before Use: Improving Few-Shot Performance of Language Models","y":2021,"s":"peer_reviewed","e":3,"v":"Proceedings of the 38th International Conference on Machine Learning (ICML)","a_":["Zhao et al."],"bl":"https://proceedings.mlr.press/v139/zhao21c.html","pc":3,"sm":[{"m":"agent draft (gemini)","l":"report 07 contextual calibration","tx":"Shows that language models suffer from majority-label, recency, and common-token biases that cause high variance across prompt formats, and corrects them via post-hoc affine calibration of output probabilities using content-free probes."}]},{"k":"zheng2023judging","a":"calibration-conformal","t":"Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena","y":2023,"s":"preprint","e":3,"x":"2306.05685","v":"Advances in Neural Information Processing Systems 36 (NeurIPS 2023 Datasets and Benchmarks Track)","a_":["Zheng et al."],"bl":"https://arxiv.org/abs/2306.05685","pc":5,"sm":[{"m":"agent draft (claude)","l":"limitations caveats","tx":"MT-Bench/Chatbot Arena: strong LLM judges approach human-human agreement on open-ended chat preferences but show position, verbosity and self-enhancement biases."},{"m":"agent draft (claude)","l":"operational translation","tx":"Strong LLM judges reach about 80% agreement with humans on chat preferences (similar to human-human), but show position bias, verbosity bias and self-enhancement bias."},{"m":"agent draft (claude)","l":"llmops","tx":"Strong LLM judges (GPT-4) reach over 80 percent agreement with human preferences, similar to inter-human agreement, but exhibit position, verbosity and self-enhancement biases and weak math/reasoning grading."}]},{"k":"zhou2023calcspar","a":"ai-operations","t":"Calcspar: A Contract-Aware LSM Store for Cloud Storage with Low Latency Spikes","y":2023,"s":"peer_reviewed","e":1,"v":"2023 USENIX Annual Technical Conference (USENIX ATC 23)","a_":["Zhou et al."],"bl":"https://www.usenix.org/conference/atc23/presentation/zhou","pc":1,"sm":[{"m":"agent draft (claude)","l":"empirical findings","tx":"Measures AWS EBS gp2/gp3/io1/io2: latency stays low until I/O pressure approaches the provisioned IOPS/throughput 'contract', then varies sharply; an LSM that rate-limits itself inside the contract avoids spikes (P99 <550 us). Relevance: throttling at a provisioned or instance cap turns into queueing latency for every I/O on the volume, including WAL fsync."}]},{"k":"zhou2024dbot","a":"ai-operations","t":"D-Bot: Database Diagnosis System using Large Language Models","y":2024,"s":"peer_reviewed","e":10,"d":"10.14778/3675034.3675043","v":"Proceedings of the VLDB Endowment (PVLDB)","a_":["Zhou et al."],"bl":"https://doi.org/10.14778/3675034.3675043","pc":6,"sm":[{"m":"agent draft (gemini)","l":"systems performance","tx":"Zhou et al. (PVLDB 2024) introduce D-Bot, an LLM-based database diagnosis system combining offline document knowledge extraction, tool-assisted metric/log retrieval, and tree-search root-cause analysis."},{"m":"agent draft (claude)","l":"bibliographic context","tx":"ACM TOS extended version of Calcspar (ATC'23): contract-aware, cost-effective LSM store for cloud block storage with low latency spikes; adds cost model. Cite with zhou2023calcspar."},{"m":"agent draft (gemini)","l":"empirical findings","tx":"Warehouse-scale characterization of TCMalloc across Google's production fleet (including Spanner, Bigtable, Monarch, F1), showing that allocator cycles, span/pageheap fragmentation, and dTLB page-walk stalls (4-10% of fleet cycles) are governed by object allocation rate (`allocs/op`) and object lifetime mixing on 2 MiB hugepages."}]},{"k":"zhu2019tools","a":"ai-operations","t":"Tools and Benchmarks for Automated Log Parsing","y":2019,"s":"peer_reviewed","e":21,"d":"10.1109/icse-seip.2019.00021","v":"2019 IEEE/ACM 41st International Conference on Software Engineering: Software Engineering in Practice (ICSE-SEIP)","a_":["Zhu et al."],"bl":"https://doi.org/10.1109/icse-seip.2019.00021","pc":2,"sm":[{"m":"agent draft (gemini)","l":"log template mining","tx":"Zhu et al. (ICSE-SEIP 2019, logparser & Loghub-2k) benchmark 13 automated log parsers across 16 open-source systems (2,000 lines per system) using Grouping Accuracy (GA: fraction of log lines whose induced partition class equals the ground-truth template's partition class) and demonstrate the necessity of domain-specific regex preprocessing (IPs, block…"}]},{"k":"zhu2023loghub","a":"ai-operations","t":"Loghub: A Large Collection of System Log Datasets for AI-driven Log Analytics","y":2023,"s":"peer_reviewed","e":2,"d":"10.1109/issre59848.2023.00071","v":"2023 IEEE 34th International Symposium on Software Reliability Engineering (ISSRE)","a_":["Zhu et al."],"bl":"https://doi.org/10.1109/issre59848.2023.00071","pc":1},{"k":"ziegler-2024","a":"foundations-se","t":"Measuring GitHub Copilot's Impact on Productivity","y":2024,"s":"peer_reviewed","e":2,"d":"10.1145/3633453","v":"Communications of the ACM","a_":["Ziegler et al."],"bl":"https://doi.org/10.1145/3633453","pc":1,"sm":[{"m":"agent draft (claude)","l":"support human ai","tx":"CACM 2024 (GitHub). Of many telemetry metrics, acceptance rate of shown suggestions best predicted developers' perceived productivity, better than persistence or retained-code measures."}]},{"k":"ziems-2024","a":"ai-operations","t":"Can Large Language Models Transform Computational Social Science?","y":2024,"s":"peer_reviewed","e":4,"d":"10.1162/coli_a_00502","v":"Computational Linguistics","a_":["Ziems et al."],"bl":"https://doi.org/10.1162/coli_a_00502"},{"k":"zong-2022","a":"foundations-viz","t":"Rich Screen Reader Experiences for Accessible Data Visualization","y":2022,"s":"peer_reviewed","e":3,"d":"10.1111/cgf.14519","v":"Computer Graphics Forum","a_":["Zong et al."],"bl":"https://doi.org/10.1111/cgf.14519"},{"k":"zong-2024","a":"foundations-se","t":"Umwelt: Accessible Structured Editing of Multi-Modal Data Representations","y":2024,"s":"peer_reviewed","e":4,"d":"10.1145/3613904.3641996","v":"Proceedings of the CHI Conference on Human Factors in Computing Systems","a_":["Zong et al."],"bl":"https://doi.org/10.1145/3613904.3641996"}],"edges":[{"id":"e0","i":"abadie-2021--CITES-->athey-2017","s":"abadie-2021","t":"athey-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1257/jel.20191450 -> 10.1257/jep.31.2.3)","m":"reference-list extractor"},{"id":"e2","i":"abadie-2021--CITES-->hainmueller-2012","s":"abadie-2021","t":"hainmueller-2012","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1257/jel.20191450 -> 10.1093/pan/mpr025)","m":"reference-list extractor"},{"id":"e3","i":"abadie-2021--EXTENDS-->abadie2010sc","s":"abadie-2021","t":"abadie2010sc","r":"EXTENDS","w":2,"why":"Abadie 2021 consolidates the 2010 synthetic-control method with feasibility conditions, placebo-in-time and placebo-in-space inference, and data requirements.","m":"agent draft (claude)"},{"id":"e4","i":"abadie2023clustering--CITES-->cameron2015practitioner","s":"abadie2023clustering","t":"cameron2015practitioner","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1093/qje/qjac038 -> 10.3368/jhr.50.2.317)","m":"reference-list extractor"},{"id":"e7","i":"abeba-2024-arxiv--EXTENDS-->raji2020closing","s":"abeba-2024-arxiv","t":"raji2020closing","r":"EXTENDS","w":2,"why":"Empirical follow-up: audits rarely reach accountability outcomes","m":"agent draft (gemini)"},{"id":"e8","i":"abedi2017rmit--CITES-->schad-2010","s":"abedi2017rmit","t":"schad-2010","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3030207.3030229 -> 10.14778/1920841.1920902)","m":"reference-list extractor"},{"id":"e9","i":"abitbol2026revolut--COMPLEMENTS-->alcott2005jevons","s":"abitbol2026revolut","t":"alcott2005jevons","r":"COMPLEMENTS","w":2,"why":"Illustrates Jevons' efficiency paradox in enterprise AI: lower per-token inference cost induces high-frequency agentic verification loops that shift the break-even point toward fixed-cost local GPUs.","m":"agent draft (gemini)"},{"id":"e10","i":"abitbol2026revolut--USES_METHOD-->hoefler-2015","s":"abitbol2026revolut","t":"hoefler-2015","r":"USES_METHOD","w":2,"why":"Follows non-parametric percentile and tail-latency reporting guidelines from Hoefler & Belli (SC'15) when comparing self-hosted GPU vs managed API throughput.","m":"agent draft (gemini)"},{"id":"e11","i":"abowd2018census--CITES-->dinur2003revealing","s":"abowd2018census","t":"dinur2003revealing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3219819.3226070 -> 10.1145/773153.773173)","m":"reference-list extractor"},{"id":"e12","i":"abowd2018census--OPERATIONALIZES-->dwork2006calibrating","s":"abowd2018census","t":"dwork2006calibrating","r":"OPERATIONALIZES","w":2,"why":"Operationalizes formal differential privacy budgets and top-down hierarchical noise injection across the 2020 US Decennial Census disclosure avoidance system.","m":"agent draft (gemini)"},{"id":"e13","i":"abraham2019datagovernance--EXTENDS-->khatri2010designing","s":"abraham2019datagovernance","t":"khatri2010designing","r":"EXTENDS","w":3,"why":"Extends Khatri & Brown's (2010) 5 decision domains into a 6-dimensional framework across structural, procedural, and relational mechanisms for structured and unstructured data.","m":"agent draft (gemini)"},{"id":"e14","i":"adams2020monarch--CITES-->bloom1970space","s":"adams2020monarch","t":"bloom1970space","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3181-3194 -> 10.1145/362686.362692)","m":"reference-list extractor"},{"id":"e15","i":"adams2020monarch--CITES-->corbett2013spanner","s":"adams2020monarch","t":"corbett2012spanner","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3181-3194 -> 10.1145/2491245)","m":"reference-list extractor"},{"id":"e17","i":"adams2020monarch--COMPLEMENTS-->luo2020lsm","s":"adams2020monarch","t":"luo-2020","r":"COMPLEMENTS","w":3,"why":"Pairs counter-reset-aware time-series window aggregation with LSM-tree compaction and write-amplification telemetry.","m":"agent draft (gemini)"},{"id":"e19","i":"adams2020monarch--EXTENDS-->pelkonen2015gorilla","s":"adams2020monarch","t":"pelkonen2015gorilla","r":"EXTENDS","w":3,"why":"Extends Gorilla's in-memory time-series storage model with explicit cumulative counter-reset semantics and zoned multi-tenant query federation.","m":"agent draft (gemini)"},{"id":"e21","i":"afroozeh-2023--CITES-->pelkonen2015gorilla","s":"afroozeh-2023","t":"pelkonen2015gorilla","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3626717 -> 10.14778/2824032.2824078)","m":"reference-list extractor"},{"id":"e22","i":"afroozeh-2023--CITES-->raasveldt2019duckdb","s":"afroozeh-2023","t":"raasveldt2019duckdb","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3626717 -> 10.1145/3299869.3320212)","m":"reference-list extractor"},{"id":"e23","i":"afroozeh-2025--CITES-->kuschewski2023btrblocks","s":"afroozeh-2025","t":"kuschewski2023btrblocks","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3749646.3749718 -> 10.1145/3589263)","m":"reference-list extractor"},{"id":"e24","i":"afroozeh-2025--CITES-->pelkonen2015gorilla","s":"afroozeh-2025","t":"pelkonen2015gorilla","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3749646.3749718 -> 10.14778/2824032.2824078)","m":"reference-list extractor"},{"id":"e25","i":"afroozeh-2025--CITES-->raasveldt2019duckdb","s":"afroozeh-2025","t":"raasveldt2019duckdb","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3749646.3749718 -> 10.1145/3299869.3320212)","m":"reference-list extractor"},{"id":"e26","i":"afroozeh-2025--CONTRASTS_WITH-->zeng2023columnar","s":"afroozeh-2025","t":"zeng2023columnar","r":"CONTRASTS_WITH","w":2,"why":"Zeng evaluates Parquet/ORC; FastLanes proposes successor addressing decode-speed gaps","m":"agent draft (gemini)"},{"id":"e27","i":"afroozeh-2025--EXTENDS-->afroozeh-2023","s":"afroozeh-2025","t":"afroozeh-2023","r":"EXTENDS","w":2,"why":"FastLanes format adopts ALP for floats","m":"agent draft (gemini)"},{"id":"e29","i":"aggarwal2024automix--USES_METHOD-->wang2023selfconsistency","s":"aggarwal2024automix","t":"wang2023selfconsistency","r":"USES_METHOD","w":2,"why":"AutoMix's surrogate verifier uses sampled self-consistency to decide escalation","m":"agent draft (qwen)"},{"id":"e30","i":"aghili2025logs--EXTENDS-->foukarakis2009deep","s":"aghili2025logs","t":"foukarakis2009deep","r":"EXTENDS","w":2,"why":"Extends protocol field anonymization concepts to software engineering event logs.","m":"agent draft (gemini)"},{"id":"e31","i":"aghili2025logs--EXTENDS-->sweeney2002kanonymity","s":"aghili2025logs","t":"sweeney2002kanonymity","r":"EXTENDS","w":2,"why":"Extends quasi-identifier anonymization and k-anonymity principles to unstructured software diagnostic logs","m":"agent draft (gemini)"},{"id":"e35","i":"ahmed2023recommending--EXTENDS-->chen-rcacopilot-2024","s":"chen2024rcacopilot","t":"ahmed2023recommending","r":"EXTENDS","w":2,"why":"Earlier LLM RCA recommendation; RCACopilot adds retrieval and handlers","m":"agent draft (gemini)"},{"id":"e37","i":"ahmed2023rootcausellm--CITES-->chen2019incidenttriage","s":"ahmed2023recommending","t":"chen2019incidenttriage","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00149 -> 10.1109/icse-seip.2019.00020)","m":"reference-list extractor"},{"id":"e39","i":"ahmed2023rootcausellm--CITES-->gao2018crashrecovery","s":"ahmed2023recommending","t":"gao2018crashrecovery","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00149 -> 10.1145/3236024.3236030)","m":"reference-list extractor"},{"id":"e40","i":"ahmed2023rootcausellm--CITES-->ghosh2022fight","s":"ahmed2023recommending","t":"ghosh2022fight","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00149 -> ghosh2022fight)","m":"reference-list extractor"},{"id":"e41","i":"ahmed2023rootcausellm--CITES-->leesatapornwongsa2016taxdc","s":"ahmed2023recommending","t":"leesatapornwongsa2016taxdc","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00149 -> 10.1145/2872362.2872374)","m":"reference-list extractor"},{"id":"e42","i":"ahmed2023rootcausellm--CITES-->liu2019cloudincidents","s":"ahmed2023recommending","t":"liu2019cloudincidents","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00149 -> 10.1145/3317550.3321438)","m":"reference-list extractor"},{"id":"e44","i":"ahmed2023rootcausellm--CITES-->yuan2014simple","s":"ahmed2023recommending","t":"yuan2014simple","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00149 -> yuan2014simple)","m":"reference-list extractor"},{"id":"e45","i":"ahmed2023rootcausellm--CITES-->zhang2021upgrade","s":"ahmed2023recommending","t":"zhang2021upgrade","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00149 -> 10.1145/3477132.3483577)","m":"reference-list extractor"},{"id":"e46","i":"ahocorasick1975--CITES-->thompson1968nfa","s":"aho1975efficient","t":"thompson1968nfa","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/360825.360855 -> 10.1145/363347.363387)","m":"reference-list extractor"},{"id":"e47","i":"ajani-2022--CITES-->kosara-2013","s":"ajani-2022","t":"kosara-2013","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2021.3068337 -> 10.1109/mc.2013.36)","m":"reference-list extractor"},{"id":"e48","i":"ajani-2022--OPERATIONALIZES-->franconeri2021science","s":"ajani-2022","t":"franconeri2021science","r":"OPERATIONALIZES","w":2,"why":"tests declutter/focus guidance from practitioner literature","m":"agent draft (claude)"},{"id":"e49","i":"akbarinasaji-2017--CITES-->giger-2010","s":"akbarinasaji-2017","t":"giger-2010","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.jss.2017.02.021 -> giger-2010)","m":"reference-list extractor"},{"id":"e50","i":"akbarinasaji-2017--CITES-->hooimeijer-2007","s":"akbarinasaji-2017","t":"hooimeijer-2007","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.jss.2017.02.021 -> hooimeijer-2007)","m":"reference-list extractor"},{"id":"e51","i":"akbarinasaji-2017--CITES-->kim2006bugfixtime","s":"akbarinasaji-2017","t":"kim2006bugfixtime","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.jss.2017.02.021 -> kim2006bugfixtime)","m":"reference-list extractor"},{"id":"e52","i":"akbarinasaji-2017--CITES-->panjer-2007","s":"akbarinasaji-2017","t":"panjer-2007","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.jss.2017.02.021 -> panjer-2007)","m":"reference-list extractor"},{"id":"e53","i":"akbarinasaji-2017--CONTRASTS_WITH-->zhang-2013","s":"akbarinasaji-2017","t":"zhang-2013","r":"CONTRASTS_WITH","w":2,"why":"Replication on open-source data contrasts with the commercial-project fix-time prediction results; time models do not transfer cleanly.","m":"agent draft (claude-sonnet-5-5)"},{"id":"e55","i":"akidau2015dataflow--EXTENDS-->akidau2013millwheel","s":"akidau2015dataflow","t":"akidau2013millwheel","r":"EXTENDS","w":2,"why":"Generalizes MillWheel's low-watermark and stateful timer abstractions into the unified Dataflow programming model.","m":"agent draft (gemini)"},{"id":"e56","i":"akidau2015dataflow--EXTENDS-->li2008outoforder","s":"akidau2015dataflow","t":"li2008outoforder","r":"EXTENDS","w":2,"why":"Builds on Li et al.'s out-of-order stream processing (OOP) and slack/punctuation watermarks for disorder handling.","m":"agent draft (gemini)"},{"id":"e57","i":"aksin2007moderncallcenter--CITES-->borst-2004","s":"aksin2007moderncallcenter","t":"borst-2004","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1937-5956.2007.tb00288.x -> 10.1287/opre.1030.0081)","m":"reference-list extractor"},{"id":"e58","i":"aksin2007moderncallcenter--CITES-->brown-2005","s":"aksin2007moderncallcenter","t":"brown-2005","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1937-5956.2007.tb00288.x -> 10.1198/016214504000001808)","m":"reference-list extractor"},{"id":"e59","i":"aksin2007moderncallcenter--CITES-->green-2007","s":"aksin2007moderncallcenter","t":"green-2007","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1937-5956.2007.tb00288.x -> 10.1111/j.1937-5956.2007.tb00164.x)","m":"reference-list extractor"},{"id":"e60","i":"aksin2007moderncallcenter--CITES-->halfin-1981","s":"aksin2007moderncallcenter","t":"halfin-1981","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1937-5956.2007.tb00288.x -> 10.1287/opre.29.3.567)","m":"reference-list extractor"},{"id":"e61","i":"aksin2007moderncallcenter--CITES-->kolesar-1998","s":"aksin2007moderncallcenter","t":"kolesar-1998","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1937-5956.2007.tb00288.x -> 10.1111/j.1937-5956.1998.tb00457.x)","m":"reference-list extractor"},{"id":"e62","i":"aksin2007moderncallcenter--EXTENDS-->gans-2003","s":"aksin2007moderncallcenter","t":"gans-2003","r":"EXTENDS","w":2,"why":"Broadens Gans-Koole-Mandelbaum's OR tutorial to multi-disciplinary interfaces (HR, marketing).","m":"agent draft (claude)"},{"id":"e63","i":"alagappan2018par--CITES-->bairavasundaram2007latent","s":"alagappan2018par","t":"bairavasundaram2007latent","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3241062 -> 10.1145/1254882.1254917)","m":"reference-list extractor"},{"id":"e64","i":"alagappan2018par--CITES-->chidambaram-2013","s":"alagappan2018par","t":"chidambaram-2013","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3241062 -> 10.1145/2517349.2522726)","m":"reference-list extractor"},{"id":"e66","i":"alagappan2018par--COMPLEMENTS-->gao2018crashrecovery","s":"alagappan2018par","t":"gao2018crashrecovery","r":"COMPLEMENTS","w":2,"why":"Pairs protocol-aware recovery for replicated storage engines with empirical failure analysis of distributed crash-recovery bugs.","m":"agent draft (gemini)"},{"id":"e67","i":"alexander-2023-arxiv--COMPLEMENTS-->laurie-2014","s":"alexander-2023-arxiv","t":"laurie-2014","r":"COMPLEMENTS","w":2,"why":"SoK situates CT in the wider log-transparency design space (integrity vs audit vs accountability)","m":"agent draft (gemini)"},{"id":"e73","i":"an2015crashinducing--CITES-->kim2011topcrashes","s":"an2015crashinducing","t":"kim2011topcrashes","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2810146.2810152 -> 10.1109/tse.2011.20)","m":"reference-list extractor"},{"id":"e74","i":"an2015crashinducing--EXTENDS-->glerum2009wer","s":"an2015crashinducing","t":"glerum2009wer","r":"EXTENDS","w":2,"why":"Links field crash buckets to inducing commits/releases","m":"agent draft (gemini)"},{"id":"e75","i":"andersen-1982--COMPLEMENTS-->kaplan-1958","s":"andersen-1982","t":"kaplan-1958","r":"COMPLEMENTS","w":2,"why":"recurrent-event counting-process regression complements single-event product-limit estimation","m":"agent draft (gemini)"},{"id":"e76","i":"andersen-1982--EXTENDS-->cox-1972","s":"andersen-1982","t":"cox-1972","r":"EXTENDS","w":2,"why":"Counting-process formulation of the Cox model, enabling time-varying covariates.","m":"agent draft (claude)"},{"id":"e77","i":"anderson2004satisfaction--COMPLEMENTS-->morgan2006metrics","s":"anderson2004satisfaction","t":"morgan2006metrics","r":"COMPLEMENTS","w":2,"why":"Both support satisfaction (not recommend-intent) as a financially meaningful customer metric.","m":"agent draft (claude)"},{"id":"e80","i":"angelopoulos2022ltt--EXTENDS-->guo2017calibration","s":"angelopoulos2022ltt","t":"guo2017calibration","r":"EXTENDS","w":3,"why":"Extends post-hoc temperature scaling to finite-sample, distribution-free family-wise error control (Learn Then Test) for risk-controlling abstention gates.","m":"agent draft (gemini)"},{"id":"e81","i":"angelopoulos2022ltt--EXTENDS-->tibshirani2019conformal","s":"angelopoulos2022ltt","t":"tibshirani2019conformal","r":"EXTENDS","w":3,"why":"Extends split-conformal distribution-free guarantees to arbitrary non-monotone risk functions via Learn-Then-Test","m":"seed curator"},{"id":"e82","i":"angelopoulos2023ppi--COMPLEMENTS-->angelopoulos2022ltt","s":"angelopoulos2023ppi","t":"angelopoulos2022ltt","r":"COMPLEMENTS","w":3,"why":"Combines small gold human labels (n=45) with large model predictions (N=253) for unbiased variance reduction","m":"seed curator"},{"id":"e83","i":"angelopoulos2023ppi--COMPLEMENTS-->shankar2024validators","s":"angelopoulos2023ppi","t":"shankar2024validators","r":"COMPLEMENTS","w":2,"why":"PPI corrects judge-based quality estimates with a small human-labelled set; validators work aligns the judge itself","m":"agent draft (claude)"},{"id":"e85","i":"aoyang2026openrca2--EXTENDS-->pham2025rcaeval","s":"aoyang2026openrca2","t":"pham2025rcaeval","r":"EXTENDS","w":2,"why":"OpenRCA 2.0 moves RCA benchmarking from outcome labels (as in RCAEval/OpenRCA) to step-wise causal process supervision.","m":"agent draft (claude)"},{"id":"e86","i":"argote-2011--CITES-->edmondson-1999","s":"argote-2011","t":"edmondson-1999","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1287/orsc.1100.0621 -> 10.2307/2666999)","m":"reference-list extractor"},{"id":"e87","i":"armbrust2018structured--CITES-->akidau2015dataflow","s":"armbrust2018structured","t":"akidau2015dataflow","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3183713.3190664 -> 10.14778/2824032.2824076)","m":"reference-list extractor"},{"id":"e88","i":"armbrust2018structured--EXTENDS-->zaharia2013dstreams","s":"armbrust2018structured","t":"zaharia2013dstreams","r":"EXTENDS","w":2,"why":"Extends D-Streams micro-batch execution with event-time watermarks, stateful aggregation, and transactional Parquet/Delta sinks.","m":"agent draft (gemini)"},{"id":"e89","i":"armbrust2021lakehouse--EXTENDS-->armbrust-2020","s":"armbrust2021lakehouse","t":"armbrust-2020","r":"EXTENDS","w":2,"why":"Generalises the Delta transaction log into the lakehouse architecture","m":"agent draft (gemini)"},{"id":"e90","i":"arndt2017tethered--OPERATIONALIZES-->carroll2006measurementerror","s":"arndt2017tethered","t":"carroll2006measurementerror","r":"OPERATIONALIZES","w":2,"why":"Arndt et al. use the validation-subsample design (event-log surrogate for all, direct observation for 14 physicians) that measurement-error theory requires to calibrate a surrogate time measure.","m":"agent draft (claude)"},{"id":"e91","i":"artstein-2008--CITES-->cohen1960kappa","s":"artstein-2008","t":"cohen1960kappa","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1162/coli.07-034-r2 -> 10.1177/001316446002000104)","m":"reference-list extractor"},{"id":"e92","i":"artstein-2008--CITES-->feinstein1990kappa","s":"artstein-2008","t":"feinstein1990kappa","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1162/coli.07-034-r2 -> 10.1016/0895-4356(90)90158-l)","m":"reference-list extractor"},{"id":"e95","i":"artstein-2008--CONTRASTS_WITH-->landis-1977","s":"artstein-2008","t":"landis-1977","r":"CONTRASTS_WITH","w":2,"why":"Rejects Landis-Koch benchmark adjectives as thresholds","m":"agent draft (gemini)"},{"id":"e98","i":"ascarza-2017--CITES-->ascarza2018silent","s":"ascarza2018silent","t":"ascarza-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s40547-017-0080-0 -> 10.1287/mksc.2017.1057)","m":"reference-list extractor"},{"id":"e99","i":"ascarza2018futility--CONTRASTS_WITH-->hochstein2023csmhealth","s":"ascarza2018futility","t":"hochstein2023csmhealth","r":"CONTRASTS_WITH","w":2,"why":"Risk-based health targeting vs uplift-based targeting of retention effort","m":"agent draft (gemini)"},{"id":"e101","i":"ascarza2018futility--EXTENDS-->ascarza-2017","s":"ascarza2018futility","t":"ascarza-2017","r":"EXTENDS","w":2,"why":"Uplift targeting as the key operational issue in the retention-management review.","m":"agent draft (gemini)"},{"id":"e102","i":"athanassoulis-2016--EXTENDS-->o’neil-1996","s":"athanassoulis-2016","t":"o’neil-1996","r":"EXTENDS","w":2,"why":"Generalizes LSM vs B-tree tradeoffs into the 3-axis RUM Conjecture (Read, Update, Memory overhead)","m":"agent draft (gemini)"},{"id":"e103","i":"athey-2017--CITES-->abadie2010sc","s":"athey-2017","t":"abadie2010sc","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1257/jep.31.2.3 -> 10.1198/jasa.2009.ap08746)","m":"reference-list extractor"},{"id":"e105","i":"athey-2017--CITES-->frangakis2002principal","s":"athey-2017","t":"frangakis2002principal","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1257/jep.31.2.3 -> 10.1111/j.0006-341x.2002.00021.x)","m":"reference-list extractor"},{"id":"e106","i":"athey-2017--CITES-->prentice1989surrogate","s":"athey-2017","t":"prentice1989surrogate","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1257/jep.31.2.3 -> 10.1002/sim.4780080407)","m":"reference-list extractor"},{"id":"e107","i":"athey-2017--COMPLEMENTS-->lipsitch-2010","s":"athey-2017","t":"lipsitch-2010","r":"COMPLEMENTS","w":2,"why":"placebo tests in econometrics are the pseudo-outcome/pseudo-exposure counterpart of epidemiological negative controls","m":"agent draft (gemini)"},{"id":"e109","i":"austin-2011--COMPLEMENTS-->lipsitch-2010","s":"austin-2011","t":"lipsitch-2010","r":"COMPLEMENTS","w":2,"why":"Propensity weighting removes measured confounding; negative controls test for the unmeasured remainder.","m":"agent draft (claude)"},{"id":"e110","i":"austin2016competingrisks--OPERATIONALIZES-->dixon2010ces","s":"austin2016competingrisks","t":"dixon2010ces","r":"OPERATIONALIZES","w":2,"why":"Applies cumulative incidence functions (CIF) under competing risks to time-to-resolution modeling when service requests terminate via multiple outcomes.","m":"agent draft (gemini)"},{"id":"e114","i":"azmat2024xb--EXTENDS-->wright2003professional","s":"azmat2024xb","t":"wright2003professional","r":"EXTENDS","w":2,"why":"Quantifies tactical fouling efficiency via Expected Booking (xB) models, formalizing Wright's analysis of professional fouls where expected card cost is lower than transition threat.","m":"agent draft (gemini)"},{"id":"e115","i":"azure2026dasv5--COMPLEMENTS-->azure2026diskperf","s":"azure2026dasv5","t":"azure2026diskperf","r":"COMPLEMENTS","w":2,"why":"Per-size VM uncached caps used with the disk/VM cap model.","m":"agent draft (claude)"},{"id":"e117","i":"azure2026diskperf--COMPLEMENTS-->azure2026disktypes","s":"azure2026diskperf","t":"azure2026disktypes","r":"COMPLEMENTS","w":2,"why":"VM uncached caps vs disk-tier caps: the lower binds.","m":"agent draft (claude)"},{"id":"e122","i":"bacon2017spannersql--CITES-->o’neil-1996","s":"bacon2017spannersql","t":"o’neil-1996","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3035918.3056103 -> 10.1007/s002360050048)","m":"reference-list extractor"},{"id":"e123","i":"bacon2017spannersql--EXTENDS-->corbett2013spanner","s":"bacon2017spannersql","t":"corbett2012spanner","r":"EXTENDS","w":2,"why":"Spanner SQL system incl. schema change at scale","m":"agent draft (claude)"},{"id":"e124","i":"balkesen2013joins--EXTENDS-->chhugani2008sorting","s":"balkesen2013joins","t":"chhugani2008sorting","r":"EXTENDS","w":2,"why":"Balkesen et al. extend Chhugani et al.'s SIMD sort kernels to full multi-core sort-merge pipelines with explicit multi-thread range partitioning.","m":"agent draft (gemini)"},{"id":"e125","i":"balmau-2018--CITES-->bloom1970space","s":"balmau-2018","t":"bloom1970space","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3380905 -> 10.1145/362686.362692)","m":"reference-list extractor"},{"id":"e126","i":"balmau-2018--CITES-->dayan-2017","s":"balmau-2018","t":"dayan-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3380905 -> 10.1145/3035918.3064054)","m":"reference-list extractor"},{"id":"e127","i":"balmau-2018--CITES-->dayan-2018","s":"balmau-2018","t":"dayan-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3380905 -> 10.1145/3183713.3196927)","m":"reference-list extractor"},{"id":"e128","i":"balmau-2018--CITES-->o’neil-1996","s":"balmau-2018","t":"o’neil-1996","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3380905 -> 10.1007/s002360050048)","m":"reference-list extractor"},{"id":"e129","i":"balmau-2018--CITES-->raju-2017","s":"balmau-2018","t":"raju-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3380905 -> 10.1145/3132747.3132765)","m":"reference-list extractor"},{"id":"e130","i":"balmau-2018--EXTENDS-->dong-2021","s":"dong-2021","t":"balmau-2018","r":"EXTENDS","w":2,"why":"Eliminates RocksDB p99 latency spikes caused by L0/memtable write stalls via compaction I/O scheduling and preemption","m":"agent draft (gemini)"},{"id":"e131","i":"bamberg2012oncall--COMPLEMENTS-->nicol2004oncall","s":"bamberg2012oncall","t":"nicol2004oncall","r":"COMPLEMENTS","w":2,"why":"Bamberg et al.'s daily-diary study adds within-person evidence on on-call days and well-being to Nicol and Botterill's review of on-call work and health.","m":"agent draft (claude)"},{"id":"e132","i":"bansal-2020--CITES-->lamkanfi-2010","s":"bansal-2020","t":"lamkanfi-2010","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3377813.3381353 -> 10.1109/msr.2010.5463284)","m":"reference-list extractor"},{"id":"e134","i":"bansal-2020--COMPLEMENTS-->chen-2019-incident-triage","s":"bansal-2020","t":"chen-2019-incident-triage","r":"COMPLEMENTS","w":2,"why":"Both from Microsoft online-service operations: empirical triage study and an automated diagnosis/triage system.","m":"agent draft (claude-sonnet-5-5)"},{"id":"e135","i":"bansal-2021--CITES-->mitchell2019modelcards","s":"bansal-2021","t":"mitchell2019modelcards","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3411764.3445717 -> 10.1145/3287560.3287596)","m":"reference-list extractor"},{"id":"e136","i":"bansal-2021--COMPLEMENTS-->lee-2004","s":"bansal-2021","t":"lee-2004","r":"COMPLEMENTS","w":2,"why":"Empirical evidence that explanations inflate reliance without calibrating it to correctness; this violates Lee & See's appropriate-reliance goal.","m":"agent draft (claude)"},{"id":"e137","i":"bansal-2021--CONTRASTS_WITH-->vasconcelos-2023","s":"bansal-2021","t":"vasconcelos-2023","r":"CONTRASTS_WITH","w":2,"why":"Bansal: explanations raise acceptance regardless of correctness; Vasconcelos: explanations reduce over-reliance when verifying is cheap.","m":"agent draft (claude)"},{"id":"e140","i":"barik2017readerrors--COMPLEMENTS-->becker2019compilererrors","s":"barik2017readerrors","t":"becker2019compilererrors","r":"COMPLEMENTS","w":2,"why":"Eye-tracking: developers do read error messages and they are as hard to read as source; supports investing in message text.","m":"agent draft (gemini)"},{"id":"e142","i":"barr2015oracle--COMPLEMENTS-->segura2016metamorphicsurvey","s":"barr2015oracle","t":"segura2016metamorphicsurvey","r":"COMPLEMENTS","w":2,"why":"MT is one answer to the oracle problem","m":"agent draft (gemini)"},{"id":"e143","i":"basak2023comparative--COMPLEMENTS-->meli2019git","s":"basak2023comparative","t":"meli2019git","r":"COMPLEMENTS","w":2,"why":"Benchmarks secret-scanning detection tools across precision and recall, complementing Meli et al.'s empirical measurement of secret leakage across public Git repositories.","m":"agent draft (gemini)"},{"id":"e144","i":"basiri2016chaos--CITES-->yuan2014simple","s":"basiri2016chaos","t":"yuan2014simple","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/ms.2016.60 -> yuan2014simple)","m":"reference-list extractor"},{"id":"e145","i":"bastani-2025--CITES-->dellacqua2023jagged","s":"bastani-2025","t":"dellacqua2023jagged","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1073/pnas.2422633122 -> 10.2139/ssrn.4573321)","m":"reference-list extractor"},{"id":"e147","i":"bastani-2025--CONTRASTS_WITH-->brynjolfsson-2025","s":"bastani-2025","t":"brynjolfsson-2025","r":"CONTRASTS_WITH","w":2,"why":"Brynjolfsson: AI speeds novices along the experience curve; Bastani: unguarded AI raises assisted performance but harms unassisted learning.","m":"agent draft (claude)"},{"id":"e148","i":"batini2009methodologies--CITES-->wang1996dataquality","s":"batini2009methodologies","t":"wang1996dataquality","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1541880.1541883 -> 10.1080/07421222.1996.11518099)","m":"reference-list extractor"},{"id":"e150","i":"bayerl2011intercoder--CITES-->artstein-2008","s":"bayerl2011intercoder","t":"artstein-2008","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1162/coli_a_00074 -> 10.1162/coli.07-034-r2)","m":"reference-list extractor"},{"id":"e151","i":"bayerl2011intercoder--CITES-->sim-2005","s":"bayerl2011intercoder","t":"sim-2005","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1162/coli_a_00074 -> 10.1093/ptj/85.3.257)","m":"reference-list extractor"},{"id":"e152","i":"bayerl2011intercoder--COMPLEMENTS-->krippendorff-2004","s":"bayerl2011intercoder","t":"krippendorff-2004","r":"COMPLEMENTS","w":2,"why":"Determinants of the agreement that alpha measures","m":"agent draft (gemini)"},{"id":"e153","i":"begg1983verification--COMPLEMENTS-->ransohoff1978spectrum","s":"begg1983verification","t":"ransohoff1978spectrum","r":"COMPLEMENTS","w":2,"why":"Jointly characterize how verification bias and clinical spectrum bias distort diagnostic sensitivity and specificity estimates.","m":"agent draft (gemini)"},{"id":"e155","i":"beller2015idetesting--CONTRASTS_WITH-->meyer2017worklife","s":"beller2015idetesting","t":"meyer2017worklife","r":"CONTRASTS_WITH","w":2,"why":"Both pair telemetry with developer self-report; Beller quantifies a twofold self-overestimate of testing time, cautioning against treating self-report as ground truth in monitored-plus-survey designs.","m":"agent draft (claude)"},{"id":"e156","i":"bendeck-2025--CITES-->dibia-2023","s":"bendeck-2025","t":"dibia-2023","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2024.3456155 -> 10.18653/v1/2023.acl-demo.11)","m":"reference-list extractor"},{"id":"e157","i":"bendeck-2025--CITES-->kong2018titles","s":"bendeck-2025","t":"kong2018titles","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2024.3456155 -> 10.1145/3173574.3174012)","m":"reference-list extractor"},{"id":"e158","i":"bendeck-2025--CITES-->stokes-2022","s":"bendeck-2025","t":"stokes-2022","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2024.3456155 -> 10.1109/tvcg.2022.3209383)","m":"reference-list extractor"},{"id":"e164","i":"bertini-2020--CITES-->heer2010crowdsourcing","s":"bertini-2020","t":"heer2010crowdsourcing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/vis47514.2020.00048 -> 10.1145/1753326.1753357)","m":"reference-list extractor"},{"id":"e165","i":"bertini-2020--CITES-->munzner2014vad","s":"bertini-2020","t":"munzner2014vad","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/vis47514.2020.00048 -> 10.1201/b17511)","m":"reference-list extractor"},{"id":"e166","i":"bertini-2020--CONTRASTS_WITH-->cleveland1984graphical","s":"bertini-2020","t":"cleveland1984graphical","r":"CONTRASTS_WITH","w":2,"why":"argues precision-ranking is not the only goal of chart design","m":"agent draft (claude)"},{"id":"e167","i":"berzak2016anchoring--EXTENDS-->lingren2014preannotation","s":"berzak2016anchoring","t":"lingren2014preannotation","r":"EXTENDS","w":2,"why":"Measures anchoring on pre-annotations and its inflation of agreement","m":"agent draft (gemini)"},{"id":"e169","i":"beschastnikh2011synoptic--CITES-->yuan2010sherlog","s":"beschastnikh2011synoptic","t":"yuan2010sherlog","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2025113.2025151 -> 10.1145/1736020.1736038)","m":"reference-list extractor"},{"id":"e170","i":"beschastnikh2011synoptic--COMPLEMENTS-->nagaraj2012distalyzer","s":"beschastnikh2011synoptic","t":"nagaraj2012distalyzer","r":"COMPLEMENTS","w":2,"why":"Synoptic enforces temporal ordering invariants within execution traces while Distalyzer performs statistical differential comparison across peer nodes.","m":"agent draft (gemini)"},{"id":"e171","i":"bevan-2006--CITES-->holmstrom-1991","s":"bevan2006targets","t":"holmstrom-1991","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1467-9299.2006.00600.x -> 10.1093/jleo/7.special_issue.24)","m":"reference-list extractor"},{"id":"e172","i":"bevan2006targets--COMPLEMENTS-->campbell1979planned","s":"bevan2006targets","t":"campbell1979planned","r":"COMPLEMENTS","w":2,"why":"Bevan and Hood provide empirical evidence of indicator gaming under NHS targets, consistent with Campbell's law on corruption of quantitative indicators used for decisions.","m":"agent draft (claude)"},{"id":"e173","i":"bevan2006targets--COMPLEMENTS-->goodhart-1984","s":"bevan2006targets","t":"goodhart-1984","r":"COMPLEMENTS","w":2,"why":"Empirical NHS evidence of Goodhart-style gaming under hard targets","m":"agent draft (gemini)"},{"id":"e175","i":"bhuiyan2025sokredos--CITES-->davis2019rethinking","s":"bhuiyan2025sokredos","t":"davis2019rethinking","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3708821.3733912 -> 10.1145/3338906.3342509)","m":"reference-list extractor"},{"id":"e176","i":"bhuiyan2025sokredos--CITES-->thompson1968nfa","s":"bhuiyan2025sokredos","t":"thompson1968nfa","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3708821.3733912 -> 10.1145/363347.363387)","m":"reference-list extractor"},{"id":"e177","i":"bhuiyan2025sokredos--EXTENDS-->davis2018redos","s":"bhuiyan2025sokredos","t":"davis2018redos","r":"EXTENDS","w":2,"why":"Extends Davis et al.'s ecosystem ReDoS study to a 93-paper systematization comparing Spencer backtracking vs Thompson NFA/DFA engines.","m":"agent draft (gemini)"},{"id":"e178","i":"bird2009fair--CONTRASTS_WITH-->sliwerski2005szz","s":"bird2009fair","t":"sliwerski2005szz","r":"CONTRASTS_WITH","w":2,"why":"Shows SZZ-style linked samples are biased","m":"agent draft (gemini)"},{"id":"e180","i":"blackburn-2025--CITES-->mytkowicz2009wrong","s":"blackburn-2025","t":"mytkowicz2009wrong","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3669940.3707217 -> 10.1145/1508244.1508275)","m":"reference-list extractor"},{"id":"e181","i":"blackburn-2025--EXTENDS-->blackburn-2006","s":"blackburn-2025","t":"blackburn-2006","r":"EXTENDS","w":2,"why":"DaCapo Chopin is the successor suite to the 2006 DaCapo benchmarks, adding latency workloads and heap-size-aware methodology.","m":"agent draft (claude)"},{"id":"e182","i":"borkin2013memorable--CITES-->shneiderman1996eyes","s":"borkin2013memorable","t":"shneiderman1996eyes","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2013.234 -> 10.1109/vl.1996.545307)","m":"reference-list extractor"},{"id":"e183","i":"borkin2013memorable--CITES-->tufte2001vdqi","s":"borkin2013memorable","t":"tufte2001vdqi","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2013.234 -> tufte2001vdqi)","m":"reference-list extractor"},{"id":"e184","i":"borkin2016beyond--CITES-->pandey-2015","s":"borkin2016beyond","t":"pandey-2015","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2015.2467732 -> 10.1145/2702123.2702608)","m":"reference-list extractor"},{"id":"e185","i":"borkin2016beyond--EXTENDS-->borkin2013memorable","s":"borkin2016beyond","t":"borkin2013memorable","r":"EXTENDS","w":2,"why":"Extends intrinsic visualization memorability (2,070 charts) with eye-tracking fixation analysis and free-text recall across 393 charts, showing title + redundancy drive comprehension.","m":"agent draft (gemini)"},{"id":"e186","i":"bornholt2021s3--CITES-->claessen2000quickcheck","s":"bornholt2021s3","t":"claessen2000quickcheck","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3477132.3483540 -> 10.1145/351240.351266)","m":"reference-list extractor"},{"id":"e187","i":"bornholt2021s3--CITES-->o’neil-1996","s":"bornholt2021s3","t":"o’neil-1996","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3477132.3483540 -> o’neil-1996)","m":"reference-list extractor"},{"id":"e189","i":"borovkov2026selective--OPERATIONALIZES-->geifman2017selective","s":"borovkov2026selective","t":"geifman2017selective","r":"OPERATIONALIZES","w":2,"why":"Production support system uses a learned abstention critic (selective prediction) trained on operator accept/correct feedback.","m":"agent draft (claude)"},{"id":"e190","i":"borst-2004--EXTENDS-->halfin-1981","s":"borst-2004","t":"halfin-1981","r":"EXTENDS","w":2,"why":"Builds cost-optimal square-root staffing on the Halfin-Whitt QED regime.","m":"agent draft (claude)"},{"id":"e191","i":"bortnikov-2018--CITES-->lu-2017","s":"bortnikov-2018","t":"lu-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3229863.3229873 -> 10.1145/3033273)","m":"reference-list extractor"},{"id":"e192","i":"bortnikov-2018--CITES-->o’neil-1996","s":"bortnikov-2018","t":"o’neil-1996","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3229863.3229873 -> 10.1007/s002360050048)","m":"reference-list extractor"},{"id":"e193","i":"bortnikov-2018--COMPLEMENTS-->dayan-2018","s":"bortnikov-2018","t":"dayan-2018","r":"COMPLEMENTS","w":2,"why":"Applies LSM compaction principles inside the MemTable (Accordion) to cut memory fragmentation and flush WAF","m":"agent draft (gemini)"},{"id":"e194","i":"bouzeghoub2004freshness--EXTENDS-->wang1996dataquality","s":"bouzeghoub2004freshness","t":"wang1996dataquality","r":"EXTENDS","w":2,"why":"Refines Wang-Strong's contextual timeliness dimension into a freshness taxonomy (currency vs timeliness) for integration pipelines.","m":"agent draft (claude)"},{"id":"e196","i":"boyer1977fast--CITES-->aho1975efficient","s":"boyer1977fast","t":"aho1975efficient","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/359842.359859 -> 10.1145/360825.360855)","m":"reference-list extractor"},{"id":"e197","i":"breck2017mltestscore--OPERATIONALIZES-->shankar2022operationalizing","s":"breck2017mltestscore","t":"shankar2022operationalizing","r":"COMPLEMENTS","w":2,"why":"Both concern production ML validation; the ML Test Score is a concrete monitoring/test rubric while the interview study reports how practitioners actually validate and version in production. (Edge direction: rubric operationalizes the practice the study describes.)","m":"agent draft (claude)"},{"id":"e199","i":"breck2019datavalidation--OPERATIONALIZES-->iso-5259","s":"breck2019datavalidation","t":"iso-5259","r":"COMPLEMENTS","w":3,"why":"Implements automated batch schema validation, drift bounds, and training-serving skew checks that operationalize ISO/IEC 5259-2/4 and ISO/IEC 42001 A.7.4/A.7.6.","m":"agent draft (gemini)"},{"id":"e200","i":"brettel-1997--COMPLEMENTS-->machado-2009","s":"brettel-1997","t":"machado-2009","r":"COMPLEMENTS","w":2,"why":"alternate dichromat simulation; preferred for tritan; Machado handles anomalous trichromacy","m":"agent draft (claude)"},{"id":"e202","i":"breu2010infoneeds--CITES-->landis-1977","s":"breu2010infoneeds","t":"landis-1977","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1718918.1718973 -> 10.2307/2529310)","m":"reference-list extractor"},{"id":"e203","i":"breu2010infoneeds--COMPLEMENTS-->zimmermann2010bugreport","s":"breu2010infoneeds","t":"bettenburg2008goodbug","r":"COMPLEMENTS","w":2,"why":"Breu et al. (CSCW 2010) quantifies the resolution-time penalty of follow-up questions caused by the T0 information deficits identified in Zimmermann et al. (2010).","m":"agent draft (gemini)"},{"id":"e204","i":"breznau2022manyanalysts--CITES-->simmons2011false","s":"breznau2022manyanalysts","t":"simmons2011false","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1073/pnas.2203150119 -> 10.1177/0956797611417632)","m":"reference-list extractor"},{"id":"e205","i":"breznau2022manyanalysts--EXTENDS-->silberzahn2018manyanalysts","s":"breznau2022manyanalysts","t":"silberzahn2018manyanalysts","r":"EXTENDS","w":2,"why":"larger many-analysts replication; unexplained analyst variance","m":"agent draft (gemini)"},{"id":"e206","i":"brodie2005symptom--CONTRASTS_WITH-->he-2018","s":"brodie2005symptom","t":"he-2018","r":"CONTRASTS_WITH","w":2,"why":"Curated symptom matching against known problems vs unsupervised log clustering correlated with KPIs.","m":"agent draft (gemini)"},{"id":"e208","i":"brooks1987nosilver--CITES-->parnas1979extension","s":"brooks1987nosilverbullet","t":"parnas1979extension","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/mc.1987.1663532 -> 10.1109/tse.1979.234169)","m":"reference-list extractor"},{"id":"e209","i":"brown-2005--COMPLEMENTS-->gans-2003","s":"brown-2005","t":"gans-2003","r":"COMPLEMENTS","w":2,"why":"Same group; empirical statistical companion to the call-center OR tutorial (arrivals, lognormal service times, abandonment).","m":"agent draft (claude)"},{"id":"e210","i":"brown-2005--USES_METHOD-->kaplan-1958","s":"brown-2005","t":"kaplan-1958","r":"USES_METHOD","w":2,"why":"Brown et al. estimate customer patience under censoring with survival-analysis (product-limit) methods.","m":"agent draft (claude)"},{"id":"e211","i":"brown2001binomial--EXTENDS-->agresti1998approximate","s":"brown2001binomial","t":"agresti1998approximate","r":"EXTENDS","w":2,"why":"Brown-Cai-DasGupta systematise Agresti-Coull's finding that approximate (Wilson/AC) intervals beat Wald and Clopper-Pearson, adding Jeffreys for small n.","m":"agent draft (claude)"},{"id":"e212","i":"brundage2020trustworthy--COMPLEMENTS-->iso-42001","s":"brundage2020trustworthy","t":"iso-42001","r":"COMPLEMENTS","w":2,"why":"Brundage et al. (2020) define institutional, software, and hardware mechanisms that turn ISO/IEC 42001 Clause 7.5/9.1/A.6.2.8 documentation into verifiable claims.","m":"agent draft (gemini)"},{"id":"e213","i":"brundage2020trustworthy--COMPLEMENTS-->mitchell2019modelcards","s":"brundage2020trustworthy","t":"mitchell2019modelcards","r":"COMPLEMENTS","w":2,"why":"Pairs institutional audit trails and red-teaming claims with standardized Model Card performance and Subgroup reporting artifacts.","m":"agent draft (gemini)"},{"id":"e214","i":"brynjolfsson-2025--CITES-->borusyak2024imputation","s":"brynjolfsson-2025","t":"borusyak2024imputation","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1093/qje/qjae044 -> 10.1093/restud/rdae007)","m":"reference-list extractor"},{"id":"e215","i":"brynjolfsson-2025--CITES-->callaway2021did","s":"brynjolfsson-2025","t":"callaway2021did","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1093/qje/qjae044 -> 10.1016/j.jeconom.2020.12.001)","m":"reference-list extractor"},{"id":"e218","i":"brynjolfsson-2025--CITES-->sun2021event","s":"brynjolfsson-2025","t":"sun2021event","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1093/qje/qjae044 -> 10.1016/j.jeconom.2020.09.006)","m":"reference-list extractor"},{"id":"e219","i":"brynjolfsson-2025--CONTRASTS_WITH-->dellacqua2023jagged","s":"brynjolfsson-2025","t":"dellacqua2023jagged","r":"CONTRASTS_WITH","w":2,"why":"Support-agent gains concentrated in novices vs correctness loss outside the AI frontier","m":"agent draft (gemini)"},{"id":"e221","i":"brynjolfsson-2025--CONTRASTS_WITH-->hussey2007steppedwedge","s":"brynjolfsson-2025","t":"hussey2007steppedwedge","r":"CONTRASTS_WITH","w":2,"why":"Quasi-experimental staggered rollout vs randomized stepped wedge; both exploit staggered adoption to separate secular trend from assistant effect","m":"agent draft (claude)"},{"id":"e222","i":"brynjolfsson-2025--CONTRASTS_WITH-->ni2026genaialibaba","s":"brynjolfsson-2025","t":"ni2026genaialibaba","r":"CONTRASTS_WITH","w":2,"why":"Both field studies of GenAI in customer service: QJE finds novice gains and better outcomes; Alibaba finds speed/ratings up, no objective quality gain, top performers worse.","m":"agent draft (claude)"},{"id":"e223","i":"buell-2017--CITES-->cameron2008cluster","s":"buell-2017","t":"cameron2008cluster","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1287/mnsc.2015.2411 -> 10.1162/rest.90.3.414)","m":"reference-list extractor"},{"id":"e225","i":"burchell-2023--CITES-->blackburn-2006","s":"burchell-2023","t":"blackburn-2006","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3617651.3622985 -> 10.1145/1167473.1167488)","m":"reference-list extractor"},{"id":"e226","i":"burchell-2023--EXTENDS-->mytkowicz-2010","s":"burchell-2023","t":"mytkowicz-2010","r":"EXTENDS","w":2,"why":"Re-runs the PLDI'10 profiler-accuracy question on modern Java samplers including async-profiler and JFR.","m":"agent draft (claude)"},{"id":"e228","i":"burchell-2025--CITES-->blackburn-2025","s":"burchell-2025","t":"blackburn-2025","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3763180 -> 10.1145/3669940.3707217)","m":"reference-list extractor"},{"id":"e229","i":"burchell-2025--CITES-->curtsinger-2015","s":"burchell-2025","t":"curtsinger-2015","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3763180 -> 10.1145/2815400.2815409)","m":"reference-list extractor"},{"id":"e230","i":"burchell-2025--CITES-->mytkowicz-2010","s":"burchell-2025","t":"mytkowicz-2010","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3763180 -> 10.1145/1806596.1806618)","m":"reference-list extractor"},{"id":"e231","i":"burchell-2025--EXTENDS-->burchell-2023","s":"burchell-2025","t":"burchell-2023","r":"EXTENDS","w":2,"why":"Adds a ground-truth approximation (machine-code-level slowdown injection) to quantify the accuracy the 2023 study could only compare across profilers.","m":"agent draft (claude)"},{"id":"e232","i":"burns2023weakstrong--COMPLEMENTS-->zheng2023judging","s":"burns2023weakstrong","t":"zheng2023judging","r":"COMPLEMENTS","w":2,"why":"Weak supervision of a strong model quantifies the ceiling that weak judges/raters impose","m":"agent draft (qwen)"},{"id":"e236","i":"buçinca-2021--CITES-->raji2020closing","s":"buçinca-2021","t":"raji2020closing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3449287 -> 10.1145/3351095.3372873)","m":"reference-list extractor"},{"id":"e237","i":"buçinca-2021--OPERATIONALIZES-->parasuraman2010complacency","s":"buçinca-2021","t":"parasuraman2010complacency","r":"OPERATIONALIZES","w":2,"why":"Cognitive forcing functions are a concrete intervention against the automation bias and complacency described by Parasuraman & Manzey.","m":"agent draft (claude)"},{"id":"e238","i":"bystrom-1995--CITES-->wood-1986","s":"bystrom-1995","t":"wood-1986","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/0306-4573(95)80035-r -> 10.1016/0749-5978(86)90044-0)","m":"reference-list extractor"},{"id":"e239","i":"bystrom-1995--COMPLEMENTS-->campbell-1988","s":"bystrom-1995","t":"campbell-1988","r":"COMPLEMENTS","w":2,"why":"Both define complexity by uncertainty of path/outcome; Bystrom adds a priori determinability as an operational test.","m":"agent draft (claude-sonnet-5-5)"},{"id":"e240","i":"cahill2008ssi--CITES-->berenson-1995","s":"cahill2008ssi","t":"berenson-1995","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1376616.1376690 -> 10.1145/223784.223785)","m":"reference-list extractor"},{"id":"e241","i":"cai-2022--CITES-->blackburn-2006","s":"cai-2022","t":"blackburn-2006","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/ispass55109.2022.00005 -> 10.1145/1167473.1167488)","m":"reference-list extractor"},{"id":"e245","i":"callaway2021did--CITES-->mackinnon-2018","s":"callaway2021did","t":"mackinnon-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.jeconom.2020.12.001 -> 10.1111/ectj.12107)","m":"reference-list extractor"},{"id":"e247","i":"callaway2021did--COMPLEMENTS-->lopezbernal2018controls","s":"callaway2021did","t":"lopezbernal2018controls","r":"COMPLEMENTS","w":2,"why":"Controlled ITS and staggered DiD are complementary designs for evaluating a rollout with a comparison group.","m":"agent draft (claude)"},{"id":"e248","i":"callaway2021did--EXTENDS-->goodmanbacon2021decomp","s":"callaway2021did","t":"goodmanbacon2021decomp","r":"EXTENDS","w":2,"why":"CS ATT(g,t) estimator is the remedy for the TWFE weighting problem Goodman-Bacon decomposes.","m":"agent draft (gemini)"},{"id":"e252","i":"campbell-1979--COMPLEMENTS-->goodhart-1984","s":"campbell1979planned","t":"goodhart-1984","r":"COMPLEMENTS","w":2,"why":"Campbell's law (social indicators corrupted under decision pressure) is the social-science counterpart of Goodhart's law.","m":"agent draft (gemini)"},{"id":"e253","i":"campbell-1988--EXTENDS-->wood-1986","s":"campbell-1988","t":"wood-1986","r":"EXTENDS","w":2,"why":"objective task-complexity attributes building on Wood's construct","m":"agent draft (gemini)"},{"id":"e261","i":"canay-2021--CITES-->mackinnon2017wild","s":"canay-2021","t":"mackinnon2017wild","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1162/rest_a_00887 -> 10.1002/jae.2508)","m":"reference-list extractor"},{"id":"e262","i":"canay-2021--COMPLEMENTS-->mackinnon-2018","s":"canay-2021","t":"mackinnon-2018","r":"COMPLEMENTS","w":2,"why":"Theory for the wild bootstrap with few large clusters.","m":"agent draft (gemini)"},{"id":"e270","i":"carmel2010followsun--CITES-->parnas1972criteria","s":"carmel2010followsun","t":"parnas1972criteria","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.2753/mis0742-1222270102 -> 10.1145/361598.361623)","m":"reference-list extractor"},{"id":"e272","i":"casper2024blackbox--CITES-->mitchell2019modelcards","s":"casper2024blackbox","t":"mitchell2019modelcards","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3630106.3659037 -> 10.1145/3287560.3287596)","m":"reference-list extractor"},{"id":"e273","i":"casper2024blackbox--CITES-->mokander2023auditingllm","s":"casper2024blackbox","t":"mokander2023auditingllm","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3630106.3659037 -> 10.1007/s43681-023-00289-2)","m":"reference-list extractor"},{"id":"e274","i":"casper2024blackbox--CITES-->nistai100_1","s":"casper2024blackbox","t":"nistai100_1","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3630106.3659037 -> 10.6028/nist.ai.100-1)","m":"reference-list extractor"},{"id":"e277","i":"casper2024blackbox--COMPLEMENTS-->raji2020closing","s":"casper2024blackbox","t":"raji2020closing","r":"COMPLEMENTS","w":2,"why":"Raji gives the internal audit artifact trail; Casper argues audits need access to such internal records beyond black-box queries.","m":"agent draft (claude)"},{"id":"e279","i":"castro2008bugreporting--EXTENDS-->broadwell2003scrash","s":"castro2008bugreporting","t":"broadwell2003scrash","r":"EXTENDS","w":2,"why":"Advances Scrash's crash-memory scrubbing by proving symbolic control-flow preservation when replacing private user inputs with redacted values.","m":"agent draft (gemini)"},{"id":"e280","i":"caveness2020tfdv--OPERATIONALIZES-->breck2019datavalidation","s":"caveness2020tfdv","t":"breck2019datavalidation","r":"OPERATIONALIZES","w":2,"why":"TFDV is the open-source tool implementing the data-validation approach","m":"agent draft (gemini)"},{"id":"e283","i":"chaparro2017missing--CITES-->breu2010infoneeds","s":"chaparro2017missing","t":"breu2010infoneeds","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3106237.3106285 -> 10.1145/1718918.1718973)","m":"reference-list extractor"},{"id":"e284","i":"chaparro2017missing--EXTENDS-->zimmermann2010bugreport","s":"chaparro2017missing","t":"bettenburg2008goodbug","r":"EXTENDS","w":2,"why":"Chaparro et al. (FSE 2017) automates detection of missing behavioral and contextual slots in bug descriptions.","m":"agent draft (gemini)"},{"id":"e285","i":"chaparro2017missing--OPERATIONALIZES-->bettenburg2008goodbug","s":"chaparro2017missing","t":"bettenburg2008goodbug","r":"OPERATIONALIZES","w":2,"why":"Automatic detection of the missing elements Bettenburg identified as most useful (S2R, expected behaviour)","m":"agent draft (claude)"},{"id":"e287","i":"charlson-1987--CITES-->cox-1972","s":"charlson-1987","t":"cox-1972","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/0021-9681(87)90171-8 -> 10.1111/j.2517-6161.1972.tb00899.x)","m":"reference-list extractor"},{"id":"e288","i":"chen-2016--CITES-->kanev-2015","s":"chen-2016","t":"kanev-2015","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2854038.2854044 -> 10.1145/2749469.2750392)","m":"reference-list extractor"},{"id":"e289","i":"chen-2016--CITES-->melnik-2010","s":"chen-2016","t":"melnik-2010","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2854038.2854044 -> 10.14778/1920841.1920886)","m":"reference-list extractor"},{"id":"e290","i":"chen-2016--COMPLEMENTS-->ren-2010","s":"chen-2016","t":"ren-2010","r":"COMPLEMENTS","w":2,"why":"AutoFDO consumes fleet-wide sampling profiles of the kind Google-Wide Profiling collects.","m":"agent draft (gemini)"},{"id":"e291","i":"chen-2019-incident-triage--CITES-->jeong2009bugtossing","s":"chen-2019-incident-triage","t":"jeong2009bugtossing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse-seip.2019.00020 -> 10.1145/1595696.1595715)","m":"reference-list extractor"},{"id":"e292","i":"chen-2019-incident-triage--CITES-->lou-2013","s":"chen-2019-incident-triage","t":"lou-2013","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse-seip.2019.00020 -> 10.1109/ase.2013.6693105)","m":"reference-list extractor"},{"id":"e297","i":"chen-2021--CITES-->chen2019incidenttriage","s":"chen-2021","t":"chen2019incidenttriage","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/ase51524.2021.9678746 -> 10.1109/icse-seip.2019.00020)","m":"reference-list extractor"},{"id":"e299","i":"chen-2021--CITES-->he-2018","s":"chen-2021","t":"he-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/ase51524.2021.9678746 -> 10.1145/3236024.3236083)","m":"reference-list extractor"},{"id":"e300","i":"chen-2021--CITES-->huang2017gray","s":"chen-2021","t":"huang2017gray","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/ase51524.2021.9678746 -> 10.1145/3102980.3103005)","m":"reference-list extractor"},{"id":"e301","i":"chen-2021--CITES-->lin2016logcluster","s":"chen-2021","t":"lin2016logcluster","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/ase51524.2021.9678746 -> lin2016logcluster)","m":"reference-list extractor"},{"id":"e304","i":"chen-2021--CITES-->zhao-2020","s":"chen-2021","t":"zhao-2020","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/ase51524.2021.9678746 -> 10.1109/infocom41043.2020.9155219)","m":"reference-list extractor"},{"id":"e305","i":"chen-2021--CITES-->zhao2020alertstorm","s":"chen-2021","t":"zhao2020alertstorm","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/ase51524.2021.9678746 -> 10.1145/3377813.3381363)","m":"reference-list extractor"},{"id":"e306","i":"chen-2021--EXTENDS-->chen2020lidar","s":"chen-2021","t":"chen2020lidar","r":"EXTENDS","w":2,"why":"graph-based aggregation after LiDAR linked-incident identification","m":"agent draft (gemini)"},{"id":"e307","i":"chen-2022--CITES-->satyanarayan-2017","s":"chen-2022","t":"satyanarayan-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2021.3114804 -> 10.1109/tvcg.2016.2599030)","m":"reference-list extractor"},{"id":"e308","i":"chen-2022--OPERATIONALIZES-->mcnutt-2020","s":"chen-2022","t":"mcnutt-2020","r":"OPERATIONALIZES","w":2,"why":"VizLinter turns mirage/misleading-chart heuristics into executable lint rules","m":"agent draft (gemini)"},{"id":"e310","i":"chen-2024--CITES-->ghosh2022fight","s":"chen2024rcacopilot","t":"ghosh2022fight","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3627703.3629553 -> 10.1145/3542929.3563482)","m":"reference-list extractor"},{"id":"e311","i":"chen-2024--CITES-->gu2023acto","s":"chen2024rcacopilot","t":"gu2023acto","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3627703.3629553 -> 10.1145/3600006.3613161)","m":"reference-list extractor"},{"id":"e312","i":"chen-2024--CITES-->liu2019cloudincidents","s":"chen2024rcacopilot","t":"liu2019cloudincidents","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3627703.3629553 -> 10.1145/3317550.3321438)","m":"reference-list extractor"},{"id":"e314","i":"chen-2024--CITES-->zhang2021upgrade","s":"chen2024rcacopilot","t":"zhang2021upgrade","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3627703.3629553 -> 10.1145/3477132.3483577)","m":"reference-list extractor"},{"id":"e315","i":"chen-2024--COMPLEMENTS-->angelopoulos2022ltt","s":"chen2024rcacopilot","t":"angelopoulos2022ltt","r":"COMPLEMENTS","w":2,"why":"Pairs RCACopilot's automated cloud incident handler matching and root-cause category prediction with distribution-free conformal risk control.","m":"agent draft (gemini)"},{"id":"e316","i":"chen-2025--CITES-->liu2024lost","s":"chen-2025","t":"liu2024lost","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2024.3456320 -> 10.1162/tacl_a_00638)","m":"reference-list extractor"},{"id":"e317","i":"chen-2025--CITES-->lo-2022","s":"chen-2025","t":"lo-2022","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2024.3456320 -> 10.1111/cgf.14559)","m":"reference-list extractor"},{"id":"e319","i":"chen-2025--CITES-->satyanarayan-2017","s":"chen-2025","t":"satyanarayan-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2024.3456320 -> 10.1109/tvcg.2016.2599030)","m":"reference-list extractor"},{"id":"e320","i":"chen-2025--CITES-->zhao2021calibrate","s":"chen-2025","t":"zhao2021calibrate","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2024.3456320 -> zhao2021calibrate)","m":"reference-list extractor"},{"id":"e322","i":"chen2014ormantipatterns--CITES-->georges2007statistically","s":"chen2014ormantipatterns","t":"georges2007statistically","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2568225.2568259 -> 10.1145/1297027.1297033)","m":"reference-list extractor"},{"id":"e324","i":"chen2016redundantdataaccess--CITES-->georges2007statistically","s":"chen2016redundantdataaccess","t":"georges2007statistically","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2016.2553039 -> 10.1145/1297027.1297033)","m":"reference-list extractor"},{"id":"e325","i":"chen2016redundantdataaccess--CITES-->qiu2013schemacoevolution","s":"chen2016redundantdataaccess","t":"qiu2013schemacoevolution","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2016.2553039 -> 10.1145/2491411.2491431)","m":"reference-list extractor"},{"id":"e326","i":"chen2016redundantdataaccess--EXTENDS-->chen2014ormantipatterns","s":"chen2016redundantdataaccess","t":"chen2014ormantipatterns","r":"EXTENDS","w":2,"why":"Extends ORM anti-pattern detection to redundant data access with impact evaluation","m":"agent draft (claude)"},{"id":"e328","i":"chen2019incidenttriage--COMPLEMENTS-->jeong2009bugtossing","s":"chen2019incidenttriage","t":"jeong2009bugtossing","r":"COMPLEMENTS","w":2,"why":"Incident reassignment at Microsoft (4-92% reassigned, up to 10x triage time) mirrors bug tossing in OSS: hand-offs are common and each adds latency.","m":"agent draft (claude)"},{"id":"e329","i":"chen2019incidenttriage--COMPLEMENTS-->lou-2013","s":"chen2019incidenttriage","t":"lou-2013","r":"COMPLEMENTS","w":2,"why":"Incident triage misrouting study and Service Analysis Studio experience report, both Microsoft incident management.","m":"agent draft (gemini)"},{"id":"e330","i":"chen2020incidental--CITES-->chen2019incidenttriage","s":"chen2020incidental","t":"chen2019incidenttriage","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3324884.3416624 -> 10.1109/icse-seip.2019.00020)","m":"reference-list extractor"},{"id":"e331","i":"chen2020incidental--CITES-->kim2006bugfixtime","s":"chen2020incidental","t":"kim2006bugfixtime","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3324884.3416624 -> 10.1145/1137983.1138027)","m":"reference-list extractor"},{"id":"e333","i":"chen2020incidental--CITES-->yin2011configerrors","s":"chen2020incidental","t":"yin2011configerrors","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3324884.3416624 -> 10.1145/2043556.2043572)","m":"reference-list extractor"},{"id":"e334","i":"chen2020lidar--CITES-->chen2019incidenttriage","s":"chen2020lidar","t":"chen2019incidenttriage","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3368089.3417055 -> 10.1109/icse-seip.2019.00020)","m":"reference-list extractor"},{"id":"e336","i":"chen2020lidar--CITES-->fu2009execution","s":"chen2020lidar","t":"fu2009execution","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3368089.3417055 -> 10.1109/icdm.2009.60)","m":"reference-list extractor"},{"id":"e337","i":"chen2020lidar--CITES-->gunawi2016cloudstop","s":"chen2020lidar","t":"gunawi2016cloudstop","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3368089.3417055 -> 10.1145/2987550.2987583)","m":"reference-list extractor"},{"id":"e338","i":"chen2020lidar--CITES-->he-2018","s":"chen2020lidar","t":"he-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3368089.3417055 -> 10.1145/3236024.3236083)","m":"reference-list extractor"},{"id":"e339","i":"chen2020lidar--CITES-->he2016experience","s":"chen2020lidar","t":"he2016experience","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3368089.3417055 -> 10.1109/issre.2016.21)","m":"reference-list extractor"},{"id":"e340","i":"chen2020lidar--CITES-->huang2017gray","s":"chen2020lidar","t":"huang2017gray","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3368089.3417055 -> 10.1145/3102980.3103005)","m":"reference-list extractor"},{"id":"e341","i":"chen2020lidar--CITES-->lin2016logcluster","s":"chen2020lidar","t":"lin2016logcluster","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3368089.3417055 -> 10.1145/2889160.2889232)","m":"reference-list extractor"},{"id":"e343","i":"chen2020lidar--EXTENDS-->chen2020incidental","s":"chen2020lidar","t":"chen2020incidental","r":"EXTENDS","w":2,"why":"From recurring-incident observation to a linking model","m":"agent draft (gemini)"},{"id":"e344","i":"chen2020lidar--EXTENDS-->lou-2013","s":"chen2020lidar","t":"lou-2013","r":"EXTENDS","w":2,"why":"same Microsoft group; 2020 generalises the 2013 incident-analytics experience report into a lifecycle framework","m":"agent draft (gemini)"},{"id":"e345","i":"chen2020strangler--OPERATIONALIZES-->parnas1972criteria","s":"chen2020strangler","t":"parnas1972criteria","r":"OPERATIONALIZES","w":2,"why":"Operationalizes Parnas's information-hiding interfaces via Strangler Fig Anti-Corruption Layer facades during incremental legacy migration.","m":"agent draft (gemini)"},{"id":"e347","i":"chen2024rcacopilot--CITES-->chen2019incidenttriage","s":"chen2024rcacopilot","t":"chen2019incidenttriage","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3627703.3629553 -> 10.1109/icse-seip.2019.00020)","m":"reference-list extractor"},{"id":"e348","i":"chen2024rcacopilot--CITES-->chen2020incidental","s":"chen2024rcacopilot","t":"chen2020incidental","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3627703.3629553 -> 10.1145/3324884.3416624)","m":"reference-list extractor"},{"id":"e349","i":"chen2024rcacopilot--CITES-->gao2018crashrecovery","s":"chen2024rcacopilot","t":"gao2018crashrecovery","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3627703.3629553 -> 10.1145/3236024.3236030)","m":"reference-list extractor"},{"id":"e350","i":"chen2024rcacopilot--COMPLEMENTS-->liu2024lost","s":"chen2024rcacopilot","t":"liu2024lost","r":"COMPLEMENTS","w":2,"why":"RCACopilot's deterministic pre-LLM log aggregation operationalizes a defense against Liu et al.'s Lost-in-the-Middle long-context degradation.","m":"agent draft (gemini)"},{"id":"e354","i":"chhugani2008sorting--EXTENDS-->graefe2006sorting","s":"chhugani2008sorting","t":"graefe2006sorting","r":"EXTENDS","w":2,"why":"Chhugani et al. adapt Graefe's database merge-sort principles to modern multi-core SIMD CPUs via register-level bitonic sort networks and L2-cache-sized multi-way merge trees.","m":"agent draft (gemini)"},{"id":"e356","i":"choetkiertikul-2019--CITES-->giger-2010","s":"choetkiertikul-2019","t":"giger-2010","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2018.2792473 -> 10.1145/1808920.1808933)","m":"reference-list extractor"},{"id":"e357","i":"choetkiertikul-2019--CITES-->hooimeijer-2007","s":"choetkiertikul-2019","t":"hooimeijer-2007","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2018.2792473 -> 10.1145/1321631.1321639)","m":"reference-list extractor"},{"id":"e359","i":"choetkiertikul-2019--CITES-->panjer-2007","s":"choetkiertikul-2019","t":"panjer-2007","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2018.2792473 -> 10.1109/msr.2007.25)","m":"reference-list extractor"},{"id":"e360","i":"choetkiertikul-2019--CITES-->shepperd-2012","s":"choetkiertikul-2019","t":"shepperd-2012","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2018.2792473 -> 10.1016/j.infsof.2011.12.008)","m":"reference-list extractor"},{"id":"e363","i":"chu-2021--CITES-->du2016spell","s":"chu-2021","t":"du2016spell","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icde51399.2021.00274 -> du2016spell)","m":"reference-list extractor"},{"id":"e364","i":"chu-2021--CITES-->fu2009execution","s":"chu-2021","t":"fu2009execution","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icde51399.2021.00274 -> 10.1109/icdm.2009.60)","m":"reference-list extractor"},{"id":"e365","i":"chu-2021--CITES-->zhu2019tools","s":"chu-2021","t":"zhu2019tools","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icde51399.2021.00274 -> 10.1109/icse-seip.2019.00021)","m":"reference-list extractor"},{"id":"e366","i":"chu-2021--EXTENDS-->he2017drain","s":"chu-2021","t":"he2017drain","r":"EXTENDS","w":2,"why":"Merges a prefix tree with a probabilistic token-transition graph (Prefix-Graph) to remove Drain's rigid equal-length assumption and resolve mid-message template branching.","m":"agent draft (gemini)"},{"id":"e367","i":"chu-2021--USES_METHOD-->morrison-1968","s":"chu-2021","t":"morrison-1968","r":"USES_METHOD","w":2,"why":"Uses a prefix trie in the first phase before merging prefix states into a probabilistic token graph.","m":"agent draft (gemini)"},{"id":"e368","i":"clayton1987ebayes--COMPLEMENTS-->spiegelhalter2005overdispersion","s":"clayton1987ebayes","t":"spiegelhalter2005overdispersion","r":"COMPLEMENTS","w":2,"why":"Partial pooling of small-area rates vs over-dispersed funnel limits for institutional comparison.","m":"agent draft (gemini)"},{"id":"e371","i":"companion_wtlp1_2026--COMPLEMENTS-->shah2026logjack","s":"vidigal2026wtllakehouse","t":"shah2026logjack","r":"COMPLEMENTS","w":2,"why":"Pairs compile-time AST log-lakehouse ingestion with LogJack's defense against indirect prompt-injection attacks embedded in untrusted log payloads.","m":"agent draft (qwen)"},{"id":"e372","i":"companion_wtlp1_2026--CONTRASTS_WITH-->he2017drain","s":"vidigal2026wtllakehouse","t":"he2017drain","r":"CONTRASTS_WITH","w":2,"why":"Extracts exact log templates from source-code ASTs at compile time and matches via Aho-Corasick and radix tries rather than Drain's online token-similarity heuristics.","m":"agent draft (gemini)"},{"id":"e373","i":"companion_wtlp1_2026--EXTENDS-->rodrigues2021clp","s":"vidigal2026wtllakehouse","t":"rodrigues2021clp","r":"EXTENDS","w":2,"why":"Extends CLP's static logtype and variable decomposition by folding consecutive in-window bursts via two-tier statistical temporal run-length encoding into Apache Parquet.","m":"agent draft (gemini)"},{"id":"e374","i":"cook1998complex--CONTRASTS_WITH-->card-2016","s":"cook1998complex","t":"card-2016","r":"CONTRASTS_WITH","w":2,"why":"Cook rejects single root cause attribution outright; the 5 Whys presumes a single linear chain.","m":"agent draft (gemini)"},{"id":"e375","i":"correll-2014--COMPLEMENTS-->padilla-2021","s":"correll-2014","t":"padilla-2021","r":"COMPLEMENTS","w":2,"why":"alternatives to error bars (gradients, violins) reviewed in uncertainty chapter","m":"agent draft (claude)"},{"id":"e376","i":"correll-2020--CITES-->correll-2014","s":"correll-2020","t":"correll-2014","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3313831.3376222 -> 10.1109/tvcg.2014.2346298)","m":"reference-list extractor"},{"id":"e377","i":"correll-2020--CITES-->kong2018titles","s":"correll-2020","t":"kong2018titles","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3313831.3376222 -> 10.1145/3173574.3174012)","m":"reference-list extractor"},{"id":"e378","i":"correll-2020--COMPLEMENTS-->isenberg2011dualscale","s":"correll-2020","t":"isenberg2011dualscale","r":"COMPLEMENTS","w":2,"why":"axis-scale manipulations and perceived effect size","m":"agent draft (claude)"},{"id":"e380","i":"costa-2021--CITES-->blackburn-2006","s":"costa-2021","t":"blackburn-2006","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2019.2925345 -> 10.1145/1167473.1167488)","m":"reference-list extractor"},{"id":"e381","i":"costa-2021--CITES-->curtsinger-2013","s":"costa-2021","t":"curtsinger-2013","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2019.2925345 -> 10.1145/2451116.2451141)","m":"reference-list extractor"},{"id":"e382","i":"costa-2021--CITES-->mytkowicz2009wrong","s":"costa-2021","t":"mytkowicz2009wrong","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2019.2925345 -> 10.1145/1508244.1508275)","m":"reference-list extractor"},{"id":"e383","i":"costa-2021--COMPLEMENTS-->georges2007statistically","s":"costa-2021","t":"georges2007statistically","r":"COMPLEMENTS","w":2,"why":"Georges gives the statistical protocol for Java measurements; Costa catalogues the JMH coding mistakes that bias the samples before statistics apply.","m":"agent draft (claude)"},{"id":"e384","i":"cox-1972--CITES-->kaplan-1958","s":"cox-1972","t":"kaplan-1958","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.2517-6161.1972.tb00899.x -> 10.1080/01621459.1958.10501452)","m":"reference-list extractor"},{"id":"e386","i":"cox2015libyear--CITES-->parnas1972criteria","s":"cox2015libyear","t":"parnas1972criteria","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse.2015.140 -> 10.1145/361598.361623)","m":"reference-list extractor"},{"id":"e387","i":"cox2015libyear--COMPLEMENTS-->zerouali2019techlagframework","s":"cox2015libyear","t":"zerouali2019techlagframework","r":"COMPLEMENTS","w":2,"why":"libyear time-distance is one of the lag measures the formal framework generalises","m":"agent draft (gemini)"},{"id":"e389","i":"crdb2026diskstall--CONTRASTS_WITH-->etcd2026hardware","s":"crdb2026diskstall","t":"etcd2026hardware","r":"CONTRASTS_WITH","w":2,"why":"CockroachDB kills the node on stall; etcd relies on election via heartbeat timeouts.","m":"agent draft (claude)"},{"id":"e390","i":"crdb2026diskstall--CONTRASTS_WITH-->gunawi-2018","s":"crdb2026diskstall","t":"gunawi2018failslow","r":"CONTRASTS_WITH","w":2,"why":"Vendor fail-stop conversion (self-terminate on long WAL sync) as a response to fail-slow disks","m":"agent draft (gemini)"},{"id":"e391","i":"crdb2026diskstall--CONTRASTS_WITH-->k8s2026probes","s":"crdb2026diskstall","t":"k8s2026probes","r":"CONTRASTS_WITH","w":2,"why":"Bounded in-process stall policy vs generic liveness probe that misses grey stalls","m":"agent draft (gemini)"},{"id":"e392","i":"crdb2026diskstall--OPERATIONALIZES-->gunawi2018failslow","s":"crdb2026diskstall","t":"gunawi2018failslow","r":"OPERATIONALIZES","w":2,"why":"Vendor policy that bounds fail-slow disk impact by terminating the node at 20 s.","m":"agent draft (claude)"},{"id":"e393","i":"crdb2026loadsplit--OPERATIONALIZES-->taft2020cockroachdb","s":"crdb2026loadsplit","t":"taft2020cockroachdb","r":"OPERATIONALIZES","w":2,"why":"QPS-threshold load-based range splitting","m":"agent draft (gemini)"},{"id":"e394","i":"crdb2026walfailover--CONTRASTS_WITH-->tikv2026slowstore","s":"crdb2026walfailover","t":"tikv2026slowstore","r":"CONTRASTS_WITH","w":2,"why":"Per-stall magnitude trigger (crash after 20 s) vs time-integrated slow score with leader eviction.","m":"agent draft (claude)"},{"id":"e399","i":"cui2025genaidevs--CITES-->dellacqua2023jagged","s":"cui2025genaidevs","t":"dellacqua2023jagged","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1287/mnsc.2025.00535 -> 10.2139/ssrn.4573321)","m":"reference-list extractor"},{"id":"e401","i":"cui2025genaidevs--CONTRASTS_WITH-->joel-2025-arxiv","s":"cui2025genaidevs","t":"joel-2025-arxiv","r":"CONTRASTS_WITH","w":2,"why":"RCT +26% tasks (mixed experience) vs RCT -19% speed for experienced OSS developers","m":"agent draft (gemini)"},{"id":"e402","i":"curtsinger-2013--CITES-->georges2007statistically","s":"curtsinger-2013","t":"georges2007statistically","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2451116.2451141 -> 10.1145/1297027.1297033)","m":"reference-list extractor"},{"id":"e403","i":"curtsinger-2013--EXTENDS-->mytkowicz2009wrong","s":"curtsinger-2013","t":"mytkowicz2009wrong","r":"EXTENDS","w":2,"why":"Stabilizer operationalizes the layout-bias fix by runtime randomization","m":"agent draft (gemini)"},{"id":"e404","i":"curtsinger-2015--CITES-->curtsinger-2013","s":"curtsinger-2015","t":"curtsinger-2013","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2815400.2815409 -> 10.1145/2451116.2451141)","m":"reference-list extractor"},{"id":"e405","i":"curtsinger-2015--CONTRASTS_WITH-->mytkowicz-2010","s":"curtsinger-2015","t":"mytkowicz-2010","r":"CONTRASTS_WITH","w":2,"why":"Mytkowicz shows sampling profilers disagree on hot methods; Coz replaces hotness with causal impact via virtual speedups","m":"agent draft (gemini)"},{"id":"e406","i":"curtsinger-2015--CONTRASTS_WITH-->ren-2010","s":"curtsinger-2015","t":"ren-2010","r":"CONTRASTS_WITH","w":2,"why":"Coz argues sampled CPU-time profiles (as in GWP) misidentify optimization targets in concurrent code; proposes causal (virtual speedup) profiling instead.","m":"agent draft (gemini)"},{"id":"e407","i":"curtsinger-2015--EXTENDS-->mytkowicz-2010","s":"curtsinger-2015","t":"mytkowicz-2010","r":"EXTENDS","w":2,"why":"Coz responds to profiler inaccuracy/misattribution by measuring causal impact of speedups rather than time share","m":"agent draft (gemini)"},{"id":"e409","i":"cutler2024cedar--USES_METHOD-->mckeeman1998differential","s":"cutler2024cedar","t":"mckeeman1998differential","r":"USES_METHOD","w":2,"why":"Cedar bridges its verified Lean 4 model and production Rust engine via millions of differential testing inputs.","m":"agent draft (gemini)"},{"id":"e410","i":"cutler2024cedar--USES_METHOD-->moura2021lean4","s":"cutler2024cedar","t":"moura2021lean4","r":"USES_METHOD","w":2,"why":"Cedar formalizes its authorization policy evaluator and type validator in Lean 4 and proves soundness theorems in the Lean 4 kernel.","m":"agent draft (gemini)"},{"id":"e412","i":"dabrowski2022appreviews--CITES-->guzman2014feature","s":"dabrowski2022appreviews","t":"guzman2014feature","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s10664-021-10065-7 -> 10.1109/re.2014.6912257)","m":"reference-list extractor"},{"id":"e414","i":"dabrowski2022appreviews--CITES-->pagano-2013","s":"dabrowski2022appreviews","t":"pagano-2013","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s10664-021-10065-7 -> 10.1109/re.2013.6636712)","m":"reference-list extractor"},{"id":"e415","i":"dabrowski2022appreviews--COMPLEMENTS-->maalej-2016","s":"dabrowski2022appreviews","t":"maalej-2016","r":"COMPLEMENTS","w":2,"why":"Systematic review of the feedback stream data-driven RE relies on.","m":"agent draft (gemini)"},{"id":"e417","i":"dai2020logram--CITES-->huang2018panorama","s":"dai2020logram","t":"huang2018panorama","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2020.3007554 -> huang2018panorama)","m":"reference-list extractor"},{"id":"e418","i":"dai2020logram--CITES-->nagaraj2012distalyzer","s":"dai2020logram","t":"nagaraj2012distalyzer","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2020.3007554 -> nagaraj2012distalyzer)","m":"reference-list extractor"},{"id":"e419","i":"dai2020logram--CITES-->yuan2012characterizing","s":"dai2020logram","t":"yuan2012characterizing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2020.3007554 -> 10.1109/icse.2012.6227202)","m":"reference-list extractor"},{"id":"e420","i":"dai2020logram--CITES-->zhu2019tools","s":"dai2020logram","t":"zhu2019tools","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2020.3007554 -> 10.1109/icse-seip.2019.00021)","m":"reference-list extractor"},{"id":"e421","i":"dai2020logram--CONTRASTS_WITH-->he2017drain","s":"dai2020logram","t":"he2017drain","r":"CONTRASTS_WITH","w":2,"why":"Replaces Drain's prefix-tree traversal with 2-gram and 3-gram corpus frequency dictionaries to separate frequent static template tokens from infrequent parameters.","m":"agent draft (gemini)"},{"id":"e422","i":"dang2012rebucket--COMPLEMENTS-->glerum2009wer","s":"dang2012rebucket","t":"glerum2009wer","r":"COMPLEMENTS","w":2,"why":"Pairs call-stack similarity clustering (ReBucket) with Windows Error Reporting's metadata bucketing to eliminate duplicate crash buckets at fleet scale.","m":"agent draft (claude)"},{"id":"e423","i":"dang2012rebucket--COMPLEMENTS-->kim2011topcrashes","s":"dang2012rebucket","t":"kim2011topcrashes","r":"COMPLEMENTS","w":2,"why":"Combines call-stack position-dependent alignment for crash deduplication with early prediction of top crashers before full release rollout.","m":"agent draft (gemini)"},{"id":"e424","i":"dang2012rebucket--EXTENDS-->brodie2005symptom","s":"dang2012rebucket","t":"brodie2005symptom","r":"EXTENDS","w":2,"why":"Refines Brodie et al.'s biological sequence alignment of call stacks by introducing both frame-depth decay and offset-shift penalty functions tailored to crash stack traces.","m":"agent draft (gemini)"},{"id":"e425","i":"dang2012rebucket--EXTENDS-->glerum2009wer","s":"dang2012rebucket","t":"glerum2009wer","r":"EXTENDS","w":2,"why":"call-stack similarity clustering refines WER-style crash bucketing","m":"agent draft (gemini)"},{"id":"e427","i":"david-2018-arxiv--EXTENDS-->goodhart-1984","s":"david-2018-arxiv","t":"goodhart-1984","r":"EXTENDS","w":2,"why":"Manheim and Garrabrant categorise the mechanisms behind Goodhart's original observation.","m":"agent draft (claude)"},{"id":"e428","i":"david-2018-arxiv--EXTENDS-->strathern-1997","s":"david-2018-arxiv","t":"strathern-1997","r":"EXTENDS","w":2,"why":"Manheim and Garrabrant formalize variants of the Goodhart/Strathern measure-as-target effect.","m":"agent draft (claude)"},{"id":"e435","i":"davis-2024--CITES-->fernandes2018quantiledotplots","s":"davis-2024","t":"fernandes2018quantiledotplots","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2022.3226463 -> 10.1145/3173574.3173718)","m":"reference-list extractor"},{"id":"e436","i":"davis-2024--CITES-->kale-2019","s":"davis-2024","t":"kale-2019","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2022.3226463 -> 10.1109/tvcg.2018.2864909)","m":"reference-list extractor"},{"id":"e437","i":"davis-2024--CITES-->munzner2014vad","s":"davis-2024","t":"munzner2014vad","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2022.3226463 -> 10.1201/b17511)","m":"reference-list extractor"},{"id":"e438","i":"davis-2024--CONTRASTS_WITH-->heer2010crowdsourcing","s":"davis-2024","t":"heer2010crowdsourcing","r":"CONTRASTS_WITH","w":2,"why":"shows average-observer rankings hide individual deviations","m":"agent draft (claude)"},{"id":"e439","i":"davis-2024--EXTENDS-->cleveland1984graphical","s":"davis-2024","t":"cleveland1984graphical","r":"EXTENDS","w":2,"why":"replicates C&M with Bayesian multilevel models; models individual differences","m":"agent draft (claude)"},{"id":"e440","i":"davis2019rethinking--CITES-->davis2018redos","s":"davis2019rethinking","t":"davis2018redos","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3338906.3342509 -> 10.1145/3236024.3236027)","m":"reference-list extractor"},{"id":"e442","i":"davis2019rethinking--EXTENDS-->thompson1968nfa","s":"davis2019rethinking","t":"thompson1968nfa","r":"EXTENDS","w":2,"why":"Applies Thompson NFA/DFA linear-time simulation to eliminate ReDoS across 95% of production regular expressions.","m":"agent draft (gemini)"},{"id":"e443","i":"dawson2007referees--COMPLEMENTS-->boyko2007referee","s":"dawson2007referees","t":"boyko2007referee","r":"COMPLEMENTS","w":2,"why":"Models yellow and red card disciplinary incidence via bivariate Poisson regression, complementing Boyko et al.'s fixed-effects separation of referee bias from home-crowd advantage.","m":"agent draft (gemini)"},{"id":"e444","i":"dayan-2017--CITES-->bloom1970space","s":"dayan-2017","t":"bloom1970space","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3035918.3064054 -> 10.1145/362686.362692)","m":"reference-list extractor"},{"id":"e445","i":"dayan-2017--CITES-->o’neil-1996","s":"dayan-2017","t":"o’neil-1996","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3035918.3064054 -> 10.1007/s002360050048)","m":"reference-list extractor"},{"id":"e446","i":"dayan-2017--OPERATIONALIZES-->athanassoulis-2016","s":"dayan-2017","t":"athanassoulis-2016","r":"OPERATIONALIZES","w":2,"why":"Navigates the RUM tradeoff via optimal per-level Bloom filter bit allocation (Monkey)","m":"agent draft (gemini)"},{"id":"e447","i":"dayan-2018--CITES-->bloom1970space","s":"dayan-2018","t":"bloom1970space","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3183713.3196927 -> 10.1145/362686.362692)","m":"reference-list extractor"},{"id":"e448","i":"dayan-2018--CITES-->o’neil-1996","s":"dayan-2018","t":"o’neil-1996","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3183713.3196927 -> 10.1007/s002360050048)","m":"reference-list extractor"},{"id":"e449","i":"dayan-2018--EXTENDS-->dayan-2017","s":"dayan-2018","t":"dayan-2017","r":"EXTENDS","w":2,"why":"Adds Fluid LSM-tree (Lazy Leveling) to Monkey's Bloom filter model to remove superfluous L_max merging","m":"agent draft (gemini)"},{"id":"e450","i":"dayan-2019--CITES-->dayan-2017","s":"dayan-2019","t":"dayan-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3299869.3319903 -> 10.1145/3035918.3064054)","m":"reference-list extractor"},{"id":"e451","i":"dayan-2019--EXTENDS-->dayan-2018","s":"dayan-2019","t":"dayan-2018","r":"EXTENDS","w":2,"why":"Extends Dostoevsky into the Wacky continuum with doubly-exponential capacity ratios (LSM-Bush)","m":"agent draft (gemini)"},{"id":"e452","i":"dayan-2022--CITES-->bloom1970space","s":"dayan-2022","t":"bloom1970space","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3551793.3551853 -> 10.1145/362686.362692)","m":"reference-list extractor"},{"id":"e453","i":"dayan-2022--CITES-->dayan-2017","s":"dayan-2022","t":"dayan-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3551793.3551853 -> 10.1145/3035918.3064054)","m":"reference-list extractor"},{"id":"e454","i":"dayan-2022--CITES-->dayan-2019","s":"dayan-2022","t":"dayan-2019","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3551793.3551853 -> 10.1145/3299869.3319903)","m":"reference-list extractor"},{"id":"e455","i":"dayan-2022--CITES-->dong-2021","s":"dayan-2022","t":"dong-2021","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3551793.3551853 -> 10.1145/3483840)","m":"reference-list extractor"},{"id":"e456","i":"dayan-2022--CITES-->o’neil-1996","s":"dayan-2022","t":"o’neil-1996","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3551793.3551853 -> 10.1007/s002360050048)","m":"reference-list extractor"},{"id":"e457","i":"dayan-2022--EXTENDS-->dayan-2018","s":"dayan-2022","t":"dayan-2018","r":"EXTENDS","w":2,"why":"Extends LSM compaction theory to partitioned SST files with boundary-aligned Spooky granulation","m":"agent draft (gemini)"},{"id":"e463","i":"dean-2013--COMPLEMENTS-->lozi-2016","s":"dean-2013","t":"lozi-2016","r":"COMPLEMENTS","w":2,"why":"Scheduler pathologies such as idle cores while threads wait (Lozi) are one host-level source of the per-server latency variability that fan-out turns into tail latency (Dean & Barroso).","m":"agent draft (claude)"},{"id":"e465","i":"decan2021semver--EXTENDS-->raemaekers2017semver","s":"decan2021semver","t":"raemaekers2017semver","r":"EXTENDS","w":2,"why":"Extends SemVer compliance analysis to Cargo and three other ecosystems, including 0.y.z dependency constraint dynamics.","m":"agent draft (gemini)"},{"id":"e467","i":"deep2020workloadcompression--CONTRASTS_WITH-->kul2018sqlclustering","s":"deep2020workloadcompression","t":"kul2018sqlclustering","r":"CONTRASTS_WITH","w":2,"why":"Selection with coverage guarantees vs similarity-metric clustering","m":"agent draft (claude)"},{"id":"e468","i":"deep2020workloadcompression--EXTENDS-->chaudhuri2002compressing","s":"deep2020workloadcompression","t":"chaudhuri2002compressing","r":"EXTENDS","w":2,"why":"Generalizes workload compression with representativity and coverage guarantees","m":"agent draft (claude)"},{"id":"e470","i":"defraeye-2016--CITES-->aksin2007moderncallcenter","s":"defraeye2016staffing","t":"aksin2007moderncallcenter","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.omega.2015.04.002 -> 10.1111/j.1937-5956.2007.tb00288.x)","m":"reference-list extractor"},{"id":"e471","i":"defraeye-2016--CITES-->borst-2004","s":"defraeye2016staffing","t":"borst-2004","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.omega.2015.04.002 -> 10.1287/opre.1030.0081)","m":"reference-list extractor"},{"id":"e472","i":"defraeye-2016--CITES-->brown-2005","s":"defraeye2016staffing","t":"brown-2005","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.omega.2015.04.002 -> 10.1198/016214504000001808)","m":"reference-list extractor"},{"id":"e473","i":"defraeye-2016--CITES-->gans-2003","s":"defraeye2016staffing","t":"gans-2003","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.omega.2015.04.002 -> 10.1287/msom.5.2.79.16071)","m":"reference-list extractor"},{"id":"e474","i":"defraeye-2016--CITES-->halfin-1981","s":"defraeye2016staffing","t":"halfin-1981","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.omega.2015.04.002 -> 10.1287/opre.29.3.567)","m":"reference-list extractor"},{"id":"e476","i":"defraeye2016staffing--EXTENDS-->green-2007","s":"defraeye2016staffing","t":"green-2007","r":"EXTENDS","w":2,"why":"Review of nonstationary staffing extends Green, Kolesar & Whitt's treatment of time-varying demand.","m":"agent draft (claude)"},{"id":"e477","i":"dehghani-2022--EXTENDS-->khatri2010designing","s":"dehghani-2022","t":"khatri2010designing","r":"EXTENDS","w":3,"why":"Decentralizes Khatri & Brown's IT decision-rights matrix into domain-oriented data ownership, data-as-a-product SLAs, and federated computational governance.","m":"agent draft (gemini)"},{"id":"e478","i":"dejong2017zerodowntime--CITES-->qiu2013schemacoevolution","s":"dejong2017zerodowntime","t":"qiu2013schemacoevolution","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse-seip.2017.5 -> 10.1145/2491411.2491431)","m":"reference-list extractor"},{"id":"e479","i":"dejong2017zerodowntime--OPERATIONALIZES-->curino2008prism","s":"dejong2017zerodowntime","t":"curino2008prism","r":"OPERATIONALIZES","w":2,"why":"Zero-downtime schema evolution for continuous deployment operationalizes graceful evolution","m":"agent draft (claude)"},{"id":"e483","i":"dellacqua2023jagged--COMPLEMENTS-->brynjolfsson-2025","s":"dellacqua2023jagged","t":"brynjolfsson-2025","r":"COMPLEMENTS","w":2,"why":"Both field experiments show GenAI gains concentrated on some workers/tasks with quality risks elsewhere (experienced agents; tasks outside the frontier).","m":"agent draft (claude)"},{"id":"e484","i":"deng2016metricdev--EXTENDS-->deng2013cuped","s":"deng2016metricdev","t":"deng2013cuped","r":"EXTENDS","w":2,"why":"Sensitivity framework generalizes CUPED's variance-reduction lever to metric design","m":"agent draft (gemini)"},{"id":"e485","i":"devisser2020longitudinal--CITES-->goddard-2012","s":"devisser2020longitudinal","t":"goddard-2012","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s12369-019-00596-x -> 10.1136/amiajnl-2011-000089)","m":"reference-list extractor"},{"id":"e486","i":"devisser2020longitudinal--CITES-->hoff2015trust","s":"devisser2020longitudinal","t":"hoff2015trust","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s12369-019-00596-x -> 10.1177/0018720814547570)","m":"reference-list extractor"},{"id":"e489","i":"devisser2020longitudinal--CITES-->parasuraman2010complacency","s":"devisser2020longitudinal","t":"parasuraman2010complacency","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s12369-019-00596-x -> 10.1177/0018720810376055)","m":"reference-list extractor"},{"id":"e490","i":"devisser2020longitudinal--EXTENDS-->lee-2004","s":"devisser2020longitudinal","t":"lee-2004","r":"EXTENDS","w":2,"why":"Adds trust violation and repair dynamics over time to Lee and See's calibration framework","m":"agent draft (gemini)"},{"id":"e491","i":"dey2020usageindependent--EXTENDS-->mockus2005customerquality","s":"dey2020usageindependent","t":"mockus2005customerquality","r":"EXTENDS","w":2,"why":"usage-independent quality metric builds on exposure-normalised customer quality","m":"agent draft (gemini)"},{"id":"e492","i":"dhaliwal2011crash--CITES-->bettenburg2008goodbug","s":"dhaliwal2011crash","t":"bettenburg2008goodbug","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icsm.2011.6080800 -> 10.1145/1453101.1453146)","m":"reference-list extractor"},{"id":"e493","i":"dhaliwal2011crash--CITES-->brodie2005symptom","s":"dhaliwal2011crash","t":"brodie2005symptom","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icsm.2011.6080800 -> 10.1109/icac.2005.49)","m":"reference-list extractor"},{"id":"e494","i":"dhaliwal2011crash--CITES-->kim2011crashgraphs","s":"dhaliwal2011crash","t":"kim2011crashgraphs","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icsm.2011.6080800 -> 10.1109/dsn.2011.5958261)","m":"reference-list extractor"},{"id":"e495","i":"dhaliwal2011crash--COMPLEMENTS-->glerum2009wer","s":"dhaliwal2011crash","t":"glerum2009wer","r":"COMPLEMENTS","w":2,"why":"Evaluates Mozilla Socorro field crash buckets and demonstrates that top-frame bucketing must be paired with top-10 call stack Levenshtein distance sub-clustering to separate distinct bugs crashing in shared utility frames.","m":"agent draft (gemini)"},{"id":"e496","i":"dibia-2023--COMPLEMENTS-->chen-2025","s":"dibia-2023","t":"chen-2025","r":"COMPLEMENTS","w":2,"why":"LLM visualization generation system vs evaluation benchmark","m":"agent draft (claude)"},{"id":"e497","i":"dietvorst2015aversion--COMPLEMENTS-->lee-2004","s":"dietvorst2015aversion","t":"lee-2004","r":"COMPLEMENTS","w":2,"why":"Aversion after observed error is the disuse side of Lee & See's calibration framework","m":"agent draft (gemini)"},{"id":"e504","i":"ding-2014--COMPLEMENTS-->chen-2020","s":"ding-2014","t":"chen2020lidar","r":"COMPLEMENTS","w":2,"why":"Historical-incident retrieval for healing actions vs linking related concurrent incidents; both Microsoft incident-management work.","m":"agent draft (gemini)"},{"id":"e505","i":"ding-2014--COMPLEMENTS-->lin2016logcluster","s":"ding-2014","t":"lin2016logcluster","r":"COMPLEMENTS","w":2,"why":"Ding et al. mine historical issue repositories to reuse past healing actions; LogCluster identifies whether the current problem is a recurring one. Together: match then reuse.","m":"agent draft (claude)"},{"id":"e507","i":"ding2020desensitization--COMPLEMENTS-->glerum2009wer","s":"ding2020desensitization","t":"glerum2009wer","r":"COMPLEMENTS","w":2,"why":"Provides formal privacy desensitization guarantees for crash artifacts before they enter WER-style stack bucketing pipelines.","m":"agent draft (gemini)"},{"id":"e508","i":"ding2020desensitization--EXTENDS-->castro2008bugreporting","s":"ding2020desensitization","t":"castro2008bugreporting","r":"EXTENDS","w":2,"why":"Extends privacy-preserving bug reporting to binary crash reports, sanitizing user-supplied inputs while preserving exact crash bucketability and fault instructions.","m":"agent draft (gemini)"},{"id":"e509","i":"dintyala2020sqlcheck--COMPLEMENTS-->patel2024acorn","s":"dintyala2020sqlcheck","t":"patel2024acorn","r":"COMPLEMENTS","w":3,"why":"Combines static SQL/DDL anti-pattern detection (SQLCheck) with predicate-agnostic selectivity and filter pushdown in vector and relational indexes (ACORN).","m":"agent draft (gemini)"},{"id":"e510","i":"dmitriev2016measuringmetrics--CITES-->deng2013cuped","s":"dmitriev2016measuringmetrics","t":"deng2013cuped","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2983323.2983356 -> 10.1145/2433396.2433413)","m":"reference-list extractor"},{"id":"e512","i":"dmitriev2016measuringmetrics--CITES-->kerr1975folly","s":"dmitriev2016measuringmetrics","t":"kerr1975folly","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2983323.2983356 -> kerr1975folly)","m":"reference-list extractor"},{"id":"e513","i":"dmitriev2016measuringmetrics--COMPLEMENTS-->deng2016metricdev","s":"dmitriev2016measuringmetrics","t":"deng2016metricdev","r":"COMPLEMENTS","w":2,"why":"Metric validation: directionality and sensitivity","m":"agent draft (gemini)"},{"id":"e516","i":"dohmen2008social--EXTENDS-->boyko2007referee","s":"dohmen2008social","t":"boyko2007referee","r":"EXTENDS","w":2,"why":"Isolates stadium architecture (running track vs steep stands) and crowd proximity as the causal mechanism driving referee stoppage-time and foul-call asymmetry.","m":"agent draft (gemini)"},{"id":"e517","i":"dong-2021--OPERATIONALIZES-->o’neil-1996","s":"dong-2021","t":"o’neil-1996","r":"OPERATIONALIZES","w":2,"why":"RocksDB is a production LSM-tree implementation.","m":"agent draft (gemini)"},{"id":"e521","i":"doshi1986vacations--EXTENDS-->gaver1962interrupted","s":"doshi1986vacations","t":"gaver1962interrupted","r":"EXTENDS","w":2,"why":"Vacation models generalize Gaver's interrupted-service queue; both give wait terms driven by the second moment of the interruption length.","m":"agent draft (claude)"},{"id":"e524","i":"drupsteen2014learning--COMPLEMENTS-->peerally-2016","s":"drupsteen2014learning","t":"peerally-2016","r":"COMPLEMENTS","w":2,"why":"Both argue RCA programmes stop at analysis and never verify the action","m":"agent draft (gemini)"},{"id":"e526","i":"du2016spell--CITES-->nagaraj2012distalyzer","s":"du2016spell","t":"nagaraj2012distalyzer","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icdm.2016.0103 -> nagaraj2012distalyzer)","m":"reference-list extractor"},{"id":"e528","i":"duan-1983-demand--CITES-->heckman1979selection","s":"duan-1983-demand","t":"heckman1979selection","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1080/07350015.1983.10509330 -> 10.2307/1912352)","m":"reference-list extractor"},{"id":"e529","i":"duan-1983-demand--EXTENDS-->cragg-1971","s":"duan-1983-demand","t":"cragg-1971","r":"EXTENDS","w":2,"why":"The RAND two-part model generalises Cragg's two-tier limited-dependent-variable model to medical spending.","m":"agent draft (claude)"},{"id":"e530","i":"duan-1983-demand--USES_METHOD-->duan-1983-smearing","s":"duan-1983-demand","t":"duan-1983-smearing","r":"USES_METHOD","w":2,"why":"The second part on the log scale is retransformed with the smearing estimator.","m":"agent draft (claude)"},{"id":"e534","i":"durner2023cloud--CITES-->armbrust-2020","s":"durner2023cloud","t":"armbrust-2020","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3611479.3611486 -> 10.14778/3415478.3415560)","m":"reference-list extractor"},{"id":"e535","i":"durner2023cloud--CITES-->dean-2013","s":"durner2023cloud","t":"dean-2013","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3611479.3611486 -> 10.1145/2408776.2408794)","m":"reference-list extractor"},{"id":"e536","i":"durner2023cloud--CITES-->leis-2014","s":"durner2023cloud","t":"leis-2014","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3611479.3611486 -> 10.1145/2588555.2610507)","m":"reference-list extractor"},{"id":"e537","i":"durner2023cloud--CITES-->melnik-2010","s":"durner2023cloud","t":"melnik-2010","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3611479.3611486 -> 10.14778/1920841.1920886)","m":"reference-list extractor"},{"id":"e539","i":"durner2023cloud--COMPLEMENTS-->zeng2023columnar","s":"durner2023cloud","t":"zeng2023columnar","r":"COMPLEMENTS","w":2,"why":"While Zeng et al. evaluate internal columnar encoding and CPU decoding trade-offs in Parquet/ORC, Durner et al. model the network/HTTP request latency, 8-16 MiB range-size sweet spot, and range-read coalescing required to serve those columnar formats efficiently from cloud object storage (S3/GCS).","m":"agent draft (gemini)"},{"id":"e540","i":"dwork2006calibrating--CITES-->sweeney2002kanonymity","s":"dwork2006calibrating","t":"sweeney2002kanonymity","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/11681878_14 -> 10.1142/s0218488502001648)","m":"reference-list extractor"},{"id":"e541","i":"dwork2014algorithmic--CITES-->dinur2003revealing","s":"dwork2014algorithmic","t":"dinur2003revealing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1561/0400000042 -> dinur2003revealing)","m":"reference-list extractor"},{"id":"e542","i":"dwork2014algorithmic--EXTENDS-->dwork2006calibrating","s":"dwork2014algorithmic","t":"dwork2006calibrating","r":"EXTENDS","w":2,"why":"Consolidates sensitivity-calibrated Laplace noise into the compositional theory of (ε, δ)-differential privacy, privacy loss random variables, and mechanism design.","m":"agent draft (gemini)"},{"id":"e544","i":"dzindolet2003reliance--COMPLEMENTS-->dietvorst2015aversion","s":"dzindolet2003reliance","t":"dietvorst2015aversion","r":"COMPLEMENTS","w":2,"why":"Both show reliance collapses after observed errors; Dzindolet adds that explaining why errors occur restores reliance","m":"agent draft (gemini)"},{"id":"e549","i":"elavsky-2022--OPERATIONALIZES-->lundgard-2022","s":"elavsky-2022","t":"lundgard-2022","r":"OPERATIONALIZES","w":2,"why":"audit heuristics incorporate text-alternative guidance","m":"agent draft (claude)"},{"id":"e550","i":"elavsky-2022--OPERATIONALIZES-->w3c2024wcag22","s":"elavsky-2022","t":"w3c2024wcag22","r":"OPERATIONALIZES","w":2,"why":"Chartability turns WCAG POUR into chart-specific heuristics","m":"agent draft (gemini)"},{"id":"e551","i":"elavsky-2023--CITES-->elavsky-2022","s":"elavsky-2023","t":"elavsky-2022","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2023.3327393 -> 10.1111/cgf.14522)","m":"reference-list extractor"},{"id":"e552","i":"elavsky-2023--CITES-->lundgard-2022","s":"elavsky-2023","t":"lundgard-2022","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2023.3327393 -> 10.1109/tvcg.2021.3114770)","m":"reference-list extractor"},{"id":"e553","i":"elavsky-2023--CITES-->satyanarayan-2017","s":"elavsky-2023","t":"satyanarayan-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2023.3327393 -> 10.1109/tvcg.2016.2599030)","m":"reference-list extractor"},{"id":"e554","i":"elavsky-2023--EXTENDS-->zong-2022","s":"elavsky-2023","t":"zong-2022","r":"EXTENDS","w":2,"why":"toolkit generalising structured screen-reader navigation of charts","m":"agent draft (claude)"},{"id":"e555","i":"elliott2017nonprob--CITES-->rubin1976missing","s":"elliott2017nonprob","t":"rubin1976missing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1214/16-sts598 -> 10.1093/biomet/63.3.581)","m":"reference-list extractor"},{"id":"e556","i":"elliott2017nonprob--EXTENDS-->heckman1979selection","s":"elliott2017nonprob","t":"heckman1979selection","r":"EXTENDS","w":2,"why":"Extends classical two-step selection correction to propensity-weighted inference combining non-probability convenience samples with probability reference surveys.","m":"agent draft (gemini)"},{"id":"e559","i":"en-18286--COMPLEMENTS-->iso-42001","s":"en-18286","t":"iso-42001","r":"COMPLEMENTS","w":3,"why":"EN 18286:2026 layers product-centric EU AI Act Article 17 QMS conformity (Annex ZA -> Arts. 11, 17, 72) onto organization-centric ISO/IEC 42001 (mapped in Annex D).","m":"agent draft (gemini)"},{"id":"e560","i":"en-18286--OPERATIONALIZES-->euaiact2024","s":"en-18286","t":"euaiact2024","r":"OPERATIONALIZES","w":3,"why":"EN 18286:2026 (Commission Implementing Decision (EU) 2026/2274) is the harmonised standard providing presumption of conformity for EU AI Act Article 17.","m":"agent draft (gemini)"},{"id":"e561","i":"enders-2001--OPERATIONALIZES-->rubin1976missing","s":"enders-2001","t":"rubin1976missing","r":"OPERATIONALIZES","w":2,"why":"FIML is unbiased under MAR as defined by Rubin; used for unread Slack channels.","m":"agent draft (claude)"},{"id":"e562","i":"es2024ragas--COMPLEMENTS-->min2023factscore","s":"es2024ragas","t":"min2023factscore","r":"COMPLEMENTS","w":2,"why":"Both decompose generations into statements and score support; RAGAS against retrieved context, FActScore against a knowledge source.","m":"agent draft (claude)"},{"id":"e564","i":"fang2022onlineschemaevo--CITES-->berenson-1995","s":"fang2022onlineschemaevo","t":"berenson-1995","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3565816.3565818 -> 10.1145/223784.223785)","m":"reference-list extractor"},{"id":"e565","i":"fang2022onlineschemaevo--CITES-->qiu2013schemacoevolution","s":"fang2022onlineschemaevo","t":"qiu2013schemacoevolution","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3565816.3565818 -> 10.1145/2491411.2491431)","m":"reference-list extractor"},{"id":"e566","i":"fang2022onlineschemaevo--COMPLEMENTS-->rae2013f1","s":"fang2022onlineschemaevo","t":"rae2013f1","r":"COMPLEMENTS","w":2,"why":"Snapshot-based online schema evolution vs F1 intermediate states","m":"agent draft (claude)"},{"id":"e569","i":"farrington-1996--EXTENDS-->page-1954","s":"farrington-1996","t":"page-1954","r":"EXTENDS","w":2,"why":"Extends Page's sequential change-detection framework to quasi-Poisson regression over seasonal, trend-drifting public-health surveillance counts.","m":"agent draft (claude)"},{"id":"e570","i":"farrington2009sccsevent--CITES-->rubin1976missing","s":"farrington2009sccsevent","t":"rubin1976missing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1093/biostatistics/kxn013 -> 10.1093/biomet/63.3.581)","m":"reference-list extractor"},{"id":"e573","i":"fernandes2018quantiledotplots--CITES-->cleveland1984graphical","s":"fernandes2018quantiledotplots","t":"cleveland1984graphical","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3173574.3173718 -> 10.1080/01621459.1984.10478080)","m":"reference-list extractor"},{"id":"e574","i":"fernandes2018quantiledotplots--EXTENDS-->kay-2016","s":"fernandes2018quantiledotplots","t":"kay-2016","r":"EXTENDS","w":2,"why":"quantile dotplots for transit decisions, building on bus arrival work","m":"agent draft (claude)"},{"id":"e575","i":"fernandes2018quantiledotplots--OPERATIONALIZES-->hullman2019pursuiterror","s":"fernandes2018quantiledotplots","t":"hullman2019pursuiterror","r":"COMPLEMENTS","w":2,"why":"CDF and quantile-dotplot displays are a tested form of the distributional uncertainty display the survey recommends.","m":"agent draft (claude)"},{"id":"e576","i":"few2013dashboard--EXTENDS-->tufte2001vdqi","s":"few2013dashboard","t":"tufte2001vdqi","r":"EXTENDS","w":2,"why":"Few applies Tufte's data-ink and integrity principles to monitoring dashboards","m":"agent draft (gemini)"},{"id":"e580","i":"fisch2024stratppi--EXTENDS-->angelopoulos2023ppi","s":"fisch2024stratppi","t":"angelopoulos2023ppi","r":"EXTENDS","w":2,"why":"Stratified PPI for hybrid human+LLM-judge evaluation","m":"agent draft (gemini)"},{"id":"e592","i":"franconeri2021science--CITES-->heer2010crowdsourcing","s":"franconeri2021science","t":"heer2010crowdsourcing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1177/15291006211051956 -> 10.1145/1753326.1753357)","m":"reference-list extractor"},{"id":"e593","i":"franconeri2021science--CITES-->ioannidis2005why","s":"franconeri2021science","t":"ioannidis2005why","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1177/15291006211051956 -> 10.1371/journal.pmed.0020124)","m":"reference-list extractor"},{"id":"e594","i":"franconeri2021science--COMPLEMENTS-->borkin2016beyond","s":"franconeri2021science","t":"borkin2016beyond","r":"COMPLEMENTS","w":2,"why":"Combines Borkin et al.'s recall/redundancy rules (takeaway titles, direct labels) with perceptual grouping and CVD-safe contrast.","m":"agent draft (gemini)"},{"id":"e595","i":"franconeri2021science--EXTENDS-->cleveland1984graphical","s":"franconeri2021science","t":"cleveland1984graphical","r":"EXTENDS","w":2,"why":"Synthesizes Cleveland & McGill's perceptual encoding hierarchy with working-memory constraints, direct labeling, and colorblind-safe design.","m":"agent draft (gemini)"},{"id":"e603","i":"fu-2023--EXTENDS-->choetkiertikul-2019","s":"fu-2023","t":"choetkiertikul-2019","r":"EXTENDS","w":2,"why":"GPT2SP replaces Deep-SE's LSTM with a transformer for story-point estimation.","m":"agent draft (claude)"},{"id":"e605","i":"gallifant-2025--CITES-->collins-2024","s":"gallifant-2025","t":"collins-2024","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1038/s41591-024-03425-5 -> 10.1136/bmj-2023-078378)","m":"reference-list extractor"},{"id":"e607","i":"gallifant-2025--COMPLEMENTS-->vasey-2022","s":"gallifant-2025","t":"vasey-2022","r":"COMPLEMENTS","w":2,"why":"TRIPOD-LLM (LLM reporting) complements DECIDE-AI (early live decision-support evaluation)","m":"agent draft (gemini)"},{"id":"e608","i":"ganesan2017redundancy--CITES-->bairavasundaram2007latent","s":"ganesan2017redundancy","t":"bairavasundaram2007latent","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3125497 -> 10.1145/1254882.1254917)","m":"reference-list extractor"},{"id":"e609","i":"ganesan2017redundancy--CITES-->o’neil-1996","s":"ganesan2017redundancy","t":"o’neil-1996","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3125497 -> o’neil-1996)","m":"reference-list extractor"},{"id":"e611","i":"ganesan2021cad--CITES-->chidambaram-2013","s":"ganesan2021cad","t":"chidambaram-2013","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3423138 -> 10.1145/2517349.2522726)","m":"reference-list extractor"},{"id":"e612","i":"ganesan2021cad--CONTRASTS_WITH-->crdb2026walfailover","s":"ganesan2021cad","t":"crdb2026walfailover","r":"CONTRASTS_WITH","w":2,"why":"Contrasts cross-datacenter durability relaxation for tail-latency reduction against strict fail-stop quorum fencing under local WAL sync stalls.","m":"agent draft (gemini)"},{"id":"e613","i":"gans-2003--CITES-->halfin-1981","s":"gans-2003","t":"halfin-1981","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1287/msom.5.2.79.16071 -> 10.1287/opre.29.3.567)","m":"reference-list extractor"},{"id":"e614","i":"gans-2003--USES_METHOD-->little-1961","s":"gans-2003","t":"little-1961","r":"USES_METHOD","w":2,"why":"Call-center tutorial relies on Little's law for queue length/wait relationships.","m":"agent draft (claude)"},{"id":"e616","i":"gao2018crashrecovery--CITES-->gunawi2014cloudbugs","s":"gao2018crashrecovery","t":"gunawi2014cloudbugs","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3236024.3236030 -> 10.1145/2670979.2670986)","m":"reference-list extractor"},{"id":"e617","i":"gao2018crashrecovery--CITES-->leesatapornwongsa2016taxdc","s":"gao2018crashrecovery","t":"leesatapornwongsa2016taxdc","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3236024.3236030 -> 10.1145/2872362.2872374)","m":"reference-list extractor"},{"id":"e619","i":"gao2018crashrecovery--COMPLEMENTS-->yuan2014simple","s":"gao2018crashrecovery","t":"yuan2014simple","r":"COMPLEMENTS","w":2,"why":"Crash-recovery bugs vs error-handling bugs in production distributed systems","m":"agent draft (gemini)"},{"id":"e620","i":"garfinkel2019reconstruction--CITES-->dwork2006calibrating","s":"garfinkel2019reconstruction","t":"dwork2006calibrating","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3287287 -> 10.1007/11681878_14)","m":"reference-list extractor"},{"id":"e621","i":"garfinkel2019reconstruction--EXTENDS-->dinur2003revealing","s":"garfinkel2019reconstruction","t":"dinur2003revealing","r":"EXTENDS","w":2,"why":"Census reconstruction operationalizes Dinur-Nissim's fundamental law of information recovery","m":"agent draft (gemini)"},{"id":"e622","i":"garfinkel2019reconstruction--OPERATIONALIZES-->dinur2003revealing","s":"garfinkel2019reconstruction","t":"dinur2003revealing","r":"OPERATIONALIZES","w":2,"why":"Demonstrates Dinur & Nissim's database reconstruction theorem empirically on published aggregate statistics, motivating formal differential privacy guarantees.","m":"agent draft (gemini)"},{"id":"e623","i":"garnett2002erlanga--CITES-->halfin-1981","s":"garnett2002erlanga","t":"halfin-1981","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1287/msom.4.3.208.7753 -> 10.1287/opre.29.3.567)","m":"reference-list extractor"},{"id":"e624","i":"garnett2002erlanga--EXTENDS-->kolesar-1998","s":"garnett2002erlanga","t":"kolesar-1998","r":"EXTENDS","w":2,"why":"Erlang-A adds abandonment to the Erlang C normal approximation","m":"agent draft (gemini)"},{"id":"e625","i":"gartner-aiready-2026--COMPLEMENTS-->iso-5259","s":"gartner-aiready-2026","t":"iso-5259","r":"COMPLEMENTS","w":3,"why":"Empirical 60% AI project abandonment rate through 2026 motivating governed semantic layers and ISO 5259 data readiness","m":"seed curator"},{"id":"e626","i":"gartner-aiready-2026--EXTENDS-->dehghani-2022","s":"gartner-aiready-2026","t":"dehghani-2022","r":"EXTENDS","w":3,"why":"Extends Data Mesh domain data products into governed agentic semantic layers (+80% agent accuracy)","m":"seed curator"},{"id":"e629","i":"geiger2013editsessions--OPERATIONALIZES-->halfaker2015sessions","s":"geiger2013editsessions","t":"halfaker2015sessions","r":"COMPLEMENTS","w":2,"why":"Both formalize bimodal inter-arrival Mixture Modeling to reconstruct active human labor sessions from timestamped event logs.","m":"agent draft (claude)"},{"id":"e632","i":"gelman-2006--COMPLEMENTS-->clayton1987ebayes","s":"gelman-2006","t":"clayton1987ebayes","r":"COMPLEMENTS","w":2,"why":"Variance-component estimation that does not collapse with few groups.","m":"agent draft (gemini)"},{"id":"e644","i":"ghosh2022fight--CITES-->gao2018crashrecovery","s":"ghosh2022fight","t":"gao2018crashrecovery","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3542929.3563482 -> 10.1145/3236024.3236030)","m":"reference-list extractor"},{"id":"e645","i":"ghosh2022fight--CITES-->gunawi2014cloudbugs","s":"ghosh2022fight","t":"gunawi2014cloudbugs","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3542929.3563482 -> 10.1145/2670979.2670986)","m":"reference-list extractor"},{"id":"e646","i":"ghosh2022fight--CITES-->gunawi2016cloudstop","s":"ghosh2022fight","t":"gunawi2016cloudstop","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3542929.3563482 -> 10.1145/2987550.2987583)","m":"reference-list extractor"},{"id":"e647","i":"ghosh2022fight--CITES-->leesatapornwongsa2016taxdc","s":"ghosh2022fight","t":"leesatapornwongsa2016taxdc","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3542929.3563482 -> 10.1145/2872362.2872374)","m":"reference-list extractor"},{"id":"e648","i":"ghosh2022fight--CITES-->liu2019cloudincidents","s":"ghosh2022fight","t":"liu2019cloudincidents","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3542929.3563482 -> 10.1145/3317550.3321438)","m":"reference-list extractor"},{"id":"e651","i":"ghosh2022fight--CITES-->zhang2021upgrade","s":"ghosh2022fight","t":"zhang2021upgrade","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3542929.3563482 -> 10.1145/3477132.3483577)","m":"reference-list extractor"},{"id":"e652","i":"ghosh2022fightincidents--COMPLEMENTS-->chen2019incidenttriage","s":"ghosh2022fight","t":"chen2019incidenttriage","r":"COMPLEMENTS","w":2,"why":"Incident root-cause mix and evidence gaps vs triage routing cost","m":"agent draft (gemini)"},{"id":"e653","i":"giger-2010--CITES-->bettenburg2008goodbug","s":"giger-2010","t":"bettenburg2008goodbug","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1808920.1808933 -> 10.1145/1453101.1453146)","m":"reference-list extractor"},{"id":"e654","i":"giger-2010--CITES-->bird2009fair","s":"giger-2010","t":"bird2009fair","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1808920.1808933 -> 10.1145/1595696.1595716)","m":"reference-list extractor"},{"id":"e655","i":"giger-2010--CITES-->hooimeijer-2007","s":"giger-2010","t":"hooimeijer-2007","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1808920.1808933 -> 10.1145/1321631.1321639)","m":"reference-list extractor"},{"id":"e656","i":"giger-2010--CITES-->kim2006bugfixtime","s":"giger-2010","t":"kim2006bugfixtime","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1808920.1808933 -> 10.1145/1137983.1138027)","m":"reference-list extractor"},{"id":"e658","i":"gilardi2023chatgpt--CITES-->bayerl2011intercoder","s":"gilardi2023chatgpt","t":"bayerl2011intercoder","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1073/pnas.2305016120 -> 10.1162/coli_a_00074)","m":"reference-list extractor"},{"id":"e660","i":"glerum-2009--CITES-->castro2008bugreporting","s":"glerum2009wer","t":"castro2008bugreporting","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1629575.1629586 -> 10.1145/1346281.1346322)","m":"reference-list extractor"},{"id":"e661","i":"glerum2009wer--COMPLEMENTS-->li2020gandalf","s":"glerum2009wer","t":"li2020gandalf","r":"COMPLEMENTS","w":2,"why":"Field-evidence ROI patterns at Microsoft: WER closes the bug-bucket-to-fix loop; Gandalf gates releases on fleet telemetry. Together they are the two peer uses a DB vendor's support-bundle pipeline should be judged against.","m":"agent draft (gemini)"},{"id":"e662","i":"glerum2009wer--EXTENDS-->brodie2005symptom","s":"glerum2009wer","t":"brodie2005symptom","r":"EXTENDS","w":2,"why":"Windows Error Reporting (WER) operationalizes automated crash symptom matching and non-blameable frame skip-lists at planetary scale across billions of field crash reports.","m":"agent draft (gemini)"},{"id":"e663","i":"glerum2009wer--OPERATIONALIZES-->kcs2016v6","s":"kcs2016v6","t":"glerum2009wer","r":"OPERATIONALIZES","w":2,"why":"Crash buckets as automated known-issue matching feeding fix prioritisation","m":"agent draft (gemini)"},{"id":"e664","i":"goddard-2012--CITES-->landis-1977","s":"goddard-2012","t":"landis-1977","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1136/amiajnl-2011-000089 -> 10.2307/2529310)","m":"reference-list extractor"},{"id":"e665","i":"goddard-2012--CITES-->lee-2004","s":"goddard-2012","t":"lee-2004","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1136/amiajnl-2011-000089 -> 10.1518/hfes.46.1.50_30392)","m":"reference-list extractor"},{"id":"e666","i":"goddard-2012--OPERATIONALIZES-->parasuraman2010complacency","s":"goddard-2012","t":"parasuraman2010complacency","r":"OPERATIONALIZES","w":2,"why":"Clinical review quantifying automation bias that the attentional model explains","m":"agent draft (gemini)"},{"id":"e667","i":"goeminne2014coevolving--COMPLEMENTS-->qiu2013schemacoevolution","s":"goeminne2014coevolving","t":"qiu2013schemacoevolution","r":"COMPLEMENTS","w":2,"why":"Co-evolution of DB-related and code changes in a data-intensive system","m":"agent draft (claude)"},{"id":"e669","i":"goodmanbacon2021decomp--CITES-->abadie2010sc","s":"goodmanbacon2021decomp","t":"abadie2010sc","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.jeconom.2021.03.014 -> 10.1198/jasa.2009.ap08746)","m":"reference-list extractor"},{"id":"e671","i":"goodmanbacon2021decomp--CITES-->sun2021event","s":"goodmanbacon2021decomp","t":"sun2021event","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.jeconom.2021.03.014 -> sun2021event)","m":"reference-list extractor"},{"id":"e672","i":"goodmanbacon2021decomp--CONTRASTS_WITH-->callaway2021did","s":"goodmanbacon2021decomp","t":"callaway2021did","r":"CONTRASTS_WITH","w":2,"why":"Bacon diagnoses TWFE bias under staggered timing; CS gives a clean ATT(g,t) estimator","m":"agent draft (gemini)"},{"id":"e674","i":"graefe2006sorting--CITES-->selinger-1979","s":"graefe2006sorting","t":"selinger-1979","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1132960.1132964 -> 10.1145/582095.582099)","m":"reference-list extractor"},{"id":"e676","i":"gramazio-2017--COMPLEMENTS-->szafir-2018","s":"gramazio-2017","t":"szafir-2018","r":"COMPLEMENTS","w":2,"why":"categorical palette generation balancing discriminability and preference","m":"agent draft (claude)"},{"id":"e677","i":"gray2005krspeed--COMPLEMENTS-->argote-2011","s":"gray2005krspeed","t":"argote-2011","r":"COMPLEMENTS","w":2,"why":"Speed-vs-learning trade-off in support repositories is a specific case of knowledge embedded in tools vs members.","m":"agent draft (claude)"},{"id":"e681","i":"green-2007--CITES-->borst-2004","s":"green-2007","t":"borst-2004","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1937-5956.2007.tb00164.x -> 10.1287/opre.1030.0081)","m":"reference-list extractor"},{"id":"e682","i":"green-2007--CITES-->brown-2005","s":"green-2007","t":"brown-2005","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1937-5956.2007.tb00164.x -> 10.1198/016214504000001808)","m":"reference-list extractor"},{"id":"e683","i":"green-2007--CITES-->gans-2003","s":"green-2007","t":"gans-2003","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1937-5956.2007.tb00164.x -> 10.1287/msom.5.2.79.16071)","m":"reference-list extractor"},{"id":"e684","i":"green-2007--CITES-->halfin-1981","s":"green-2007","t":"halfin-1981","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1937-5956.2007.tb00164.x -> 10.1287/opre.29.3.567)","m":"reference-list extractor"},{"id":"e686","i":"green-2007--EXTENDS-->kolesar-1998","s":"green-2007","t":"kolesar-1998","r":"EXTENDS","w":2,"why":"Same authors extend stationary Erlang-C staffing insights to time-varying demand.","m":"agent draft (claude)"},{"id":"e692","i":"greenland2005multiplebias--CITES-->frangakis2002principal","s":"greenland2005multiplebias","t":"frangakis2002principal","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1467-985x.2004.00349.x -> 10.1111/j.0006-341x.2002.00021.x)","m":"reference-list extractor"},{"id":"e693","i":"greenland2005multiplebias--CITES-->heckman1979selection","s":"greenland2005multiplebias","t":"heckman1979selection","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1467-985x.2004.00349.x -> 10.2307/1912352)","m":"reference-list extractor"},{"id":"e694","i":"greenland2005multiplebias--CITES-->rubin1976missing","s":"greenland2005multiplebias","t":"rubin1976missing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1467-985x.2004.00349.x -> 10.1093/biomet/63.3.581)","m":"reference-list extractor"},{"id":"e695","i":"greenland2005multiplebias--COMPLEMENTS-->carroll2006measurementerror","s":"greenland2005multiplebias","t":"carroll2006measurementerror","r":"COMPLEMENTS","w":2,"why":"Carroll et al. correct covariate measurement error when an internal validation subsample exists; Greenland's Monte Carlo bias analysis bounds unmeasured confounding and error via prior distributions.","m":"agent draft (claude)"},{"id":"e696","i":"greenland2005multiplebias--COMPLEMENTS-->saisana-2005","s":"greenland2005multiplebias","t":"saisana-2005","r":"COMPLEMENTS","w":2,"why":"Pairs probabilistic multi-bias Monte Carlo simulation with global variance-based Sobol sensitivity indices for composite indicators.","m":"agent draft (claude)"},{"id":"e697","i":"gregg-2016--COMPLEMENTS-->gregg-2012","s":"gregg-2016","t":"gregg2012use","r":"COMPLEMENTS","w":2,"why":"USE method locates the saturated resource; flame graphs (on-CPU and off-CPU) then explain which code paths consume or wait on it.","m":"agent draft (claude)"},{"id":"e698","i":"gregg-2016--USES_METHOD-->ren-2010","s":"gregg-2016","t":"ren-2010","r":"USES_METHOD","w":2,"why":"Flame graphs visualize aggregated sampled stack profiles of the kind GWP collects continuously (perf/pprof stack samples).","m":"agent draft (gemini)"},{"id":"e699","i":"groen2017qualityexperts--CITES-->pagano-2013","s":"groen2017qualityexperts","t":"pagano-2013","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/re.2017.73 -> 10.1109/re.2013.6636712)","m":"reference-list extractor"},{"id":"e701","i":"groen2017qualityexperts--COMPLEMENTS-->guzman2014feature","s":"groen2017qualityexperts","t":"guzman2014feature","r":"COMPLEMENTS","w":2,"why":"Feedback channels reveal non-functional quality, not just features.","m":"agent draft (gemini)"},{"id":"e707","i":"gunawi-2014--CITES-->do-2013","s":"gunawi2014cloudbugs","t":"do-2013","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2670979.2670986 -> 10.1145/2523616.2523627)","m":"reference-list extractor"},{"id":"e709","i":"gunawi-2014--CITES-->yin2011configerrors","s":"gunawi2014cloudbugs","t":"yin2011configerrors","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2670979.2670986 -> 10.1145/2043556.2043572)","m":"reference-list extractor"},{"id":"e710","i":"gunawi-2016--CITES-->bairavasundaram2007latent","s":"gunawi2016cloudstop","t":"bairavasundaram2007latent","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2987550.2987583 -> 10.1145/1254882.1254917)","m":"reference-list extractor"},{"id":"e711","i":"gunawi-2016--CITES-->do-2013","s":"gunawi2016cloudstop","t":"do-2013","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2987550.2987583 -> 10.1145/2523616.2523627)","m":"reference-list extractor"},{"id":"e712","i":"gunawi-2016--CITES-->gunawi2014cloudbugs","s":"gunawi2016cloudstop","t":"gunawi2014cloudbugs","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2987550.2987583 -> 10.1145/2670979.2670986)","m":"reference-list extractor"},{"id":"e713","i":"gunawi-2016--CITES-->yin2011configerrors","s":"gunawi2016cloudstop","t":"yin2011configerrors","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2987550.2987583 -> 10.1145/2043556.2043572)","m":"reference-list extractor"},{"id":"e714","i":"gunawi-2016--COMPLEMENTS-->gunawi2018failslow","s":"gunawi2016cloudstop","t":"gunawi2018failslow","r":"COMPLEMENTS","w":2,"why":"Outage-level and hardware fail-slow-level field studies from the same group.","m":"agent draft (claude)"},{"id":"e715","i":"gunawi2018failslow--COMPLEMENTS-->huang-2017","s":"gunawi2018failslow","t":"huang2017gray","r":"COMPLEMENTS","w":2,"why":"Gray failure gives the observability framing (detector vs app view); fail-slow at scale gives the hardware-incident evidence that slow components are a common cause.","m":"agent draft (claude)"},{"id":"e716","i":"gunawi2014cloudbugs--CONTRASTS_WITH-->glerum2009wer","s":"gunawi2014cloudbugs","t":"glerum2009wer","r":"CONTRASTS_WITH","w":2,"why":"Issue-tracker corpus vs automatic field-report corpus; different selection mechanisms","m":"agent draft (gemini)"},{"id":"e717","i":"hao2016tailatstore--COMPLEMENTS-->gunawi-2018","s":"hao2016tailatstore","t":"gunawi2018failslow","r":"COMPLEMENTS","w":2,"why":"Fleet frequency/severity of slow I/O (Hao) vs root causes and cascades of fail-slow hardware (Gunawi)","m":"agent draft (gemini)"},{"id":"e727","i":"guzman2014feature--EXTENDS-->pagano-2013","s":"guzman2014feature","t":"pagano-2013","r":"EXTENDS","w":2,"why":"Moves from review-level to feature-level feedback analysis.","m":"agent draft (gemini)"},{"id":"e730","i":"haerem-2015--CITES-->campbell-1988","s":"haerem-2015","t":"campbell-1988","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.5465/amr.2013.0350 -> 10.5465/amr.1988.4306775)","m":"reference-list extractor"},{"id":"e731","i":"haerem-2015--CITES-->liu-2012","s":"haerem-2015","t":"liu-2012","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.5465/amr.2013.0350 -> 10.1016/j.ergon.2012.09.001)","m":"reference-list extractor"},{"id":"e732","i":"haerem-2015--EXTENDS-->wood-1986","s":"haerem-2015","t":"wood-1986","r":"EXTENDS","w":2,"why":"Replaces additive component/coordinative counts with pathways in the task event network.","m":"agent draft (claude-sonnet-5-5)"},{"id":"e735","i":"halfaker2015sessions--CONTRASTS_WITH-->killick2012optimal","s":"halfaker2015sessions","t":"killick2012optimal","r":"CONTRASTS_WITH","w":2,"why":"Gap-threshold sessionisation of events vs penalised changepoints on sampled series.","m":"agent draft (gemini)"},{"id":"e740","i":"hao2016tailatstore--OPERATIONALIZES-->lu2023perseus","s":"lu2023perseus","t":"hao2016tailatstore","r":"OPERATIONALIZES","w":2,"why":"Peer-median slow-drive rule used by fail-slow detection.","m":"agent draft (gemini)"},{"id":"e742","i":"hauer2020meaningful--COMPLEMENTS-->beyer2018workbook","s":"hauer2020meaningful","t":"beyer2018workbook","r":"COMPLEMENTS","w":2,"why":"Windowed user-uptime complements request-based SLOs/burn-rate alerts by separating many short from few long outages.","m":"agent draft (claude)"},{"id":"e743","i":"hauer2020meaningful--OPERATIONALIZES-->beyer2018workbook","s":"hauer2020meaningful","t":"beyer2018workbook","r":"OPERATIONALIZES","w":2,"why":"Windowed availability as a reporting form for SLO compliance.","m":"agent draft (gemini)"},{"id":"e748","i":"hayes-2007--CITES-->krippendorff-2004","s":"hayes-2007","t":"krippendorff-2004","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1080/19312450709336664 -> 10.1111/j.1468-2958.2004.tb00738.x)","m":"reference-list extractor"},{"id":"e749","i":"hayes-2007--EXTENDS-->landis-1977","s":"hayes-2007","t":"landis-1977","r":"EXTENDS","w":2,"why":"Krippendorff alpha guidance generalises beyond kappa's two-rater categorical setting with CIs and missing data.","m":"agent draft (claude-sonnet-5-5)"},{"id":"e750","i":"hazem-2026-arxiv--EXTENDS-->ziems-2024","s":"hazem-2026-arxiv","t":"ziems-2024","r":"EXTENDS","w":2,"why":"Re-runs the Ziems et al. zero-shot CSS annotation benchmark with a typed decision model vs 19 LLMs.","m":"agent draft (claude)"},{"id":"e751","i":"hazem-2026-arxiv--USES_METHOD-->chen2024frugalgpt","s":"hazem-2026-arxiv","t":"chen2024frugalgpt","r":"USES_METHOD","w":2,"why":"Confidence-gated cascade (decision model first, escalate low-confidence to LLM) matches LLM accuracy at 25-50% cost, the FrugalGPT cascade pattern.","m":"agent draft (claude)"},{"id":"e753","i":"he-2018--CITES-->fu2009execution","s":"he-2018","t":"fu2009execution","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3236024.3236083 -> 10.1109/icdm.2009.60)","m":"reference-list extractor"},{"id":"e757","i":"he-2018--EXTENDS-->lin2016logcluster","s":"he-2018","t":"lin2016logcluster","r":"EXTENDS","w":2,"why":"Log3C extends LogCluster's log-sequence clustering by correlating clusters with KPIs to rank impactful problems (same Microsoft group).","m":"agent draft (claude)"},{"id":"e758","i":"he2016experience--CITES-->beschastnikh2011synoptic","s":"he2016experience","t":"beschastnikh2011synoptic","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/issre.2016.21 -> 10.1145/2025113.2025151)","m":"reference-list extractor"},{"id":"e761","i":"he2016experience--CITES-->fu2009execution","s":"he2016experience","t":"fu2009execution","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/issre.2016.21 -> 10.1109/icdm.2009.60)","m":"reference-list extractor"},{"id":"e763","i":"he2016experience--CITES-->lin2016logcluster","s":"he2016experience","t":"lin2016logcluster","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/issre.2016.21 -> lin2016logcluster)","m":"reference-list extractor"},{"id":"e764","i":"he2016experience--CITES-->oliner2007supercomputers","s":"he2016experience","t":"oliner2007supercomputers","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/issre.2016.21 -> 10.1109/dsn.2007.103)","m":"reference-list extractor"},{"id":"e767","i":"he2017drain--CITES-->oliner2007supercomputers","s":"he2017drain","t":"oliner2007supercomputers","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icws.2017.13 -> 10.1109/dsn.2007.103)","m":"reference-list extractor"},{"id":"e769","i":"he2017drain--COMPLEMENTS-->du2016spell","s":"he2017drain","t":"du2016spell","r":"COMPLEMENTS","w":2,"why":"Drain uses fixed-depth prefix trees with token similarity whereas Spell uses streaming longest common subsequence (LCS) for online log template mining.","m":"agent draft (gemini)"},{"id":"e770","i":"he2017drain--COMPLEMENTS-->guo2017calibration","s":"he2017drain","t":"guo2017calibration","r":"COMPLEMENTS","w":3,"why":"Combines fixed-depth parse-tree log template clustering with post-hoc probability calibration for log-based incident classification.","m":"agent draft (gemini)"},{"id":"e771","i":"he2017drain--COMPLEMENTS-->kleinberg2002bursty","s":"he2017drain","t":"kleinberg2002bursty","r":"COMPLEMENTS","w":2,"why":"Drain collapses high-volume log lines into canonical templates while Kleinberg's burst model motivates temporal windowing and burst deduplication over template streams.","m":"agent draft (gemini)"},{"id":"e772","i":"he2017drain--CONTRASTS_WITH-->du2016spell","s":"he2017drain","t":"du2016spell","r":"CONTRASTS_WITH","w":2,"why":"Avoids Spell's quadratic O(|m|*|t|) LCS dynamic programming by bounding parse-tree depth and routing via token count and leading non-digit tokens.","m":"agent draft (gemini)"},{"id":"e773","i":"he2017drain--CONTRASTS_WITH-->keiichi-2016-arxiv","s":"he2017drain","t":"keiichi-2016-arxiv","r":"CONTRASTS_WITH","w":2,"why":"Uses exact token matching and similarity thresholds in a fixed-depth tree rather than LenMa's word character-length cosine similarity vectors.","m":"agent draft (gemini)"},{"id":"e774","i":"he2017drain--EXTENDS-->fu2009execution","s":"he2017drain","t":"fu2009execution","r":"EXTENDS","w":2,"why":"Replaces offline clustering with a bounded fixed-depth parse tree for streaming O(1) online log template extraction.","m":"agent draft (gemini)"},{"id":"e775","i":"he2017drain--USES_METHOD-->morrison-1968","s":"he2017drain","t":"morrison-1968","r":"USES_METHOD","w":2,"why":"Adapts radix/prefix tree routing (PATRICIA) into a bounded-depth token prefix tree with wildcard transitions for online log template clustering.","m":"agent draft (gemini)"},{"id":"e776","i":"heckman1979selection--COMPLEMENTS-->lee2009bounds","s":"heckman1979selection","t":"lee2009bounds","r":"COMPLEMENTS","w":2,"why":"Selection model vs bounds under sample selection","m":"agent draft (gemini)"},{"id":"e777","i":"heer2010crowdsourcing--EXTENDS-->cleveland1984graphical","s":"heer2010crowdsourcing","t":"cleveland1984graphical","r":"EXTENDS","w":2,"why":"crowdsourced replication of the C&M graphical-perception ranking","m":"agent draft (claude)"},{"id":"e778","i":"heinrich2018requirements--EXTENDS-->batini2009methodologies","s":"heinrich2018requirements","t":"batini2009methodologies","r":"EXTENDS","w":2,"why":"Adds formal requirements (bounded, interval-scaled, sound aggregation) that the surveyed DQ metrics must satisfy.","m":"agent draft (claude)"},{"id":"e779","i":"helske-2021--CITES-->fernandes2018quantiledotplots","s":"helske-2021","t":"fernandes2018quantiledotplots","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2021.3073466 -> 10.1145/3173574.3173718)","m":"reference-list extractor"},{"id":"e781","i":"helske-2021--CITES-->hullman2019pursuiterror","s":"helske-2021","t":"hullman2019pursuiterror","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2021.3073466 -> 10.1109/tvcg.2018.2864889)","m":"reference-list extractor"},{"id":"e782","i":"helske-2021--CITES-->ioannidis2005why","s":"helske-2021","t":"ioannidis2005why","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2021.3073466 -> 10.1371/journal.pmed.0020124)","m":"reference-list extractor"},{"id":"e783","i":"helske-2021--CITES-->kale-2019","s":"helske-2021","t":"kale-2019","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2021.3073466 -> 10.1109/tvcg.2018.2864909)","m":"reference-list extractor"},{"id":"e785","i":"helske-2021--CITES-->simmons2011false","s":"helske-2021","t":"simmons2011false","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2021.3073466 -> 10.1177/0956797611417632)","m":"reference-list extractor"},{"id":"e787","i":"helske-2021--OPERATIONALIZES-->spiegelhalter2005funnel","s":"helske-2021","t":"spiegelhalter2005funnel","r":"OPERATIONALIZES","w":2,"why":"Both argue against dichotomous colouring of point estimates; funnel limits are the operational form for unit comparison","m":"agent draft (gemini)"},{"id":"e789","i":"herbsleb2003distributed--CITES-->parnas1972criteria","s":"herbsleb2003distributed","t":"parnas1972criteria","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2003.1205177 -> 10.1145/361598.361623)","m":"reference-list extractor"},{"id":"e791","i":"hernán-2010--CITES-->cole-2004","s":"hernán-2010","t":"cole-2004","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1097/ede.0b013e3181c1ea43 -> 10.1016/j.cmpb.2003.10.004)","m":"reference-list extractor"},{"id":"e792","i":"hernán-2010--CITES-->hernan2004selection","s":"hernán-2010","t":"hernan2004selection","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1097/ede.0b013e3181c1ea43 -> 10.1097/01.ede.0000135174.63482.43)","m":"reference-list extractor"},{"id":"e793","i":"hernán-2010--CONTRASTS_WITH-->cox-1972","s":"hernán-2010","t":"cox-1972","r":"CONTRASTS_WITH","w":2,"why":"Argues Cox hazard ratios are poor effect summaries; prefer standardised survival curves","m":"agent draft (claude)"},{"id":"e794","i":"hernán-2016--CITES-->lipsitch-2010","s":"hernán-2016","t":"lipsitch-2010","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1093/aje/kwv254 -> 10.1097/ede.0b013e3181d61eeb)","m":"reference-list extractor"},{"id":"e795","i":"hernán-2016--CITES-->suissa2008immortal","s":"hernán-2016","t":"suissa2008immortal","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1093/aje/kwv254 -> 10.1093/aje/kwm324)","m":"reference-list extractor"},{"id":"e796","i":"hernán-2016--COMPLEMENTS-->hernan2004selection","s":"hernán-2016","t":"hernan2004selection","r":"COMPLEMENTS","w":2,"why":"time-zero alignment in target-trial emulation avoids the collider selection structures catalogued in 2004","m":"agent draft (gemini)"},{"id":"e798","i":"herschel2017provsurvey--CITES-->wang1996dataquality","s":"herschel2017provsurvey","t":"wang1996dataquality","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s00778-017-0486-1 -> 10.1080/07421222.1996.11518099)","m":"reference-list extractor"},{"id":"e799","i":"herschel2017provsurvey--COMPLEMENTS-->pimentel2019jupyterrepro","s":"herschel2017provsurvey","t":"pimentel2019jupyterrepro","r":"COMPLEMENTS","w":2,"why":"workflow provenance is the remedy for notebook non-reproducibility","m":"agent draft (gemini)"},{"id":"e800","i":"herschel2017provsurvey--EXTENDS-->buneman-2001","s":"herschel2017provsurvey","t":"buneman-2001","r":"EXTENDS","w":2,"why":"Survey builds on why/where provenance","m":"agent draft (gemini)"},{"id":"e801","i":"hidalgo2020slo--EXTENDS-->beyer2018workbook","s":"hidalgo2020slo","t":"beyer2018workbook","r":"EXTENDS","w":2,"why":"Book-length treatment of SLI/SLO/error-budget practice from the workbook","m":"agent draft (gemini)"},{"id":"e802","i":"hidalgo2020slo--OPERATIONALIZES-->mogul2019nines","s":"hidalgo2020slo","t":"mogul2019nines","r":"OPERATIONALIZES","w":2,"why":"Error budgets + burn-rate alerting operationalize frequency-vs-duration aware availability targets","m":"agent draft (gemini)"},{"id":"e805","i":"hoefler-2015--COMPLEMENTS-->alcott2005jevons","s":"hoefler-2015","t":"alcott2005jevons","r":"COMPLEMENTS","w":2,"why":"Provides the statistical benchmarking discipline (non-parametric quantiles, full-loop cost accounting) needed to measure whether unit-efficiency gains reduce or rebound total system resource consumption.","m":"agent draft (gemini)"},{"id":"e806","i":"hoefler-2015--COMPLEMENTS-->kalibera2013rigorous","s":"hoefler-2015","t":"kalibera2013rigorous","r":"COMPLEMENTS","w":2,"why":"Both prescribe statistically sound reporting (CIs, nonparametric stats) for performance results","m":"agent draft (gemini)"},{"id":"e808","i":"hoff2015trust--EXTENDS-->lee-2004","s":"hoff2015trust","t":"lee-2004","r":"EXTENDS","w":2,"why":"Three-layer model of trust in automation built on Lee and See","m":"agent draft (gemini)"},{"id":"e809","i":"hoffswell2020responsive--CITES-->satyanarayan-2017","s":"hoffswell2020responsive","t":"satyanarayan-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3313831.3376777 -> satyanarayan-2017)","m":"reference-list extractor"},{"id":"e810","i":"hofman-2020--CITES-->fernandes2018quantiledotplots","s":"hofman-2020","t":"fernandes2018quantiledotplots","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3313831.3376454 -> 10.1145/3173574.3173718)","m":"reference-list extractor"},{"id":"e811","i":"hofman-2020--CITES-->kay-2016","s":"hofman-2020","t":"kay-2016","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3313831.3376454 -> 10.1145/2858036.2858558)","m":"reference-list extractor"},{"id":"e812","i":"hofman-2020--EXTENDS-->kale-2019","s":"hofman-2020","t":"kale-2019","r":"EXTENDS","w":2,"why":"uses HOPs as alternative to CI displays","m":"agent draft (claude)"},{"id":"e814","i":"holmstrom-1991--COMPLEMENTS-->kerr-1975","s":"holmstrom-1991","t":"kerr1975folly","r":"COMPLEMENTS","w":2,"why":"Kerr describes the folly of rewarding A while hoping for B; Holmstrom and Milgrom formalise it as multitask incentive distortion.","m":"agent draft (claude)"},{"id":"e815","i":"hooimeijer-2007--CITES-->kim2006bugfixtime","s":"hooimeijer-2007","t":"kim2006bugfixtime","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1321631.1321639 -> 10.1145/1137983.1138027)","m":"reference-list extractor"},{"id":"e816","i":"hooimeijer-2007--CITES-->weiss-2007","s":"hooimeijer-2007","t":"weiss-2007","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1321631.1321639 -> 10.1109/msr.2007.13)","m":"reference-list extractor"},{"id":"e817","i":"hooimeijer-2007--COMPLEMENTS-->bettenburg2008goodbug","s":"hooimeijer-2007","t":"bettenburg2008goodbug","r":"COMPLEMENTS","w":2,"why":"Earlier report-quality model predicting triage-time outcomes from surface features","m":"agent draft (claude)"},{"id":"e818","i":"hooimeijer-2007--COMPLEMENTS-->panjer-2007","s":"hooimeijer-2007","t":"panjer-2007","r":"COMPLEMENTS","w":2,"why":"Report-quality features and lifetime classification on Bugzilla-style data from the same period; both highlight entanglement of report attributes with time.","m":"agent draft (claude-sonnet-5-5)"},{"id":"e819","i":"horspool1980practical--EXTENDS-->boyer1977fast","s":"horspool1980practical","t":"boyer1977fast","r":"EXTENDS","w":3,"why":"Horspool simplifies Boyer-Moore to a rightmost-character bad-character shift table, forming the basis of CPython's fastsearch (str.find / 'in') literal prefilter.","m":"agent draft (gemini)"},{"id":"e826","i":"hu2026cbstail--COMPLEMENTS-->gulati2009parda","s":"hu2026cbstail","t":"gulati2009parda","r":"COMPLEMENTS","w":2,"why":"Both mitigate shared-backend noisy neighbours by throttling (dual-bucket vs latency-driven queue depth).","m":"agent draft (claude)"},{"id":"e827","i":"hu2026cbstail--EXTENDS-->zhang2024ebsglory","s":"hu2026cbstail","t":"zhang2024ebsglory","r":"EXTENDS","w":2,"why":"Same Alibaba EBS architecture; root-causes its tail spikes to bursty major-client VDs and proxy scheduling.","m":"agent draft (claude)"},{"id":"e831","i":"huang-2015--COMPLEMENTS-->yin2011configerrors","s":"huang-2015","t":"yin2011configerrors","r":"COMPLEMENTS","w":2,"why":"Validation tooling responding to the empirical config-error taxonomy","m":"agent draft (gemini)"},{"id":"e832","i":"huang-2017--CITES-->gunawi2016cloudstop","s":"huang2017gray","t":"gunawi2016cloudstop","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3102980.3103005 -> 10.1145/2987550.2987583)","m":"reference-list extractor"},{"id":"e833","i":"huang2017gray--CONTRASTS_WITH-->treynor-2017","s":"huang2017gray","t":"treynor-2017","r":"CONTRASTS_WITH","w":2,"why":"Gray failure breaks the binary up/down dependency model assumed in availability calculus","m":"agent draft (gemini)"},{"id":"e834","i":"huang2018panorama--EXTENDS-->huang2017gray","s":"huang2018panorama","t":"huang2017gray","r":"EXTENDS","w":2,"why":"Panorama turns the gray-failure observer idea into in-situ observers in code.","m":"agent draft (gemini)"},{"id":"e835","i":"huang2018panorama--OPERATIONALIZES-->huang2017gray","s":"huang2018panorama","t":"huang2017gray","r":"OPERATIONALIZES","w":2,"why":"Operationalizes gray-failure differential observability by correlating cross-component request observations to detect partial liveness degradation.","m":"agent draft (claude)"},{"id":"e836","i":"huang2020tidb--CITES-->corbett2013spanner","s":"huang2020tidb","t":"corbett2012spanner","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3415478.3415535 -> 10.1145/2491245)","m":"reference-list extractor"},{"id":"e837","i":"huang2020tidb--CITES-->o’neil-1996","s":"huang2020tidb","t":"o’neil-1996","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3415478.3415535 -> 10.1007/s002360050048)","m":"reference-list extractor"},{"id":"e838","i":"huang2020tidb--COMPLEMENTS-->taft2020cockroachdb","s":"huang2020tidb","t":"taft2020cockroachdb","r":"COMPLEMENTS","w":2,"why":"Multi-Raft LSM distributed SQL architecture (TiDB/TiKV) with Raft learner columnar replicas","m":"agent draft (gemini)"},{"id":"e840","i":"hubert1985ari--COMPLEMENTS-->rousseeuw1987silhouettes","s":"hubert1985ari","t":"rousseeuw1987silhouettes","r":"COMPLEMENTS","w":2,"why":"External vs internal cluster validation","m":"agent draft (claude)"},{"id":"e841","i":"hughes-2016--EXTENDS-->claessen2000quickcheck","s":"hughes-2016","t":"claessen2000quickcheck","r":"EXTENDS","w":2,"why":"industrial experience report on QuickCheck property-based testing","m":"agent draft (gemini)"},{"id":"e842","i":"hullman-2011--CITES-->cleveland1984graphical","s":"hullman-2011","t":"cleveland1984graphical","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2011.255 -> 10.1080/01621459.1984.10478080)","m":"reference-list extractor"},{"id":"e843","i":"hullman-2011--CITES-->segel2010narrative","s":"hullman-2011","t":"segel2010narrative","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2011.255 -> 10.1109/tvcg.2010.179)","m":"reference-list extractor"},{"id":"e844","i":"hullman-2011--CITES-->tufte2001vdqi","s":"hullman-2011","t":"tufte2001vdqi","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2011.255 -> tufte2001vdqi)","m":"reference-list extractor"},{"id":"e845","i":"hullman-2011--EXTENDS-->kong2019titlemisalignment","s":"kong2019titlemisalignment","t":"hullman-2011","r":"EXTENDS","w":2,"why":"framing via titles and colour steers reading beyond the data","m":"agent draft (gemini)"},{"id":"e846","i":"hullman-2020--CITES-->correll-2014","s":"hullman-2020","t":"correll-2014","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2019.2934287 -> 10.1109/tvcg.2014.2346298)","m":"reference-list extractor"},{"id":"e847","i":"hullman-2020--CITES-->fernandes2018quantiledotplots","s":"hullman-2020","t":"fernandes2018quantiledotplots","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2019.2934287 -> 10.1145/3173574.3173718)","m":"reference-list extractor"},{"id":"e848","i":"hullman-2020--CITES-->kale-2019","s":"hullman-2020","t":"kale-2019","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2019.2934287 -> kale-2019)","m":"reference-list extractor"},{"id":"e849","i":"hullman-2020--CITES-->kay-2016","s":"hullman-2020","t":"kay-2016","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2019.2934287 -> 10.1145/2858036.2858558)","m":"reference-list extractor"},{"id":"e850","i":"hullman-2020--COMPLEMENTS-->hullman2019pursuiterror","s":"hullman-2020","t":"hullman2019pursuiterror","r":"COMPLEMENTS","w":2,"why":"why authors omit uncertainty vs how it is evaluated","m":"agent draft (claude)"},{"id":"e851","i":"hullman2015hops--CITES-->correll-2014","s":"hullman2015hops","t":"correll-2014","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1371/journal.pone.0142444 -> 10.1109/tvcg.2014.2346298)","m":"reference-list extractor"},{"id":"e852","i":"hullman2015hops--CITES-->heer2010crowdsourcing","s":"hullman2015hops","t":"heer2010crowdsourcing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1371/journal.pone.0142444 -> 10.1145/1753326.1753357)","m":"reference-list extractor"},{"id":"e853","i":"hullman2015hops--COMPLEMENTS-->kale-2019","s":"hullman2015hops","t":"kale-2019","r":"COMPLEMENTS","w":2,"why":"Both test HOPs; 2015 on ordering reliability, 2019 on trend judgements","m":"agent draft (gemini)"},{"id":"e854","i":"hullman2019pursuiterror--CITES-->ioannidis2005why","s":"hullman2019pursuiterror","t":"ioannidis2005why","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2018.2864889 -> 10.1371/journal.pmed.0020124)","m":"reference-list extractor"},{"id":"e856","i":"huo2023semparser--CITES-->chu-2021","s":"huo2023semparser","t":"chu-2021","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00082 -> 10.1109/icde51399.2021.00274)","m":"reference-list extractor"},{"id":"e857","i":"huo2023semparser--CITES-->cohen1960kappa","s":"huo2023semparser","t":"cohen1960kappa","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00082 -> 10.1177/001316446002000104)","m":"reference-list extractor"},{"id":"e858","i":"huo2023semparser--CITES-->dai2020logram","s":"huo2023semparser","t":"dai2020logram","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00082 -> 10.1109/tse.2020.3007554)","m":"reference-list extractor"},{"id":"e859","i":"huo2023semparser--CITES-->fu2009execution","s":"huo2023semparser","t":"fu2009execution","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00082 -> 10.1109/icdm.2009.60)","m":"reference-list extractor"},{"id":"e860","i":"huo2023semparser--CITES-->he2016experience","s":"huo2023semparser","t":"he2016experience","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00082 -> 10.1109/issre.2016.21)","m":"reference-list extractor"},{"id":"e862","i":"huo2023semparser--CITES-->keiichi-2016-arxiv","s":"huo2023semparser","t":"keiichi-2016-arxiv","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00082 -> keiichi-2016-arxiv)","m":"reference-list extractor"},{"id":"e863","i":"huo2023semparser--CITES-->lin2016logcluster","s":"huo2023semparser","t":"lin2016logcluster","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00082 -> lin2016logcluster)","m":"reference-list extractor"},{"id":"e866","i":"huo2023semparser--CONTRASTS_WITH-->he2017drain","s":"huo2023semparser","t":"he2017drain","r":"CONTRASTS_WITH","w":2,"why":"SemParser recovers parameter semantics that syntax-only parsers like Drain discard.","m":"agent draft (gemini)"},{"id":"e868","i":"hutwagner-2003--COMPLEMENTS-->glerum2009wer","s":"hutwagner-2003","t":"glerum2009wer","r":"COMPLEMENTS","w":2,"why":"Aberration-detection timeliness (EARS) plus bucket reach ranking (WER) define fleet early-detection metrics.","m":"agent draft (claude)"},{"id":"e869","i":"hutwagner-2003--CONTRASTS_WITH-->farrington-1996","s":"hutwagner-2003","t":"farrington-1996","r":"CONTRASTS_WITH","w":2,"why":"Contrasts short-baseline EARS aberration statistics (C1-C3) against long-baseline seasonal quasi-Poisson regression when historical surveillance baselines are unavailable.","m":"agent draft (claude)"},{"id":"e870","i":"huyse-1999--COMPLEMENTS-->shippee-2012","s":"huyse-1999","t":"shippee-2012","r":"COMPLEMENTS","w":2,"why":"INTERMED operationalises a profile of complexity domains; cumulative complexity gives the conceptual burden/capacity framing.","m":"agent draft (claude-sonnet-5-5)"},{"id":"e871","i":"ibrahim-2016--CITES-->aksin2007moderncallcenter","s":"ibrahim2016forecasting","t":"aksin2007moderncallcenter","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.ijforecast.2015.11.012 -> 10.1111/j.1937-5956.2007.tb00288.x)","m":"reference-list extractor"},{"id":"e872","i":"ibrahim-2016--CITES-->green-2007","s":"ibrahim2016forecasting","t":"green-2007","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.ijforecast.2015.11.012 -> 10.1111/j.1937-5956.2007.tb00164.x)","m":"reference-list extractor"},{"id":"e873","i":"ibrahim-2016--CITES-->kim-2014","s":"ibrahim2016forecasting","t":"kim-2014","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.ijforecast.2015.11.012 -> 10.1287/msom.2014.0490)","m":"reference-list extractor"},{"id":"e874","i":"ibrahim2016forecasting--EXTENDS-->brown-2005","s":"ibrahim2016forecasting","t":"brown-2005","r":"EXTENDS","w":2,"why":"Survey builds on Brown et al.'s statistical analysis of call-center arrivals (time-varying, overdispersed) toward forecasting methods.","m":"agent draft (claude)"},{"id":"e875","i":"ibrahim2016forecasting--EXTENDS-->gans-2003","s":"ibrahim2016forecasting","t":"gans-2003","r":"EXTENDS","w":2,"why":"Updates the forecasting component of the Gans et al. forecasting-staffing-scheduling chain with a survey of arrival models.","m":"agent draft (claude)"},{"id":"e877","i":"inoue2015simd--CITES-->balkesen2013joins","s":"inoue2015simd","t":"balkesen2013joins","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/2809974.2809988 -> 10.14778/2732219.2732227)","m":"reference-list extractor"},{"id":"e878","i":"inoue2015simd--EXTENDS-->chhugani2008sorting","s":"inoue2015simd","t":"chhugani2008sorting","r":"EXTENDS","w":2,"why":"Inoue and Taura solve the random-access record-rearrangement bottleneck in SIMD sorting of Arrays of Structures (AoS) by combining packed key-index SIMD sorting with cache-blocked record permutation.","m":"agent draft (gemini)"},{"id":"e880","i":"isenberg2011dualscale--CITES-->heer2010crowdsourcing","s":"isenberg2011dualscale","t":"heer2010crowdsourcing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2011.160 -> 10.1145/1753326.1753357)","m":"reference-list extractor"},{"id":"e883","i":"iso-5259--COMPLEMENTS-->johansson-2017","s":"iso-5259","t":"johansson-2017","r":"COMPLEMENTS","w":2,"why":"Provides data-quality lineage, missing-AIS interpolation bounds, and sensor-reconciliation controls for regulated vessel telemetry pipelines.","m":"agent draft (gemini)"},{"id":"e884","i":"iso-5259--OPERATIONALIZES-->iso-42001","s":"iso-5259","t":"iso-42001","r":"OPERATIONALIZES","w":3,"why":"ISO/IEC 5259-1..5:2025 operationalizes ISO/IEC 42001:2023 Annex A.7 (Data for AI systems) across quality, lineage, and governance","m":"seed curator"},{"id":"e885","i":"iso-5259--OPERATIONALIZES-->khatri2010designing","s":"iso-5259","t":"khatri2010designing","r":"OPERATIONALIZES","w":3,"why":"ISO/IEC 5259-1..5 operationalizes Khatri & Brown's Data Quality, Metadata, and Data Lifecycle decision domains for analytics and machine learning.","m":"agent draft (gemini)"},{"id":"e893","i":"jalkanen-2012--EXTENDS-->jalkanen-2009","s":"jalkanen-2012","t":"jalkanen-2009","r":"EXTENDS","w":2,"why":"Extends the STEAM AIS vessel-telemetry emission model from NOx and SOx to engine-load-dependent particulate matter (PM2.5, sulfate, organic carbon) and CO.","m":"agent draft (gemini)"},{"id":"e894","i":"jeong-2009--CITES-->bettenburg2008goodbug","s":"jeong2009bugtossing","t":"bettenburg2008goodbug","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1595696.1595715 -> 10.1145/1453101.1453146)","m":"reference-list extractor"},{"id":"e895","i":"jeong-2009--CITES-->hooimeijer-2007","s":"jeong2009bugtossing","t":"hooimeijer-2007","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1595696.1595715 -> 10.1145/1321631.1321639)","m":"reference-list extractor"},{"id":"e896","i":"jeong-2009--CITES-->panjer-2007","s":"jeong2009bugtossing","t":"panjer-2007","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1595696.1595715 -> 10.1109/msr.2007.25)","m":"reference-list extractor"},{"id":"e897","i":"jeong2009bugtossing--CITES-->weiss-2007","s":"jeong2009bugtossing","t":"weiss-2007","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1595696.1595715 -> 10.1109/msr.2007.13)","m":"reference-list extractor"},{"id":"e899","i":"jeyakumar2019explainit--CITES-->pelkonen2015gorilla","s":"jeyakumar2019explainit","t":"pelkonen2015gorilla","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3299869.3314048 -> 10.14778/2824032.2824078)","m":"reference-list extractor"},{"id":"e900","i":"jeyakumar2019explainit--EXTENDS-->yoon2016dbsherlock","s":"jeyakumar2019explainit","t":"yoon2016dbsherlock","r":"EXTENDS","w":2,"why":"Extends causal database/telemetry performance diagnosis with a declarative SQL-like hypothesis specification language","m":"agent draft (gemini)"},{"id":"e901","i":"ji2026selective--CITES-->xu2002prefix","s":"ji2026selective","t":"xu2002prefix","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.jisa.2026.104550 -> xu2002prefix)","m":"reference-list extractor"},{"id":"e902","i":"ji2026selective--COMPLEMENTS-->aghili2025logs","s":"ji2026selective","t":"aghili2025logs","r":"COMPLEMENTS","w":2,"why":"Ji et al. (2026) operationalize selective field-level protection in structured logs for the sensitive field taxonomy identified by Aghili et al. (2025).","m":"agent draft (gemini)"},{"id":"e904","i":"jia2024aicreativity--EXTENDS-->brynjolfsson-2025","s":"brynjolfsson-2025","t":"jia2024aicreativity","r":"EXTENDS","w":2,"why":"Field AI-in-service experiment where gains accrue to higher-skill workers, unlike the novice-skewed support-chat gains","m":"agent draft (gemini)"},{"id":"e906","i":"jiang2024lilac--EXTENDS-->xu2024divlog","s":"jiang2024lilac","t":"xu2024divlog","r":"EXTENDS","w":2,"why":"Wraps DivLog-style ICL LLM log parsing in an online Adaptive Parsing Cache (prefix tree) so LLM inference is invoked only once per distinct template (~280 calls/dataset) rather than per line.","m":"agent draft (gemini)"},{"id":"e907","i":"jiang2024lilac--SHARES_BENCHMARK-->loghub2_2024","s":"jiang2024lilac","t":"jiang2024loghub2","r":"SHARES_BENCHMARK","w":2,"why":"LILAC evaluates LLM log parsing with an adaptive cache on Loghub-2.0.","m":"agent draft (claude)"},{"id":"e908","i":"jiang2024lilac--USES_METHOD-->morrison-1968","s":"jiang2024lilac","t":"morrison-1968","r":"USES_METHOD","w":2,"why":"Implements its Adaptive Parsing Cache as a prefix/radix tree with wildcard nodes to match cached templates in O(|tokens|) before falling back to the LLM.","m":"agent draft (gemini)"},{"id":"e909","i":"jiang2024loghub2--EXTENDS-->khan2022guidelines","s":"jiang2024loghub2","t":"khan2022guidelines","r":"EXTENDS","w":2,"why":"Scales Khan et al.'s GA/PA/FGA/FTA evaluation framework from 2k-line samples to 50.4M lines across 14 systems (Loghub-2.0), exposing timeout and >=5-parameter accuracy collapse.","m":"agent draft (gemini)"},{"id":"e910","i":"jiang2024loghub2--EXTENDS-->zhu2019tools","s":"jiang2024loghub2","t":"zhu2019tools","r":"EXTENDS","w":2,"why":"Expands Loghub-2k (32k lines) by >1,500x to Loghub-2.0 (50.4M lines, 3,488 ground-truth templates) and shows that 9 of 15 parsers fail to complete within 12 hours.","m":"agent draft (gemini)"},{"id":"e915","i":"joel-2025-arxiv--CONTRASTS_WITH-->ziegler-2024","s":"joel-2025-arxiv","t":"ziegler-2024","r":"CONTRASTS_WITH","w":2,"why":"Perceived productivity (what acceptance rate predicts) diverged from measured speed: experts felt 20% faster but were 19% slower.","m":"agent draft (claude)"},{"id":"e916","i":"johanssen2019feedback--CITES-->guzman2014feature","s":"johanssen2019feedback","t":"guzman2014feature","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/re.2019.00026 -> 10.1109/re.2014.6912257)","m":"reference-list extractor"},{"id":"e917","i":"johanssen2019feedback--COMPLEMENTS-->maalej-2016","s":"johanssen2019feedback","t":"maalej-2016","r":"COMPLEMENTS","w":2,"why":"Practice gap: feedback rarely linked to usage data.","m":"agent draft (gemini)"},{"id":"e918","i":"johansson-2017--CITES-->jalkanen-2009","s":"johansson-2017","t":"jalkanen-2009","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.atmosenv.2017.08.042 -> 10.5194/acp-9-9209-2009)","m":"reference-list extractor"},{"id":"e919","i":"johansson-2017--EXTENDS-->jalkanen-2012","s":"johansson-2017","t":"jalkanen-2012","r":"EXTENDS","w":3,"why":"Scales bottom-up AIS transponder trajectory and Admiralty propulsion-law modeling (STEAM3) to global fleet CO2 and pollutant inventories.","m":"agent draft (gemini)"},{"id":"e922","i":"joyner-2022--CITES-->jung-2022","s":"joyner-2022","t":"jung-2022","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3491102.3517630 -> 10.1109/tvcg.2021.3114846)","m":"reference-list extractor"},{"id":"e923","i":"joyner-2022--CITES-->kong2018titles","s":"joyner-2022","t":"kong2018titles","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3491102.3517630 -> kong2018titles)","m":"reference-list extractor"},{"id":"e924","i":"joyner-2022--COMPLEMENTS-->elavsky-2022","s":"joyner-2022","t":"elavsky-2022","r":"COMPLEMENTS","w":2,"why":"why designers skip accessibility; Chartability is the checklist answer","m":"agent draft (gemini)"},{"id":"e925","i":"jung-2022--CITES-->kim2021accessiblevis","s":"jung-2022","t":"kim2021accessiblevis","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2021.3114846 -> 10.1111/cgf.14298)","m":"reference-list extractor"},{"id":"e926","i":"jung-2022--CITES-->kong2019titlemisalignment","s":"jung-2022","t":"kong2019titlemisalignment","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2021.3114846 -> 10.1145/3290605.3300576)","m":"reference-list extractor"},{"id":"e927","i":"jung-2022--COMPLEMENTS-->lundgard-2022","s":"jung-2022","t":"lundgard-2022","r":"COMPLEMENTS","w":2,"why":"empirical alt-text guidance for blind readers alongside four-level semantic model","m":"agent draft (claude)"},{"id":"e928","i":"just2014mutants--EXTENDS-->jia2011mutation","s":"just2014mutants","t":"jia2011mutation","r":"EXTENDS","w":2,"why":"Empirically validates Jia & Harman's mutation testing Coupling Effect across 357 real faults, controlling for coverage.","m":"agent draft (gemini)"},{"id":"e929","i":"k8s2026probes--COMPLEMENTS-->huang2017gray","s":"k8s2026probes","t":"huang2017gray","r":"COMPLEMENTS","w":2,"why":"Liveness probes observe from inside the pod; differential observability decides whether they see a disk stall.","m":"agent draft (claude)"},{"id":"e939","i":"kale-2019--CITES-->fernandes2018quantiledotplots","s":"kale-2019","t":"fernandes2018quantiledotplots","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2018.2864909 -> 10.1145/3173574.3173718)","m":"reference-list extractor"},{"id":"e940","i":"kale-2019--CITES-->ioannidis2005why","s":"kale-2019","t":"ioannidis2005why","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2018.2864909 -> 10.1371/journal.pmed.0020124)","m":"reference-list extractor"},{"id":"e941","i":"kale-2019--EXTENDS-->hullman2019pursuiterror","s":"kale-2019","t":"hullman2019pursuiterror","r":"EXTENDS","w":2,"why":"HOP evaluation within the uncertainty-vis evaluation agenda","m":"agent draft (gemini)"},{"id":"e942","i":"kale-2021--CITES-->correll-2014","s":"kale-2021","t":"correll-2014","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2020.3030335 -> 10.1109/tvcg.2014.2346298)","m":"reference-list extractor"},{"id":"e943","i":"kale-2021--CITES-->fernandes2018quantiledotplots","s":"kale-2021","t":"fernandes2018quantiledotplots","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2020.3030335 -> 10.1145/3173574.3173718)","m":"reference-list extractor"},{"id":"e944","i":"kale-2021--CITES-->kale-2019","s":"kale-2021","t":"kale-2019","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2020.3030335 -> 10.1109/tvcg.2018.2864909)","m":"reference-list extractor"},{"id":"e945","i":"kale-2021--COMPLEMENTS-->hofman-2020","s":"kale-2021","t":"hofman-2020","r":"COMPLEMENTS","w":2,"why":"reasoning strategies with uncertainty displays for effect sizes","m":"agent draft (claude)"},{"id":"e946","i":"kalibera2013rigorous--CITES-->blackburn-2006","s":"kalibera2013rigorous","t":"blackburn-2006","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2464157.2464160 -> 10.1145/1167473.1167488)","m":"reference-list extractor"},{"id":"e947","i":"kalibera2013rigorous--CITES-->curtsinger-2013","s":"kalibera2013rigorous","t":"curtsinger-2013","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2464157.2464160 -> 10.1145/2451116.2451141)","m":"reference-list extractor"},{"id":"e948","i":"kalibera2013rigorous--CITES-->mytkowicz2009wrong","s":"kalibera2013rigorous","t":"mytkowicz2009wrong","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2464157.2464160 -> 10.1145/1508244.1508275)","m":"reference-list extractor"},{"id":"e949","i":"kalibera2013rigorous--EXTENDS-->georges2007statistically","s":"kalibera2013rigorous","t":"georges2007statistically","r":"EXTENDS","w":2,"why":"Adds variance-driven repetition planning across levels and CI on ratios to Georges et al.'s CI methodology","m":"agent draft (gemini)"},{"id":"e950","i":"kamra2008anomalousaccess--COMPLEMENTS-->ma2018queryforecast","s":"kamra2008anomalousaccess","t":"ma2018queryforecast","r":"COMPLEMENTS","w":2,"why":"access-profile anomaly detection vs template-arrival forecasting: both build per-template workload fingerprints from query logs","m":"agent draft (gemini)"},{"id":"e951","i":"kanev-2015--CITES-->dean-2013","s":"kanev-2015","t":"dean-2013","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2749469.2750392 -> 10.1145/2408776.2408794)","m":"reference-list extractor"},{"id":"e952","i":"kanev-2015--CITES-->melnik-2010","s":"kanev-2015","t":"melnik-2010","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2749469.2750392 -> 10.14778/1920841.1920886)","m":"reference-list extractor"},{"id":"e953","i":"kanev-2015--CITES-->yasin-2014","s":"kanev-2015","t":"yasin-2014","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2749469.2750392 -> 10.1109/ispass.2014.6844459)","m":"reference-list extractor"},{"id":"e954","i":"kanev-2015--COMPLEMENTS-->ren-2010","s":"kanev-2015","t":"ren-2010","r":"COMPLEMENTS","w":2,"why":"GWP provides the continuous-profiling infrastructure; Kanev et al. analyse its fleet data to quantify the datacenter tax","m":"agent draft (gemini)"},{"id":"e956","i":"kapoor2023leakage--CITES-->gebru2021datasheets","s":"kapoor2023leakage","t":"gebru2021datasheets","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.patter.2023.100804 -> 10.1145/3458723)","m":"reference-list extractor"},{"id":"e960","i":"kapoor2023leakage--CONTRASTS_WITH-->montgomery2018escalationfe","s":"kapoor2023leakage","t":"montgomery2018escalationfe","r":"CONTRASTS_WITH","w":2,"why":"escalation predictors must exclude evidence collected after the escalation (temporal leakage)","m":"agent draft (gemini)"},{"id":"e964","i":"kcs2016v6--COMPLEMENTS-->maalej-2016","s":"kcs2016v6","t":"maalej-2016","r":"COMPLEMENTS","w":2,"why":"Support knowledge evolve loop as a product-feedback data stream","m":"agent draft (gemini)"},{"id":"e966","i":"keiichi-2016-arxiv--CONTRASTS_WITH-->du2016spell","s":"keiichi-2016-arxiv","t":"du2016spell","r":"CONTRASTS_WITH","w":2,"why":"Clusters log messages by cosine similarity over word character-length vectors (LenMa) rather than token-level LCS alignment.","m":"agent draft (gemini)"},{"id":"e967","i":"keiningham2007nps--COMPLEMENTS-->morgan2006metrics","s":"keiningham2007nps","t":"morgan2006metrics","r":"COMPLEMENTS","w":2,"why":"Both independently find NPS/recommend-intent no better (or worse) than satisfaction measures at predicting firm performance.","m":"agent draft (claude)"},{"id":"e968","i":"keiningham2007nps--CONTRASTS_WITH-->dixon2010ces","s":"keiningham2007nps","t":"dixon2010ces","r":"CONTRASTS_WITH","w":2,"why":"Both challenge single-metric loyalty claims; Keiningham finds NPS no better than other satisfaction metrics, Dixon proposes effort as the better driver.","m":"agent draft (claude)"},{"id":"e970","i":"kerbiriou2019pugz--OPERATIONALIZES-->mytkowicz2014fsm","s":"kerbiriou2019pugz","t":"mytkowicz2014fsm","r":"OPERATIONALIZES","w":2,"why":"pugz breaks the sequential state dependency of single-member Deflate bitstreams by speculatively synchronizing on dynamic Huffman block headers in parallel chunks and resolving 32 KiB LZ77 window back-references in a second stage.","m":"agent draft (gemini)"},{"id":"e971","i":"kerr1975folly--COMPLEMENTS-->strathern-1997","s":"kerr1975folly","t":"strathern-1997","r":"COMPLEMENTS","w":2,"why":"Both describe how measures used as targets distort behaviour (reward misalignment; Goodhart).","m":"agent draft (claude)"},{"id":"e972","i":"kersten2006mylyn--CITES-->gonzalez2004multitasking","s":"kersten2006mylyn","t":"gonzalez2004multitasking","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1181775.1181777 -> 10.1145/985692.985707)","m":"reference-list extractor"},{"id":"e973","i":"kersten2006mylyn--CITES-->parnas1972criteria","s":"kersten2006mylyn","t":"parnas1972criteria","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1181775.1181777 -> 10.1145/361598.361623)","m":"reference-list extractor"},{"id":"e974","i":"kersten2006mylyn--COMPLEMENTS-->xia2018comprehension","s":"kersten2006mylyn","t":"xia2018comprehension","r":"COMPLEMENTS","w":2,"why":"Mylyn attributes interaction to an active task; Xia et al. time-slice interaction into activity categories with idle thresholds. Combined, they define task-attributed active time from telemetry.","m":"agent draft (claude)"},{"id":"e975","i":"khan-2022--CITES-->dai2020logram","s":"khan2022guidelines","t":"dai2020logram","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3510003.3510101 -> 10.1109/tse.2020.3007554)","m":"reference-list extractor"},{"id":"e976","i":"khan-2022--CITES-->fu2009execution","s":"khan2022guidelines","t":"fu2009execution","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3510003.3510101 -> 10.1109/icdm.2009.60)","m":"reference-list extractor"},{"id":"e977","i":"khan-2022--CITES-->he2017drain","s":"khan2022guidelines","t":"he2017drain","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3510003.3510101 -> 10.1109/icws.2017.13)","m":"reference-list extractor"},{"id":"e978","i":"khan2022guidelines--REFUTES-->zhu2019tools","s":"khan2022guidelines","t":"zhu2019tools","r":"REFUTES","w":2,"why":"Proves that Grouping Accuracy (GA) in Zhu et al. (ICSE-SEIP 2019) masks severe template-identification errors and introduces message-level Parsing Accuracy (PA) and template-level FTA.","m":"agent draft (gemini)","tr":1},{"id":"e979","i":"khan2022guidelines--USES_METHOD-->loghub2_2024","s":"jiang2024loghub2","t":"khan2022guidelines","r":"USES_METHOD","w":2,"why":"Both evaluate template-identification accuracy; Loghub-2.0 adds large-scale datasets and template-level metrics.","m":"agent draft (gemini)"},{"id":"e981","i":"khatri2010designing--CITES-->redman1998dataquality","s":"khatri2010designing","t":"redman1998dataquality","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1629175.1629210 -> 10.1145/269012.269025)","m":"reference-list extractor"},{"id":"e982","i":"khatri2010designing--CITES-->wang1996dataquality","s":"khatri2010designing","t":"wang1996dataquality","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1629175.1629210 -> 10.1080/07421222.1996.11518099)","m":"reference-list extractor"},{"id":"e984","i":"khomh2012fasterreleases--CONTRASTS_WITH-->mockus2005customerquality","s":"khomh2012fasterreleases","t":"mockus2005customerquality","r":"CONTRASTS_WITH","w":2,"why":"release cadence and adoption timing confound field-quality comparison","m":"agent draft (gemini)"},{"id":"e985","i":"khomh2015rapidreleases--CITES-->sliwerski2005szz","s":"khomh2015rapidreleases","t":"sliwerski2005szz","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s10664-014-9308-x -> 10.1145/1083142.1083147)","m":"reference-list extractor"},{"id":"e986","i":"khomh2015rapidreleases--EXTENDS-->khomh2012fasterreleases","s":"khomh2015rapidreleases","t":"khomh2012fasterreleases","r":"EXTENDS","w":2,"why":"Journal extension of the Firefox rapid-release study","m":"agent draft (gemini)"},{"id":"e993","i":"killick2012optimal--COMPLEMENTS-->guo2017calibration","s":"killick2012optimal","t":"guo2017calibration","r":"COMPLEMENTS","w":3,"why":"Combines exact linear-time PELT regime changepoint detection with post-hoc probability calibration.","m":"agent draft (gemini)"},{"id":"e994","i":"kim-2014--CITES-->aksin2007moderncallcenter","s":"kim-2014","t":"aksin2007moderncallcenter","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1287/msom.2014.0490 -> 10.1111/j.1937-5956.2007.tb00288.x)","m":"reference-list extractor"},{"id":"e995","i":"kim-2014--EXTENDS-->brown-2005","s":"kim-2014","t":"brown-2005","r":"EXTENDS","w":2,"why":"Kim and Whitt refine the NHPP arrival test of Brown et al., showing pitfalls (rounding, overdispersion from pooling).","m":"agent draft (claude)"},{"id":"e996","i":"kim-2018--CITES-->heer2010crowdsourcing","s":"kim-2018","t":"heer2010crowdsourcing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/cgf.13409 -> 10.1145/1753326.1753357)","m":"reference-list extractor"},{"id":"e997","i":"kim-2018--CITES-->munzner2014vad","s":"kim-2018","t":"munzner2014vad","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/cgf.13409 -> 10.1201/b17511)","m":"reference-list extractor"},{"id":"e998","i":"kim-2018--CITES-->satyanarayan-2017","s":"kim-2018","t":"satyanarayan-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/cgf.13409 -> 10.1109/tvcg.2016.2599030)","m":"reference-list extractor"},{"id":"e999","i":"kim-2018--CONTRASTS_WITH-->cleveland1984graphical","s":"kim-2018","t":"cleveland1984graphical","r":"CONTRASTS_WITH","w":2,"why":"encoding effectiveness depends on task and data distribution, not a fixed ranking","m":"agent draft (claude)"},{"id":"e1000","i":"kim-2024--CITES-->satyanarayan-2017","s":"kim-2024","t":"satyanarayan-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3613904.3642442 -> 10.1109/tvcg.2016.2599030)","m":"reference-list extractor"},{"id":"e1001","i":"kim-2024--COMPLEMENTS-->zong-2024","s":"kim-2024","t":"zong-2024","r":"COMPLEMENTS","w":2,"why":"declarative sonification grammar alongside multi-modal representations","m":"agent draft (claude)"},{"id":"e1002","i":"kim2006bugfixtime--COMPLEMENTS-->kaplan-1958","s":"kim2006bugfixtime","t":"kaplan-1958","r":"COMPLEMENTS","w":2,"why":"Skewed, right-censored bug fix times are best summarized with survival estimators.","m":"agent draft (claude)"},{"id":"e1003","i":"kim2011crashgraphs--CITES-->glerum2009wer","s":"kim2011crashgraphs","t":"glerum2009wer","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/dsn.2011.5958261 -> 10.1145/1629575.1629586)","m":"reference-list extractor"},{"id":"e1004","i":"kim2011crashgraphs--COMPLEMENTS-->dang2012rebucket","s":"kim2011crashgraphs","t":"dang2012rebucket","r":"COMPLEMENTS","w":2,"why":"Aggregates multiple crash stacks within a bucket into a caller-callee transition graph to expose common fault convergence subgraphs across multi-node crashes.","m":"agent draft (gemini)"},{"id":"e1005","i":"kim2011topcrashes--CITES-->bettenburg2008goodbug","s":"kim2011topcrashes","t":"bettenburg2008goodbug","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2011.20 -> 10.1145/1453101.1453146)","m":"reference-list extractor"},{"id":"e1006","i":"kim2021accessiblevis--CITES-->kim-2018","s":"kim2021accessiblevis","t":"kim-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/cgf.14298 -> 10.1111/cgf.13409)","m":"reference-list extractor"},{"id":"e1007","i":"kim2021accessiblevis--CITES-->kong2019titlemisalignment","s":"kim2021accessiblevis","t":"kong2019titlemisalignment","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/cgf.14298 -> 10.1145/3290605.3300576)","m":"reference-list extractor"},{"id":"e1008","i":"kim2021accessiblevis--CITES-->shneiderman1996eyes","s":"kim2021accessiblevis","t":"shneiderman1996eyes","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/cgf.14298 -> 10.1109/vl.1996.545307)","m":"reference-list extractor"},{"id":"e1009","i":"kim2021accessiblevis--COMPLEMENTS-->elavsky-2022","s":"kim2021accessiblevis","t":"elavsky-2022","r":"COMPLEMENTS","w":2,"why":"design-space survey vs Chartability audit heuristics","m":"agent draft (claude)"},{"id":"e1010","i":"kim2021captions--COMPLEMENTS-->kong2018titles","s":"kim2021captions","t":"kong2018titles","r":"COMPLEMENTS","w":2,"why":"captions vs chart salience shape takeaways","m":"agent draft (claude)"},{"id":"e1011","i":"kim2021responsivecomm--CITES-->hullman-2020","s":"kim2021responsivecomm","t":"hullman-2020","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/cgf.14321 -> 10.1109/tvcg.2019.2934287)","m":"reference-list extractor"},{"id":"e1012","i":"kim2021responsivecomm--EXTENDS-->hoffswell2020responsive","s":"kim2021responsivecomm","t":"hoffswell2020responsive","r":"EXTENDS","w":2,"why":"empirical patterns/trade-offs of responsive transformations in communicative charts","m":"agent draft (claude)"},{"id":"e1013","i":"kim2024imnotsure--CITES-->bansal-2021","s":"kim2024imnotsure","t":"bansal-2021","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3630106.3658941 -> 10.1145/3411764.3445717)","m":"reference-list extractor"},{"id":"e1014","i":"kim2024imnotsure--CITES-->buçinca-2021","s":"kim2024imnotsure","t":"buçinca-2021","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3630106.3658941 -> 10.1145/3449287)","m":"reference-list extractor"},{"id":"e1015","i":"kim2024imnotsure--CITES-->hofman-2020","s":"kim2024imnotsure","t":"hofman-2020","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3630106.3658941 -> 10.1145/3313831.3376454)","m":"reference-list extractor"},{"id":"e1017","i":"kim2024imnotsure--CITES-->mitchell2019modelcards","s":"kim2024imnotsure","t":"mitchell2019modelcards","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3630106.3658941 -> 10.1145/3287560.3287596)","m":"reference-list extractor"},{"id":"e1018","i":"kim2024imnotsure--CITES-->nistai100_1","s":"kim2024imnotsure","t":"nistai100_1","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3630106.3658941 -> 10.6028/nist.ai.100-1)","m":"reference-list extractor"},{"id":"e1019","i":"kim2024imnotsure--CITES-->tian2023justask","s":"kim2024imnotsure","t":"tian2023justask","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3630106.3658941 -> 10.18653/v1/2023.emnlp-main.330)","m":"reference-list extractor"},{"id":"e1020","i":"kim2024imnotsure--CITES-->vasconcelos-2023","s":"kim2024imnotsure","t":"vasconcelos-2023","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3630106.3658941 -> 10.1145/3579605)","m":"reference-list extractor"},{"id":"e1021","i":"kim2024imnotsure--COMPLEMENTS-->zhang2020confidenceexplanation","s":"kim2024imnotsure","t":"zhang2020confidenceexplanation","r":"COMPLEMENTS","w":2,"why":"Both show uncertainty/confidence display calibrates reliance; Kim adds natural-language first-person hedges for LLMs.","m":"agent draft (claude)"},{"id":"e1023","i":"kim2025anarchy--CITES-->dong-2021","s":"kim2025anarchy","t":"dong-2021","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3725688.3725719 -> 10.1145/3483840)","m":"reference-list extractor"},{"id":"e1024","i":"kim2025anarchy--CITES-->huang2020tidb","s":"kim2025anarchy","t":"huang2020tidb","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3725688.3725719 -> 10.14778/3415478.3415535)","m":"reference-list extractor"},{"id":"e1025","i":"kim2025anarchy--COMPLEMENTS-->szlang2025snowflakeworkload","s":"kim2025anarchy","t":"szlang2025snowflakeworkload","r":"COMPLEMENTS","w":2,"why":"Feature/extension usage evidence for product decisions from vendor data","m":"agent draft (claude)"},{"id":"e1032","i":"knespel2023rapidgzip--EXTENDS-->kerbiriou2019pugz","s":"knespel2023rapidgzip","t":"kerbiriou2019pugz","r":"EXTENDS","w":2,"why":"rapidgzip generalizes pugz's two-stage speculative parallel gzip decompression from ASCII-only FASTQ streams to arbitrary binary/UTF-8 gzip streams using cache-prefetched 16-bit marker resolution and seek-point indexing.","m":"agent draft (gemini)"},{"id":"e1035","i":"knox2014complaints--COMPLEMENTS-->ascarza2018silent","s":"knox2014complaints","t":"ascarza2018silent","r":"COMPLEMENTS","w":2,"why":"Complainers carry elevated churn risk recoverable by good recovery; silent leavers are the complementary blind spot.","m":"agent draft (claude)"},{"id":"e1043","i":"kolesar-1998--COMPLEMENTS-->halfin-1981","s":"kolesar-1998","t":"halfin-1981","r":"COMPLEMENTS","w":2,"why":"Normal approximation to Erlang C gives practitioner square-root staffing consistent with Halfin-Whitt asymptotics.","m":"agent draft (claude)"},{"id":"e1044","i":"kong2018titles--EXTENDS-->borkin2016beyond","s":"kong2018titles","t":"borkin2016beyond","r":"EXTENDS","w":2,"why":"Builds on Borkin et al.'s finding that titles dominate eye fixations to quantify how title framing governs viewer recall.","m":"agent draft (gemini)"},{"id":"e1046","i":"kong2019titlemisalignment--EXTENDS-->kong2018titles","s":"kong2019titlemisalignment","t":"kong2018titles","r":"EXTENDS","w":2,"why":"title-visualization misalignment effects on trust and recall","m":"agent draft (claude)"},{"id":"e1047","i":"koole-2023--CITES-->aksin2007moderncallcenter","s":"koole-2023","t":"aksin2007moderncallcenter","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1287/stsy.2021.0008 -> 10.1111/j.1937-5956.2007.tb00288.x)","m":"reference-list extractor"},{"id":"e1048","i":"koole-2023--CITES-->brown-2005","s":"koole-2023","t":"brown-2005","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1287/stsy.2021.0008 -> 10.1198/016214504000001808)","m":"reference-list extractor"},{"id":"e1049","i":"koole-2023--CITES-->gans-2003","s":"koole-2023","t":"gans-2003","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1287/stsy.2021.0008 -> 10.1287/msom.5.2.79.16071)","m":"reference-list extractor"},{"id":"e1050","i":"koole-2023--CITES-->halfin-1981","s":"koole-2023","t":"halfin-1981","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1287/stsy.2021.0008 -> 10.1287/opre.29.3.567)","m":"reference-list extractor"},{"id":"e1051","i":"koole-2023--CITES-->ibrahim2016forecasting","s":"koole-2023","t":"ibrahim2016forecasting","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1287/stsy.2021.0008 -> 10.1016/j.ijforecast.2015.11.012)","m":"reference-list extractor"},{"id":"e1052","i":"koole-2023--CITES-->kaplan-1958","s":"koole-2023","t":"kaplan-1958","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1287/stsy.2021.0008 -> 10.1080/01621459.1958.10501452)","m":"reference-list extractor"},{"id":"e1053","i":"koole-2023--CITES-->kim-2014","s":"koole-2023","t":"kim-2014","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1287/stsy.2021.0008 -> 10.1287/msom.2014.0490)","m":"reference-list extractor"},{"id":"e1054","i":"krippendorff-2004--CITES-->cohen1960kappa","s":"krippendorff-2004","t":"cohen1960kappa","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1468-2958.2004.tb00738.x -> 10.1177/001316446002000104)","m":"reference-list extractor"},{"id":"e1058","i":"kuiper-2025--CITES-->leis-2015","s":"kuiper-2025","t":"leis-2015","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3742728.3742762 -> 10.14778/2850583.2850594)","m":"reference-list extractor"},{"id":"e1059","i":"kuiper-2025--CITES-->neumann-2011","s":"kuiper-2025","t":"neumann-2011","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3742728.3742762 -> 10.14778/2002938.2002940)","m":"reference-list extractor"},{"id":"e1060","i":"kuiper-2025--EXTENDS-->raasveldt2019duckdb","s":"kuiper-2025","t":"raasveldt2019duckdb","r":"EXTENDS","w":2,"why":"Out-of-core hash join implemented in DuckDB","m":"agent draft (gemini)"},{"id":"e1062","i":"kumar2012webcost--CITES-->aksin2007moderncallcenter","s":"kumar2012webcost","t":"aksin2007moderncallcenter","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1287/isre.1110.0390 -> 10.1111/j.1937-5956.2007.tb00288.x)","m":"reference-list extractor"},{"id":"e1063","i":"kumar2012webcost--CITES-->gans-2003","s":"kumar2012webcost","t":"gans-2003","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1287/isre.1110.0390 -> 10.1287/msom.5.2.79.16071)","m":"reference-list extractor"},{"id":"e1064","i":"kumar2012webcost--CONTRASTS_WITH-->xu2024ragkgcs","s":"kumar2012webcost","t":"xu2024ragkgcs","r":"CONTRASTS_WITH","w":2,"why":"Kumar&Telang: self-service deflection is heterogeneous and can raise contacts; Xu et al. report vendor-internal A/B gains from RAG over tickets. Independent, query-type-stratified evaluation is missing.","m":"agent draft (claude)"},{"id":"e1066","i":"kuschewski2023btrblocks--COMPLEMENTS-->zeng2023columnar","s":"kuschewski2023btrblocks","t":"zeng2023columnar","r":"COMPLEMENTS","w":2,"why":"Both argue decode speed over compression ratio on modern storage","m":"agent draft (gemini)"},{"id":"e1067","i":"kwon-2023--CITES-->dayan-2017","s":"kwon-2023","t":"dayan-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3582695 -> 10.1145/3035918.3064054)","m":"reference-list extractor"},{"id":"e1068","i":"kwon-2023--CITES-->dayan-2018","s":"kwon-2023","t":"dayan-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3582695 -> 10.1145/3183713.3196927)","m":"reference-list extractor"},{"id":"e1069","i":"kwon-2023--CITES-->dong-2021","s":"kwon-2023","t":"dong-2021","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3582695 -> dong-2021)","m":"reference-list extractor"},{"id":"e1070","i":"kwon-2023--CITES-->raju-2017","s":"kwon-2023","t":"raju-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3582695 -> 10.1145/3132747.3132765)","m":"reference-list extractor"},{"id":"e1071","i":"kwon-2023--EXTENDS-->balmau-2018","s":"kwon-2023","t":"balmau-2018","r":"EXTENDS","w":2,"why":"Co-designs LSM compaction scheduling with NVMe PLM deterministic windows beyond software-only SILK","m":"agent draft (gemini)"},{"id":"e1074","i":"laaber2019cloud--CITES-->georges2007statistically","s":"laaber2019cloud","t":"georges2007statistically","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s10664-019-09681-1 -> 10.1145/1297027.1297033)","m":"reference-list extractor"},{"id":"e1075","i":"laaber2019cloud--COMPLEMENTS-->kalibera2013rigorous","s":"laaber2019cloud","t":"kalibera2013rigorous","r":"COMPLEMENTS","w":2,"why":"Empirical cloud variability motivates variance-driven repetition and interleaved trials","m":"agent draft (gemini)"},{"id":"e1076","i":"laaber2019cloud--EXTENDS-->mytkowicz2009wrong","s":"laaber2019cloud","t":"mytkowicz2009wrong","r":"EXTENDS","w":2,"why":"Extends environmental measurement bias analysis to public cloud VMs, motivating paired single-instance execution schedules.","m":"agent draft (gemini)"},{"id":"e1077","i":"laaber2019cloud--USES_METHOD-->abedi2017rmit","s":"laaber2019cloud","t":"abedi2017rmit","r":"USES_METHOD","w":2,"why":"Laaber et al. (2019) adopt Abedi & Brecht (2017) interleaved trials on the same cloud instance to measure microbenchmark variability.","m":"agent draft (gemini)"},{"id":"e1078","i":"lamb2024datafusion--CITES-->armbrust-2020","s":"lamb2024datafusion","t":"armbrust-2020","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3626246.3653368 -> 10.14778/3415478.3415560)","m":"reference-list extractor"},{"id":"e1079","i":"lamb2024datafusion--CITES-->graefe2006sorting","s":"lamb2024datafusion","t":"graefe2006sorting","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3626246.3653368 -> 10.1145/1132960.1132964)","m":"reference-list extractor"},{"id":"e1081","i":"lamb2024datafusion--CITES-->leis-2014","s":"lamb2024datafusion","t":"leis-2014","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3626246.3653368 -> 10.1145/2588555.2610507)","m":"reference-list extractor"},{"id":"e1082","i":"lamb2024datafusion--CITES-->melnik-2010","s":"lamb2024datafusion","t":"melnik-2010","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3626246.3653368 -> 10.14778/1920841.1920886)","m":"reference-list extractor"},{"id":"e1083","i":"lamb2024datafusion--COMPLEMENTS-->raasveldt2019duckdb","s":"lamb2024datafusion","t":"raasveldt2019duckdb","r":"COMPLEMENTS","w":2,"why":"Both DataFusion (Rust/Arrow-native) and DuckDB (C++/vectorized) interoperate zero-copy in Python via the Apache Arrow C Data Interface (`__arrow_c_stream__` / `__arrow_c_array__`).","m":"agent draft (gemini)"},{"id":"e1084","i":"lamb2024datafusion--OPERATIONALIZES-->palkar2018weld","s":"lamb2024datafusion","t":"palkar2018weld","r":"OPERATIONALIZES","w":2,"why":"DataFusion and `arrow-rs` operationalize cross-library pipeline fusion over standardized Arrow memory buffers.","m":"agent draft (gemini)"},{"id":"e1085","i":"lamb2024datafusion--USES_METHOD-->zeng2023columnar","s":"lamb2024datafusion","t":"zeng2023columnar","r":"USES_METHOD","w":2,"why":"Relies on Apache Parquet dictionary/RLE encoding, zone-map page indexes, and late materialization evaluated by Zeng et al. (PVLDB 2023).","m":"agent draft (gemini)"},{"id":"e1086","i":"lamkanfi-2010--CITES-->bettenburg2008goodbug","s":"lamkanfi-2010","t":"bettenburg2008goodbug","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/msr.2010.5463284 -> 10.1145/1453101.1453146)","m":"reference-list extractor"},{"id":"e1087","i":"lamkanfi-2010--CITES-->jeong2009bugtossing","s":"lamkanfi-2010","t":"jeong2009bugtossing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/msr.2010.5463284 -> jeong2009bugtossing)","m":"reference-list extractor"},{"id":"e1088","i":"lamkanfi-2010--COMPLEMENTS-->tian-2016","s":"lamkanfi-2010","t":"tian-2016","r":"COMPLEMENTS","w":2,"why":"Severity is predictable from text at intake (Lamkanfi) yet the labels are noisy (Tian).","m":"agent draft (claude-sonnet-5-5)"},{"id":"e1093","i":"langdale2019simdjson--CITES-->mytkowicz2014fsm","s":"langdale2019simdjson","t":"mytkowicz2014fsm","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s00778-019-00578-5 -> 10.1145/2541940.2541988)","m":"reference-list extractor"},{"id":"e1094","i":"langdale2019simdjson--COMPLEMENTS-->muhlbauer2013instant","s":"langdale2019simdjson","t":"muhlbauer2013instant","r":"COMPLEMENTS","w":2,"why":"Mühlbauer et al.'s coarse-grained speculative byte-range chunking across threads complements simdjson's fine-grained 64-byte Stage-1 SIMD structural bitmask indexing within each chunk.","m":"agent draft (gemini)"},{"id":"e1095","i":"langdale2019simdjson--EXTENDS-->li2017mison","s":"langdale2019simdjson","t":"li2017mison","r":"EXTENDS","w":3,"why":"Extends Mison's speculative structural bitmask indexing to branchless two-stage SIMD JSON parsing at multi-gigabyte-per-second throughput.","m":"agent draft (gemini)"},{"id":"e1098","i":"le-2023--CITES-->dai2020logram","s":"le-2023","t":"dai2020logram","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00204 -> 10.1109/tse.2020.3007554)","m":"reference-list extractor"},{"id":"e1099","i":"le-2023--CITES-->du2016spell","s":"le-2023","t":"du2016spell","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00204 -> 10.1109/icdm.2016.0103)","m":"reference-list extractor"},{"id":"e1100","i":"le-2023--CITES-->fu2009execution","s":"le-2023","t":"fu2009execution","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00204 -> 10.1109/icdm.2009.60)","m":"reference-list extractor"},{"id":"e1101","i":"le-2023--CITES-->he2016experience","s":"le-2023","t":"he2016experience","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00204 -> 10.1109/issre.2016.21)","m":"reference-list extractor"},{"id":"e1102","i":"le-2023--CITES-->keiichi-2016-arxiv","s":"le-2023","t":"keiichi-2016-arxiv","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00204 -> keiichi-2016-arxiv)","m":"reference-list extractor"},{"id":"e1103","i":"le-2023--CITES-->le2022logad","s":"le-2023","t":"le2022logad","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00204 -> 10.1145/3510003.3510155)","m":"reference-list extractor"},{"id":"e1105","i":"le-2023--CITES-->wei2021logreducer","s":"le-2023","t":"wei2021logreducer","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00204 -> wei2021logreducer)","m":"reference-list extractor"},{"id":"e1107","i":"le-2023--CITES-->zhu2019tools","s":"le-2023","t":"zhu2019tools","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00204 -> 10.1109/icse-seip.2019.00021)","m":"reference-list extractor"},{"id":"e1108","i":"le-2023--CONTRASTS_WITH-->he2017drain","s":"le-2023","t":"he2017drain","r":"CONTRASTS_WITH","w":2,"why":"Frames log parsing as prompt-tuned few-shot RoBERTa token classification (LogPPT) rather than unsupervised prefix-tree clustering.","m":"agent draft (gemini)"},{"id":"e1114","i":"lee2009bounds--CITES-->rubin1976missing","s":"lee2009bounds","t":"rubin1976missing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1467-937x.2009.00536.x -> 10.1093/biomet/63.3.581)","m":"reference-list extractor"},{"id":"e1115","i":"lee2009bounds--COMPLEMENTS-->frangakis2002principal","s":"lee2009bounds","t":"frangakis2002principal","r":"COMPLEMENTS","w":2,"why":"Lee trimming bounds bound the always-selected principal-stratum effect under monotonicity","m":"agent draft (gemini)"},{"id":"e1116","i":"lee2023eadro_icse--CITES-->he2017drain","s":"lee2023eadro_icse","t":"he2017drain","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse48619.2023.00150 -> 10.1109/icws.2017.13)","m":"reference-list extractor"},{"id":"e1123","i":"leeflang-2009--CITES-->ransohoff1978spectrum","s":"leeflang-2009","t":"ransohoff1978spectrum","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.jclinepi.2008.04.007 -> 10.1056/nejm197810262991705)","m":"reference-list extractor"},{"id":"e1126","i":"leesatapornwongsa2016taxdc--COMPLEMENTS-->gunawi2014cloudbugs","s":"leesatapornwongsa2016taxdc","t":"gunawi2014cloudbugs","r":"COMPLEMENTS","w":2,"why":"Same group; TaxDC deepens the distributed-concurrency slice of the Cloud Bug Study","m":"agent draft (gemini)"},{"id":"e1128","i":"leinonen2023llmerrors--CITES-->barik2017readerrors","s":"leinonen2023llmerrors","t":"barik2017readerrors","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3545945.3569770 -> 10.1109/icse.2017.59)","m":"reference-list extractor"},{"id":"e1129","i":"leinonen2023llmerrors--COMPLEMENTS-->becker2019compilererrors","s":"leinonen2023llmerrors","t":"becker2019compilererrors","r":"COMPLEMENTS","w":2,"why":"LLM-generated explanations as an alternative to hand-enhanced messages.","m":"agent draft (gemini)"},{"id":"e1130","i":"leis-2014--COMPLEMENTS-->raasveldt2019duckdb","s":"leis-2014","t":"raasveldt2019duckdb","r":"COMPLEMENTS","w":2,"why":"DuckDB adopts morsel-driven parallel scheduling over row groups","m":"agent draft (gemini)"},{"id":"e1131","i":"leis-2015--EXTENDS-->selinger-1979","s":"leis-2015","t":"selinger-1979","r":"EXTENDS","w":2,"why":"Empirically evaluates Selinger-style cost-based optimization; cardinality estimation dominates.","m":"agent draft (gemini)"},{"id":"e1132","i":"lepers-2019--CITES-->o’neil-1996","s":"lepers-2019","t":"o’neil-1996","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3341301.3359628 -> o’neil-1996)","m":"reference-list extractor"},{"id":"e1133","i":"lepers-2019--CITES-->raju-2017","s":"lepers-2019","t":"raju-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3341301.3359628 -> 10.1145/3132747.3132765)","m":"reference-list extractor"},{"id":"e1134","i":"lepers-2019--COMPLEMENTS-->balmau-2018","s":"lepers-2019","t":"balmau-2018","r":"COMPLEMENTS","w":2,"why":"Shows how NVMe parallelism shifts KV store bottlenecks from disk bandwidth to CPU/memory-copy and compaction interference","m":"agent draft (gemini)"},{"id":"e1137","i":"lex-2014--CITES-->heer2010crowdsourcing","s":"lex-2014","t":"heer2010crowdsourcing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2014.2346248 -> 10.1145/1753326.1753357)","m":"reference-list extractor"},{"id":"e1138","i":"lex-2014--CONTRASTS_WITH-->kosara-2013","s":"lex-2014","t":"kosara-2013","r":"CONTRASTS_WITH","w":2,"why":"analytic set-overlap display vs narrative presentation","m":"agent draft (gemini)"},{"id":"e1141","i":"li2007tcloseness--CITES-->sweeney2002kanonymity","s":"li2007tcloseness","t":"sweeney2002kanonymity","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icde.2007.367856 -> 10.1142/s0218488502001648)","m":"reference-list extractor"},{"id":"e1142","i":"li2007tcloseness--EXTENDS-->machanavajjhala2007ldiversity","s":"li2007tcloseness","t":"machanavajjhala2007ldiversity","r":"EXTENDS","w":2,"why":"Extends l-diversity by bounding the Earth Mover's Distance (t-closeness) between equivalence-class sensitive attributes and the global population distribution.","m":"agent draft (gemini)"},{"id":"e1145","i":"li2017mison--CITES-->melnik-2010","s":"li2017mison","t":"melnik-2010","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3115404.3115416 -> 10.14778/1920841.1920886)","m":"reference-list extractor"},{"id":"e1146","i":"li2017mison--CITES-->muhlbauer2013instant","s":"li2017mison","t":"muhlbauer2013instant","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3115404.3115416 -> 10.14778/2556549.2556555)","m":"reference-list extractor"},{"id":"e1158","i":"lin-2016--CITES-->fu2009execution","s":"lin2016logcluster","t":"fu2009execution","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2889160.2889232 -> 10.1109/icdm.2009.60)","m":"reference-list extractor"},{"id":"e1160","i":"lin-2016--CITES-->wu2014crashlocator","s":"lin2016logcluster","t":"wu2014crashlocator","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2889160.2889232 -> 10.1145/2610384.2610386)","m":"reference-list extractor"},{"id":"e1161","i":"lin2013agnostic--COMPLEMENTS-->deng2013cuped","s":"lin2013agnostic","t":"deng2013cuped","r":"COMPLEMENTS","w":2,"why":"Lin 2013 justifies regression adjustment (with interactions) as CUPED's design-based foundation","m":"agent draft (gemini)"},{"id":"e1162","i":"lin2015cowic--CITES-->fu2009execution","s":"lin2015cowic","t":"fu2009execution","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/ccgrid.2015.45 -> 10.1109/icdm.2009.60)","m":"reference-list extractor"},{"id":"e1163","i":"lin2015cowic--CITES-->oliner2007supercomputers","s":"lin2015cowic","t":"oliner2007supercomputers","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/ccgrid.2015.45 -> 10.1109/dsn.2007.103)","m":"reference-list extractor"},{"id":"e1166","i":"lin2016logcluster--OPERATIONALIZES-->brodie2005symptom","s":"lin2016logcluster","t":"brodie2005symptom","r":"OPERATIONALIZES","w":2,"why":"LogCluster's knowledge base of previously seen log clusters operationalizes IBM's symptom-database idea for online services.","m":"agent draft (claude)"},{"id":"e1167","i":"lin2025sleep--EXTENDS-->snell2024scaling","s":"lin2025sleep","t":"snell2024scaling","r":"EXTENDS","w":2,"why":"Moves test-time compute scaling offline into sleep-time pre-computation over predictable repository/log contexts.","m":"agent draft (gemini)"},{"id":"e1169","i":"lingren2014preannotation--COMPLEMENTS-->goddard-2012","s":"lingren2014preannotation","t":"goddard-2012","r":"COMPLEMENTS","w":2,"why":"Pre-annotation bias in annotation is an instance of automation bias","m":"agent draft (gemini)"},{"id":"e1172","i":"lipsitch-2010--COMPLEMENTS-->hernan2004selection","s":"lipsitch-2010","t":"hernan2004selection","r":"COMPLEMENTS","w":2,"why":"Negative controls detect the selection structures Hernan describes.","m":"agent draft (gemini)"},{"id":"e1174","i":"liu-2012--CITES-->wood-1986","s":"liu-2012","t":"wood-1986","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.ergon.2012.09.001 -> 10.1016/0749-5978(86)90044-0)","m":"reference-list extractor"},{"id":"e1175","i":"liu-2012--EXTENDS-->campbell-1988","s":"liu-2012","t":"campbell-1988","r":"EXTENDS","w":2,"why":"review and conceptualisation framework","m":"agent draft (gemini)"},{"id":"e1176","i":"liu-2018--CITES-->cleveland1984graphical","s":"liu-2018","t":"cleveland1984graphical","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3173574.3174172 -> 10.1080/01621459.1984.10478080)","m":"reference-list extractor"},{"id":"e1177","i":"liu-2018--CITES-->szafir-2018","s":"liu-2018","t":"szafir-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3173574.3174172 -> 10.1109/tvcg.2017.2744359)","m":"reference-list extractor"},{"id":"e1178","i":"liu-2018--CONTRASTS_WITH-->borland-2007","s":"liu-2018","t":"borland-2007","r":"CONTRASTS_WITH","w":2,"why":"empirical comparison quantifying rainbow weakness and multi-hue alternatives","m":"agent draft (claude)"},{"id":"e1180","i":"liu2015opprentice--CITES-->davis2006prroc","s":"liu2015opprentice","t":"davis2006prroc","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2815675.2815679 -> 10.1145/1143844.1143874)","m":"reference-list extractor"},{"id":"e1181","i":"liu2019cloudincidents--EXTENDS-->gunawi2014cloudbugs","s":"liu2019cloudincidents","t":"gunawi2014cloudbugs","r":"EXTENDS","w":2,"why":"Moves from issue-tracker bug study to production-incident causes in cloud systems.","m":"agent draft (claude)"},{"id":"e1192","i":"lo-2022--CITES-->correll-2020","s":"lo-2022","t":"correll-2020","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/cgf.14559 -> 10.1145/3313831.3376222)","m":"reference-list extractor"},{"id":"e1193","i":"lo-2022--COMPLEMENTS-->mcnutt-2020","s":"lo-2022","t":"mcnutt-2020","r":"COMPLEMENTS","w":2,"why":"taxonomy of misleading designs vs pipeline-level mirages","m":"agent draft (claude)"},{"id":"e1194","i":"loghub2_2024--EXTENDS-->zhu2023loghub","s":"jiang2024loghub2","t":"zhu2023loghub","r":"EXTENDS","w":2,"why":"Loghub-2.0 re-annotates larger datasets from Loghub for template-level parser evaluation.","m":"agent draft (claude)"},{"id":"e1196","i":"lopezbernal2017its--COMPLEMENTS-->callaway2021did","s":"lopezbernal2017its","t":"callaway2021did","r":"COMPLEMENTS","w":2,"why":"ITS for single-start interventions; staggered DiD when bot rollout differs by region/product","m":"agent draft (gemini)"},{"id":"e1197","i":"lopezbernal2017its--USES_METHOD-->newey1987hac","s":"lopezbernal2017its","t":"newey1987hac","r":"USES_METHOD","w":2,"why":"HAC SEs are a standard ITS autocorrelation remedy","m":"agent draft (gemini)"},{"id":"e1198","i":"lopezbernal2018controls--CITES-->abadie2010sc","s":"lopezbernal2018controls","t":"abadie2010sc","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1093/ije/dyy135 -> 10.1198/jasa.2009.ap08746)","m":"reference-list extractor"},{"id":"e1199","i":"lopezbernal2018controls--EXTENDS-->lopezbernal2017its","s":"lopezbernal2018controls","t":"lopezbernal2017its","r":"EXTENDS","w":2,"why":"adds control series to single-group ITS","m":"agent draft (gemini)"},{"id":"e1200","i":"lopezbernal2018controls--EXTENDS-->wagner2002segmented","s":"lopezbernal2018controls","t":"wagner2002segmented","r":"EXTENDS","w":2,"why":"controls and placebos for segmented ITS","m":"agent draft (gemini)"},{"id":"e1202","i":"lu-2017--CITES-->o’neil-1996","s":"lu-2017","t":"o’neil-1996","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3033273 -> 10.1007/s002360050048)","m":"reference-list extractor"},{"id":"e1203","i":"lu-2017--COMPLEMENTS-->dong-2021","s":"lu-2017","t":"dong-2021","r":"COMPLEMENTS","w":2,"why":"Key-value separation (WiscKey / BlobDB) reduces LSM compaction WAF in RocksDB","m":"agent draft (gemini)"},{"id":"e1204","i":"lu2023perseus--CONTRASTS_WITH-->panda2019iaso","s":"lu2023perseus","t":"panda2019iaso","r":"CONTRASTS_WITH","w":2,"why":"Perseus uses device latency/throughput; IASO uses service-level timeouts seen by peers.","m":"agent draft (gemini)"},{"id":"e1205","i":"lu2023perseus--EXTENDS-->gunawi2018failslow","s":"lu2023perseus","t":"gunawi2018failslow","r":"EXTENDS","w":2,"why":"Perseus operationalises fail-slow detection at fleet scale","m":"agent draft (gemini)"},{"id":"e1206","i":"lu2023perseus--EXTENDS-->hao2016tailatstore","s":"lu2023perseus","t":"hao2016tailatstore","r":"EXTENDS","w":2,"why":"Perseus starts from peer evaluation (2x median, as in Tail at Store) and replaces it with a latency-vs-throughput regression bound.","m":"agent draft (gemini)"},{"id":"e1207","i":"lu2023perseus--OPERATIONALIZES-->gunawi2018failslow","s":"lu2023perseus","t":"gunawi2018failslow","r":"OPERATIONALIZES","w":2,"why":"Perseus turns the fail-slow phenomenon characterised by Gunawi et al. into a production detector using latency-vs-throughput peer regression.","m":"agent draft (claude)"},{"id":"e1208","i":"lu2025xinda--CONTRASTS_WITH-->lu2023perseus","s":"lu2025xinda","t":"lu2023perseus","r":"CONTRASTS_WITH","w":2,"why":"Xinda: slow-fault impact is non-monotonic and threshold-driven; Perseus scores slowdown x duration as a monotone risk.","m":"agent draft (claude)"},{"id":"e1209","i":"lu2025xinda--EXTENDS-->do-2013","s":"lu2025xinda","t":"do-2013","r":"EXTENDS","w":2,"why":"Xinda revisits limpware-style slow-fault tolerance 12 years after Limplock on modern systems.","m":"agent draft (gemini)"},{"id":"e1213","i":"lundgard-2022--CITES-->hullman-2011","s":"lundgard-2022","t":"hullman-2011","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2021.3114770 -> 10.1109/tvcg.2011.255)","m":"reference-list extractor"},{"id":"e1214","i":"lundgard-2022--CITES-->kim2021accessiblevis","s":"lundgard-2022","t":"kim2021accessiblevis","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2021.3114770 -> 10.1111/cgf.14298)","m":"reference-list extractor"},{"id":"e1215","i":"lundgard-2022--CITES-->kim2021captions","s":"lundgard-2022","t":"kim2021captions","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2021.3114770 -> 10.1145/3411764.3445443)","m":"reference-list extractor"},{"id":"e1216","i":"lundgard-2022--CITES-->kong2018titles","s":"lundgard-2022","t":"kong2018titles","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2021.3114770 -> 10.1145/3173574.3174012)","m":"reference-list extractor"},{"id":"e1217","i":"lundgard-2022--CITES-->kong2019titlemisalignment","s":"lundgard-2022","t":"kong2019titlemisalignment","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2021.3114770 -> 10.1145/3290605.3300576)","m":"reference-list extractor"},{"id":"e1218","i":"lundgard-2022--CITES-->nuñez-2018","s":"lundgard-2022","t":"nuñez-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2021.3114770 -> 10.1371/journal.pone.0199239)","m":"reference-list extractor"},{"id":"e1219","i":"lundgard-2022--CITES-->satyanarayan-2017","s":"lundgard-2022","t":"satyanarayan-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2021.3114770 -> 10.1109/tvcg.2016.2599030)","m":"reference-list extractor"},{"id":"e1220","i":"lundgard-2022--COMPLEMENTS-->elavsky-2022","s":"lundgard-2022","t":"elavsky-2022","r":"COMPLEMENTS","w":2,"why":"semantic levels of chart descriptions vs Chartability audit heuristics","m":"agent draft (gemini)"},{"id":"e1221","i":"luo2020lsm--CITES-->bloom1970space","s":"luo-2020","t":"bloom1970space","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s00778-019-00555-y -> 10.1145/362686.362692)","m":"reference-list extractor"},{"id":"e1222","i":"luo-2020--CITES-->bortnikov-2018","s":"luo-2020","t":"bortnikov-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3430915.3430916 -> 10.14778/3229863.3229873)","m":"reference-list extractor"},{"id":"e1223","i":"luo-2020--CITES-->dayan-2018","s":"luo-2020","t":"dayan-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3430915.3430916 -> 10.1145/3183713.3196927)","m":"reference-list extractor"},{"id":"e1224","i":"luo-2020--CITES-->dayan-2019","s":"luo-2020","t":"dayan-2019","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3430915.3430916 -> 10.1145/3299869.3319903)","m":"reference-list extractor"},{"id":"e1225","i":"luo2020lsm--CITES-->o’neil-1996","s":"luo-2020","t":"o’neil-1996","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s00778-019-00555-y -> 10.1007/s002360050048)","m":"reference-list extractor"},{"id":"e1226","i":"luo-2020--CITES-->raju-2017","s":"luo-2020","t":"raju-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3430915.3430916 -> 10.1145/3132747.3132765)","m":"reference-list extractor"},{"id":"e1229","i":"luo-2020--EXTENDS-->dayan-2017","s":"luo-2020","t":"dayan-2017","r":"EXTENDS","w":2,"why":"Extends Monkey's static memory model to dynamic partitioning between write buffers and read caches","m":"agent draft (gemini)"},{"id":"e1230","i":"luo2020lsm--OPERATIONALIZES-->guo2017calibration","s":"luo-2020","t":"guo2017calibration","r":"OPERATIONALIZES","w":3,"why":"Combines LSM-tree storage engine telemetry indicators with calibrated log-odds likelihood-ratio scoring.","m":"agent draft (gemini)"},{"id":"e1231","i":"ma2018queryforecast--CITES-->yoon2016dbsherlock","s":"ma2018queryforecast","t":"yoon2016dbsherlock","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3183713.3196908 -> 10.1145/2882903.2915218)","m":"reference-list extractor"},{"id":"e1235","i":"maalej-2016--CITES-->bettenburg2008goodbug","s":"maalej-2016","t":"bettenburg2008goodbug","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/ms.2015.153 -> 10.1145/1453101.1453146)","m":"reference-list extractor"},{"id":"e1236","i":"machanavajjhala2007ldiversity--CITES-->dinur2003revealing","s":"machanavajjhala2007ldiversity","t":"dinur2003revealing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1217299.1217302 -> 10.1145/773153.773173)","m":"reference-list extractor"},{"id":"e1237","i":"machanavajjhala2007ldiversity--EXTENDS-->sweeney2002kanonymity","s":"machanavajjhala2007ldiversity","t":"sweeney2002kanonymity","r":"EXTENDS","w":2,"why":"Extends k-anonymity with intra-equivalence-class l-diversity to defend against homogeneity and background-knowledge attribute disclosure attacks.","m":"agent draft (gemini)"},{"id":"e1239","i":"mackinnon-2018--CITES-->cameron2008cluster","s":"mackinnon-2018","t":"cameron2008cluster","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/ectj.12107 -> 10.1162/rest.90.3.414)","m":"reference-list extractor"},{"id":"e1240","i":"mackinnon-2018--CITES-->cameron2015practitioner","s":"mackinnon-2018","t":"cameron2015practitioner","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/ectj.12107 -> 10.3368/jhr.50.2.317)","m":"reference-list extractor"},{"id":"e1242","i":"mackinnon-2018--EXTENDS-->mackinnon2017wild","s":"mackinnon-2018","t":"mackinnon2017wild","r":"EXTENDS","w":2,"why":"Few treated clusters case of the wild cluster bootstrap.","m":"agent draft (gemini)"},{"id":"e1243","i":"mackinnon2017wild--EXTENDS-->cameron2008bootstrap","s":"mackinnon2017wild","t":"cameron2008cluster","r":"EXTENDS","w":2,"why":"Wild cluster bootstrap behaviour with few/unbalanced clusters.","m":"agent draft (gemini)"},{"id":"e1244","i":"maclure1991casecrossover--COMPLEMENTS-->hernan2004selection","s":"maclure1991casecrossover","t":"hernan2004selection","r":"COMPLEMENTS","w":2,"why":"Self-controlled case-crossover designs eliminate time-invariant between-subject confounding while Hernán et al. formalize conditioning-induced selection bias.","m":"agent draft (gemini)"},{"id":"e1246","i":"majumdar2018partition--CITES-->brewer-2000","s":"majumdar2018partition","t":"brewer-2000","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3158134 -> 10.1145/564585.564601)","m":"reference-list extractor"},{"id":"e1252","i":"manski-2000--EXTENDS-->lee2009bounds","s":"lee2009bounds","t":"manski-2000","r":"EXTENDS","w":2,"why":"Both tighten worst-case selection bounds by monotonicity; MTS/MIV on outcomes vs Lee's monotone selection into observation.","m":"agent draft (gemini)"},{"id":"e1255","i":"mark2016emailduration--EXTENDS-->gonzalez2004multitasking","s":"mark2016emailduration","t":"gonzalez2004multitasking","r":"EXTENDS","w":2,"why":"Same research line (Mark): from shadowed task fragmentation (~3 min per event, ~11 min per working sphere) to logged email duration (~1.5 h/day, 77 checks/day) with a 5-min inactivity cutoff.","m":"agent draft (claude)"},{"id":"e1262","i":"mcduff2025ddxllm--CITES-->landis-1977","s":"mcduff2025ddxllm","t":"landis-1977","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1038/s41586-025-08869-4 -> 10.2307/2529310)","m":"reference-list extractor"},{"id":"e1263","i":"mcduff2025ddxllm--EXTENDS-->kanjee2023gpt4ddx","s":"mcduff2025ddxllm","t":"kanjee2023gpt4ddx","r":"EXTENDS","w":2,"why":"Extends LLM DDx evaluation on NEJM cases with top-k accuracy and an assisted-clinician arm.","m":"agent draft (claude)"},{"id":"e1264","i":"mckeeman1998differential--COMPLEMENTS-->jia2011mutation","s":"mckeeman1998differential","t":"jia2011mutation","r":"COMPLEMENTS","w":2,"why":"Differential testing provides a reference oracle across implementations while mutation testing measures the fault-detection sensitivity of the differential test suite.","m":"agent draft (gemini)"},{"id":"e1266","i":"mcnutt-2020--CITES-->kong2018titles","s":"mcnutt-2020","t":"kong2018titles","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3313831.3376420 -> 10.1145/3173574.3174012)","m":"reference-list extractor"},{"id":"e1267","i":"mcnutt-2020--CITES-->satyanarayan-2017","s":"mcnutt-2020","t":"satyanarayan-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3313831.3376420 -> 10.1109/tvcg.2016.2599030)","m":"reference-list extractor"},{"id":"e1268","i":"meng2018paradoxes--CITES-->rubin1976missing","s":"meng2018paradoxes","t":"rubin1976missing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1214/18-aoas1161sf -> 10.1093/biomet/63.3.581)","m":"reference-list extractor"},{"id":"e1269","i":"meng2018paradoxes--COMPLEMENTS-->elliott2017nonprob","s":"meng2018paradoxes","t":"elliott2017nonprob","r":"COMPLEMENTS","w":2,"why":"Data defect correlation quantifies why a large biased fleet sample needs a reference sample.","m":"agent draft (gemini)"},{"id":"e1270","i":"meng2018paradoxes--COMPLEMENTS-->hernan2004selection","s":"meng2018paradoxes","t":"hernan2004selection","r":"COMPLEMENTS","w":2,"why":"Meng's data-defect correlation identity quantifies the effective-sample-size collapse caused by the selection mechanisms Hernán et al. formalize.","m":"agent draft (gemini)"},{"id":"e1272","i":"meyer2017worklife--CITES-->gonzalez2004multitasking","s":"meyer2017worklife","t":"gonzalez2004multitasking","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2017.2656886 -> 10.1145/985692.985707)","m":"reference-list extractor"},{"id":"e1273","i":"meyer2017worklife--CITES-->iqbal2007disruption","s":"meyer2017worklife","t":"iqbal2007disruption","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2017.2656886 -> 10.1145/1240624.1240730)","m":"reference-list extractor"},{"id":"e1274","i":"meyer2017worklife--CITES-->mark2016emailduration","s":"meyer2017worklife","t":"mark2016emailduration","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2017.2656886 -> 10.1145/2858036.2858262)","m":"reference-list extractor"},{"id":"e1275","i":"miranda-2003--CONTRASTS_WITH-->tanabe-2004","s":"miranda-2003","t":"tanabe-2004","r":"CONTRASTS_WITH","w":2,"why":"NAS measures realised workload from timed activities, ESI predicts resource need and acuity at triage; effort vs ex-ante prediction.","m":"agent draft (claude-sonnet-5-5)"},{"id":"e1276","i":"miranda-2003--EXTENDS-->reis-1996","s":"miranda-2003","t":"reis-1996","r":"EXTENDS","w":2,"why":"NAS replaces TISS-28 items with activity items weighted by measured nursing time.","m":"agent draft (claude)"},{"id":"e1277","i":"mitchell2019modelcards--COMPLEMENTS-->gebru2021datasheets","s":"mitchell2019modelcards","t":"gebru2021datasheets","r":"COMPLEMENTS","w":2,"why":"Model cards document models; datasheets document datasets; together they cover ISO 42001 A.6.2.7 and A.7 documentation evidence.","m":"agent draft (claude)"},{"id":"e1278","i":"mockus-2008--CITES-->mockus2005customerquality","s":"mockus-2008","t":"mockus2005customerquality","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1368088.1368190 -> 10.1145/1062455.1062506)","m":"reference-list extractor"},{"id":"e1279","i":"mockus-2008--COMPLEMENTS-->dey2020usageindependent","s":"mockus-2008","t":"dey2020usageindependent","r":"COMPLEMENTS","w":2,"why":"Both normalise field defects by deployment/usage exposure","m":"agent draft (gemini)"},{"id":"e1280","i":"mockus2005customerquality--EXTENDS-->musa1993opprofile","s":"mockus2005customerquality","t":"musa1993opprofile","r":"EXTENDS","w":2,"why":"field failure probability depends on deployment/usage profile; operational profile is the test-side counterpart","m":"agent draft (gemini)"},{"id":"e1281","i":"mockus2005customerquality--OPERATIONALIZES-->goel-1979","s":"mockus2005customerquality","t":"goel-1979","r":"OPERATIONALIZES","w":2,"why":"field quality per install aligned to deployment time operationalises reliability growth in the field","m":"agent draft (gemini)"},{"id":"e1282","i":"mockus2005customerquality--USES_METHOD-->kaplan-1958","s":"mockus2005customerquality","t":"kaplan-1958","r":"USES_METHOD","w":2,"why":"Field-defect timing by release can be modelled as censored time-to-event","m":"agent draft (gemini)"},{"id":"e1283","i":"mogul2019nines--COMPLEMENTS-->hauer2020meaningful","s":"mogul2019nines","t":"hauer2020meaningful","r":"COMPLEMENTS","w":2,"why":"Both argue single availability percentages hide duration structure.","m":"agent draft (gemini)"},{"id":"e1284","i":"mogul2019nines--EXTENDS-->treynor-2017","s":"mogul2019nines","t":"treynor-2017","r":"EXTENDS","w":2,"why":"Nines-are-not-enough adds outage-duration distribution to Treynor's MTBF/MTTR availability calculus","m":"agent draft (gemini)"},{"id":"e1286","i":"mokander2023auditingllm--CITES-->gebru2021datasheets","s":"mokander2023auditingllm","t":"gebru2021datasheets","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s43681-023-00289-2 -> 10.1145/3458723)","m":"reference-list extractor"},{"id":"e1288","i":"mokander2023auditingllm--COMPLEMENTS-->iso-42001","s":"mokander2023auditingllm","t":"iso-42001","r":"COMPLEMENTS","w":2,"why":"Three-layered LLM auditing (governance, model, application) maps onto ISO/IEC 42001 A.10 (suppliers), A.6.2.4 (V&V), and A.6.2.6 (operation and monitoring).","m":"agent draft (gemini)"},{"id":"e1289","i":"mokander2023auditingllm--EXTENDS-->brundage2020trustworthy","s":"mokander2023auditingllm","t":"brundage2020trustworthy","r":"EXTENDS","w":2,"why":"Extends Brundage et al.'s institutional verification mechanisms into a three-layered audit architecture covering governance, model, and application audits for large language models.","m":"agent draft (gemini)"},{"id":"e1290","i":"mokander2023auditingllm--EXTENDS-->raji2020closing","s":"mokander2023auditingllm","t":"raji2020closing","r":"EXTENDS","w":2,"why":"Extends single-organisation internal audit to a three-layer (governance/model/application) scheme specific to LLMs.","m":"agent draft (claude)"},{"id":"e1292","i":"montgomery2017escalation--COMPLEMENTS-->chen-2020","s":"montgomery2017escalation","t":"chen2020lidar","r":"COMPLEMENTS","w":2,"why":"Escalation prediction from customer context and incident linking are complementary support-analytics uses of ticket plus environment data.","m":"agent draft (claude)"},{"id":"e1293","i":"montgomery2017escalation--COMPLEMENTS-->knox2014complaints","s":"montgomery2017escalation","t":"knox2014complaints","r":"COMPLEMENTS","w":2,"why":"Escalation prediction in enterprise software support operationalizes complaint/escalation risk at account level.","m":"agent draft (claude)"},{"id":"e1294","i":"montgomery2018escalationfe--CITES-->davis2006prroc","s":"montgomery2018escalationfe","t":"davis2006prroc","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s00766-018-0292-3 -> 10.1145/1143844.1143874)","m":"reference-list extractor"},{"id":"e1295","i":"montgomery2018escalationfe--EXTENDS-->montgomery2017escalation","s":"montgomery2018escalationfe","t":"montgomery2017escalation","r":"EXTENDS","w":2,"why":"Journal extension of the RE'17 IBM escalation study with full feature engineering and evaluation.","m":"agent draft (gemini)"},{"id":"e1296","i":"montgomery2020escalationfe--COMPLEMENTS-->montgomery2017escalation","s":"montgomery2018escalationfe","t":"montgomery2017escalation","r":"COMPLEMENTS","w":2,"why":"journal extension of the RE'17 escalation study","m":"agent draft (gemini)"},{"id":"e1299","i":"moritz-2019--CITES-->heer2010crowdsourcing","s":"moritz-2019","t":"heer2010crowdsourcing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2018.2865240 -> 10.1145/1753326.1753357)","m":"reference-list extractor"},{"id":"e1300","i":"moritz-2019--CITES-->kim-2018","s":"moritz-2019","t":"kim-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2018.2865240 -> 10.1111/cgf.13409)","m":"reference-list extractor"},{"id":"e1301","i":"moritz-2019--CITES-->munzner2014vad","s":"moritz-2019","t":"munzner2014vad","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2018.2865240 -> 10.1201/b17511)","m":"reference-list extractor"},{"id":"e1302","i":"moritz-2019--CITES-->pandey-2015","s":"moritz-2019","t":"pandey-2015","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2018.2865240 -> 10.1145/2702123.2702608)","m":"reference-list extractor"},{"id":"e1303","i":"moritz-2019--CITES-->satyanarayan-2017","s":"moritz-2019","t":"satyanarayan-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2018.2865240 -> 10.1109/tvcg.2016.2599030)","m":"reference-list extractor"},{"id":"e1304","i":"moritz-2019--CITES-->szafir-2018","s":"moritz-2019","t":"szafir-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2018.2865240 -> 10.1109/tvcg.2017.2744359)","m":"reference-list extractor"},{"id":"e1305","i":"moritz-2019--COMPLEMENTS-->chen-2022","s":"moritz-2019","t":"chen-2022","r":"COMPLEMENTS","w":2,"why":"constraint-based design knowledge vs lint-and-fix framework","m":"agent draft (gemini)"},{"id":"e1306","i":"morris-2019--OPERATIONALIZES-->padilla-2021","s":"morris-2019","t":"padilla-2021","r":"COMPLEMENTS","w":2,"why":"KM bands + at-risk tables operationalize interval display for survival curves","m":"agent draft (gemini)"},{"id":"e1309","i":"muhlbauer2013instant--CITES-->neumann-2011","s":"muhlbauer2013instant","t":"neumann-2011","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/2556549.2556555 -> 10.14778/2002938.2002940)","m":"reference-list extractor"},{"id":"e1310","i":"muhlbauer2013instant--OPERATIONALIZES-->mytkowicz2014fsm","s":"muhlbauer2013instant","t":"mytkowicz2014fsm","r":"COMPLEMENTS","w":2,"why":"Instant Loading operationalizes speculative FSM boundary synchronization for multi-core CSV/text bulk loading by jumping to byte-slice offsets and scanning forward to the next unquoted record delimiter.","m":"agent draft (gemini)"},{"id":"e1312","i":"muralidhar-2025--OPERATIONALIZES-->shneiderman1996eyes","s":"muralidhar-2025","t":"shneiderman1996eyes","r":"OPERATIONALIZES","w":2,"why":"Operationalizes overview-first / details-on-demand progressive disclosure for AI diagnostic interfaces.","m":"agent draft (gemini)"},{"id":"e1313","i":"muse2020sqlsmells--CITES-->kaplan-1958","s":"muse2020sqlsmells","t":"kaplan-1958","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3379597.3387467 -> 10.1080/01621459.1958.10501452)","m":"reference-list extractor"},{"id":"e1314","i":"muse2020sqlsmells--COMPLEMENTS-->dintyala2020sqlcheck","s":"muse2020sqlsmells","t":"dintyala2020sqlcheck","r":"COMPLEMENTS","w":2,"why":"SQL code smell prevalence in repos complements SQLCheck detection","m":"agent draft (claude)"},{"id":"e1316","i":"muthen-1984--COMPLEMENTS-->takane-1987","s":"muthen-1984","t":"takane-1987","r":"COMPLEMENTS","w":2,"why":"Muthen's SEM with categorical indicators uses the same latent-response (tetrachoric/polychoric) formulation.","m":"agent draft (claude)"},{"id":"e1317","i":"muthen-1989--CITES-->muthen-1984","s":"muthen-1989","t":"muthen-1984","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/bf02296397 -> 10.1007/bf02294210)","m":"reference-list extractor"},{"id":"e1318","i":"muthen-1989--COMPLEMENTS-->swaminathan-rogers-1990","s":"muthen-1989","t":"swaminathan-rogers-1990","r":"COMPLEMENTS","w":2,"why":"MIMIC/multi-group latent-variable DIF and observed-score logistic-regression DIF are alternative detectors for the same item bias question.","m":"agent draft (claude-sonnet-5-5)"},{"id":"e1319","i":"mytkowicz-2010--CITES-->georges2007statistically","s":"mytkowicz-2010","t":"georges2007statistically","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1806596.1806618 -> 10.1145/1297027.1297033)","m":"reference-list extractor"},{"id":"e1320","i":"mytkowicz-2010--COMPLEMENTS-->gregg-2016","s":"mytkowicz-2010","t":"gregg-2016","r":"COMPLEMENTS","w":2,"why":"Sampling-profiler bias (yield-point sampling) caveats apply to the stack samples flame graphs render; triangulate profilers before trusting a hot frame.","m":"agent draft (gemini)"},{"id":"e1321","i":"mytkowicz-2010--EXTENDS-->mytkowicz2009wrong","s":"mytkowicz-2010","t":"mytkowicz2009wrong","r":"EXTENDS","w":2,"why":"Same group; moves from benchmark measurement bias to profiler attribution bias","m":"agent draft (gemini)"},{"id":"e1322","i":"mytkowicz2009wrong--CITES-->georges2007statistically","s":"mytkowicz2009wrong","t":"georges2007statistically","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1508244.1508275 -> 10.1145/1297027.1297033)","m":"reference-list extractor"},{"id":"e1323","i":"mytkowicz2009wrong--COMPLEMENTS-->curtsinger-2013","s":"mytkowicz2009wrong","t":"curtsinger-2013","r":"COMPLEMENTS","w":2,"why":"Stabilizer operationalizes a remedy for the layout/setup measurement bias identified by Mytkowicz et al.","m":"agent draft (gemini)"},{"id":"e1324","i":"mytkowicz2014fsm--EXTENDS-->blelloch1989scans","s":"mytkowicz2014fsm","t":"blelloch1989scans","r":"EXTENDS","w":2,"why":"Mytkowicz et al. generalize parallel prefix scans to finite-state machines by composing state-transition functions across input chunks and exploiting state convergence.","m":"agent draft (gemini)"},{"id":"e1332","i":"nakatumba2010resourcebehavior--USES_METHOD-->vanderaalst2016processmining","s":"vanderaalst2016processmining","t":"nakatumba2010resourcebehavior","r":"USES_METHOD","w":2,"why":"Applies process-mining organizational/time perspectives (formalised in van der Aalst's text) to estimate how resource service time varies with workload.","m":"agent draft (claude)"},{"id":"e1336","i":"nash2022void--CONTRASTS_WITH-->beyer2016sre","s":"nash2022void","t":"beyer2016sre","r":"CONTRASTS_WITH","w":2,"why":"MTTR-style metrics critiqued vs SRE postmortem metric use","m":"agent draft (gemini)"},{"id":"e1340","i":"ni2026genaialibaba--COMPLEMENTS-->yiwei-2026-arxiv","s":"ni2026genaialibaba","t":"yiwei-2026-arxiv","r":"COMPLEMENTS","w":2,"why":"Same Alibaba setting: assistant vs agentic AI; both find speed gains without objective-quality gains","m":"agent draft (gemini)"},{"id":"e1341","i":"ni2026genaialibaba--EXTENDS-->brynjolfsson-2025","s":"ni2026genaialibaba","t":"brynjolfsson-2025","r":"EXTENDS","w":2,"why":"RCT replication in chat support; adds objective-quality (retrial) null and top-performer quality decline via workflow disruption.","m":"agent draft (gemini)"},{"id":"e1343","i":"nicol2004oncall--COMPLEMENTS-->beyer2016sre","s":"nicol2004oncall","t":"beyer2016sre","r":"COMPLEMENTS","w":2,"why":"Health evidence on on-call burden complements SRE's operational on-call load limits (<=25% time on-call, <=2 incidents/12h shift).","m":"agent draft (claude)"},{"id":"e1352","i":"noda-2023--EXTENDS-->forsgren-2021","s":"noda-2023","t":"forsgren-2021","r":"EXTENDS","w":2,"why":"DevEx narrows SPACE to feedback loops, cognitive load and flow, measured through paired perception and workflow data.","m":"agent draft (claude)"},{"id":"e1353","i":"norman2010diagerror--CITES-->graber2005diagerror","s":"norman2010diagerror","t":"graber2005diagerror","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1365-2923.2009.03507.x -> 10.1001/archinte.165.13.1493)","m":"reference-list extractor"},{"id":"e1354","i":"norman2010diagerror--CONTRASTS_WITH-->croskerry2009universal","s":"norman2010diagerror","t":"croskerry2009universal","r":"CONTRASTS_WITH","w":2,"why":"Norman & Eva attribute diagnostic error mainly to knowledge deficits; Croskerry emphasises dual-process biases and debiasing.","m":"agent draft (claude)"},{"id":"e1360","i":"noufaily-2012--CITES-->spiegelhalter2012regulation","s":"noufaily-2012","t":"spiegelhalter2012regulation","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1002/sim.5595 -> 10.1111/j.1467-985x.2011.01010.x)","m":"reference-list extractor"},{"id":"e1361","i":"noufaily-2012--EXTENDS-->farrington-1996","s":"noufaily-2012","t":"farrington-1996","r":"EXTENDS","w":2,"why":"Extends the Farrington quasi-Poisson outbreak detection regression with negative-binomial dispersion modeling and improved baseline down-weighting of past aberrations.","m":"agent draft (gemini)"},{"id":"e1363","i":"nsvif2026instruction--COMPLEMENTS-->vidigal2026wtlformal","s":"nsvif2026instruction","t":"vidigal2026wtlformal","r":"COMPLEMENTS","w":2,"why":"Complements machine-checked Lean 4 parser verification with neuro-symbolic constraint verification for instruction-following agents.","m":"agent draft (gemini)"},{"id":"e1364","i":"nuñez-2018--CITES-->brettel-1997","s":"nuñez-2018","t":"brettel-1997","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1371/journal.pone.0199239 -> 10.1364/josaa.14.002647)","m":"reference-list extractor"},{"id":"e1365","i":"nuñez-2018--CITES-->szafir-2018","s":"nuñez-2018","t":"szafir-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1371/journal.pone.0199239 -> 10.1109/tvcg.2017.2744359)","m":"reference-list extractor"},{"id":"e1366","i":"nuñez-2018--USES_METHOD-->machado-2009","s":"nuñez-2018","t":"machado-2009","r":"USES_METHOD","w":2,"why":"cividis optimised under CVD simulation","m":"agent draft (claude)"},{"id":"e1367","i":"obermeyer-2019--CONTRASTS_WITH-->tanabe-2004","s":"obermeyer-2019","t":"tanabe-2004","r":"CONTRASTS_WITH","w":2,"why":"resource-based proxies of need can encode allocation bias","m":"agent draft (gemini)"},{"id":"e1369","i":"ong2025routellm--EXTENDS-->chen2024frugalgpt","s":"ong2025routellm","t":"chen2024frugalgpt","r":"EXTENDS","w":2,"why":"RouteLLM learns routers from preference data and benchmarks directly against FrugalGPT-style cascades","m":"agent draft (qwen)"},{"id":"e1372","i":"padilla-2021--CITES-->hullman2019pursuiterror","s":"padilla-2021","t":"hullman2019pursuiterror","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1002/9781118445112.stat08296 -> 10.1109/tvcg.2018.2864889)","m":"reference-list extractor"},{"id":"e1373","i":"padilla-2021--CITES-->kale-2019","s":"padilla-2021","t":"kale-2019","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1002/9781118445112.stat08296 -> 10.1109/tvcg.2018.2864909)","m":"reference-list extractor"},{"id":"e1374","i":"padilla-2021--CITES-->munzner2014vad","s":"padilla-2021","t":"munzner2014vad","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1002/9781118445112.stat08296 -> 10.1201/b17511)","m":"reference-list extractor"},{"id":"e1376","i":"pagano-2013--CITES-->zimmermann2010bugreport","s":"pagano-2013","t":"bettenburg2008goodbug","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/re.2013.6636712 -> 10.1109/tse.2010.63)","m":"reference-list extractor"},{"id":"e1377","i":"pagano-2013--COMPLEMENTS-->maalej-2016","s":"pagano-2013","t":"maalej-2016","r":"COMPLEMENTS","w":2,"why":"Explicit user feedback (reviews) as one of the two data streams of data-driven RE","m":"agent draft (gemini)"},{"id":"e1381","i":"palkar2018weld--CITES-->neumann-2011","s":"palkar2018weld","t":"neumann-2011","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3213880.3213890 -> 10.14778/2002938.2002940)","m":"reference-list extractor"},{"id":"e1382","i":"panda2019iaso--OPERATIONALIZES-->gunawi2018failslow","s":"panda2019iaso","t":"gunawi2018failslow","r":"OPERATIONALIZES","w":2,"why":"IASO turns fail-slow evidence into detection+mitigation in distributed storage","m":"agent draft (gemini)"},{"id":"e1383","i":"panda2019iaso--OPERATIONALIZES-->huang2017gray","s":"panda2019iaso","t":"huang2017gray","r":"OPERATIONALIZES","w":2,"why":"IASO detects gray (fail-slow) nodes from peers' observations, the differential-observability idea.","m":"agent draft (gemini)"},{"id":"e1384","i":"pandey-2015--CITES-->borland-2007","s":"pandey-2015","t":"borland-2007","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2702123.2702608 -> 10.1109/mcg.2007.323435)","m":"reference-list extractor"},{"id":"e1385","i":"pandey-2015--CITES-->cleveland1984graphical","s":"pandey-2015","t":"cleveland1984graphical","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2702123.2702608 -> 10.1080/01621459.1984.10478080)","m":"reference-list extractor"},{"id":"e1386","i":"pandey-2015--CITES-->segel2010narrative","s":"pandey-2015","t":"segel2010narrative","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2702123.2702608 -> 10.1109/tvcg.2010.179)","m":"reference-list extractor"},{"id":"e1387","i":"pandey-2015--CITES-->tufte2001vdqi","s":"pandey-2015","t":"tufte2001vdqi","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2702123.2702608 -> tufte2001vdqi)","m":"reference-list extractor"},{"id":"e1388","i":"pandey-2015--COMPLEMENTS-->correll-2020","s":"pandey-2015","t":"correll-2020","r":"COMPLEMENTS","w":2,"why":"empirical effect of deceptive encodings incl. truncated axes","m":"agent draft (gemini)"},{"id":"e1390","i":"pang2006devil--EXTENDS-->xu2002prefix","s":"pang2006devil","t":"xu2002prefix","r":"EXTENDS","w":2,"why":"Applies prefix-preserving IP pseudonymization alongside multi-layer payload/header transformation and highlights operational verification traps.","m":"agent draft (gemini)"},{"id":"e1392","i":"panjer-2007--CITES-->kim2006bugfixtime","s":"panjer-2007","t":"kim2006bugfixtime","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/msr.2007.25 -> 10.1145/1137983.1138027)","m":"reference-list extractor"},{"id":"e1393","i":"panjer-2007--COMPLEMENTS-->giger-2010","s":"panjer-2007","t":"giger-2010","r":"COMPLEMENTS","w":2,"why":"Both predict bug lifetime/fix time from report attributes; modest accuracy supports treating time as a weak difficulty proxy.","m":"agent draft (claude-sonnet-5-5)"},{"id":"e1396","i":"pappalardo2019events--COMPLEMENTS-->ridder1994down","s":"pappalardo2019events","t":"ridder1994down","r":"COMPLEMENTS","w":2,"why":"Provides multi-league spatio-temporal event logs (fouls, cards, passes, shots) enabling match-level replication of red-card and foul-impunity models.","m":"agent draft (gemini)"},{"id":"e1397","i":"parasuraman2010complacency--COMPLEMENTS-->euaiact2024","s":"parasuraman2010complacency","t":"euaiact2024","r":"COMPLEMENTS","w":3,"why":"Provides the human-factors empirical foundation for EU AI Act Article 14(4)(b) automation bias mitigation and ISO/IEC 42001 Annex A.9 responsible use.","m":"agent draft (gemini)"},{"id":"e1401","i":"parnas1972criteria--COMPLEMENTS-->brooks1987nosilver","s":"parnas1972criteria","t":"brooks1987nosilverbullet","r":"COMPLEMENTS","w":2,"why":"Parnas's information-hiding module decomposition provides the structural mechanism to isolate Brooks's essential domain complexity from accidental representation complexity.","m":"agent draft (gemini)"},{"id":"e1402","i":"parnas1979extension--EXTENDS-->parnas1972criteria","s":"parnas1979extension","t":"parnas1972criteria","r":"EXTENDS","w":2,"why":"Extends information-hiding modules (1972) with the acyclic 'uses' hierarchy required for software extension and subset contraction (1979).","m":"agent draft (gemini)"},{"id":"e1405","i":"patel2015vibration--CITES-->ioannidis2005why","s":"patel2015vibration","t":"ioannidis2005why","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.jclinepi.2015.05.029 -> 10.1371/journal.pmed.0020124)","m":"reference-list extractor"},{"id":"e1406","i":"patel2015vibration--COMPLEMENTS-->young2017modeluncertainty","s":"patel2015vibration","t":"young2017modeluncertainty","r":"COMPLEMENTS","w":2,"why":"both quantify estimate instability across model specs","m":"agent draft (gemini)"},{"id":"e1408","i":"paul2022planencoders--CITES-->leis-2015","s":"paul2022planencoders","t":"leis-2015","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3503585.3503600 -> 10.14778/2850583.2850594)","m":"reference-list extractor"},{"id":"e1413","i":"peerally-2016--COMPLEMENTS-->card-2016","s":"peerally-2016","t":"card-2016","r":"COMPLEMENTS","w":2,"why":"Both BMJ QS 2016 critiques of RCA practice: Card targets the 5 Whys technique, Peerally the wider RCA process (weak actions, no recurrence tracking).","m":"agent draft (gemini)"},{"id":"e1414","i":"pereira2021ranking--EXTENDS-->pereira2017energy","s":"pereira2021ranking","t":"pereira2017energy","r":"EXTENDS","w":2,"why":"Extends the cross-language RAPL energy, runtime, and peak-memory benchmark corpus across 27 programming languages and normalizes energy-delay profiles.","m":"agent draft (gemini)"},{"id":"e1415","i":"pezeshkpour2024order--COMPLEMENTS-->zheng2023judging","s":"pezeshkpour2024order","t":"zheng2023judging","r":"COMPLEMENTS","w":2,"why":"Pezeshkpour & Hruschka (2024) quantify multiple-choice option-order sensitivity alongside Zheng et al.'s (2023) pairwise position bias in LLM-as-a-Judge.","m":"agent draft (gemini)"},{"id":"e1416","i":"pezeshkpour2024order--EXTENDS-->zhao2021calibrate","s":"pezeshkpour2024order","t":"zhao2021calibrate","r":"EXTENDS","w":2,"why":"Pezeshkpour & Hruschka (2024) extend contextual calibration (Zhao et al., 2021) to option-order sensitivity in LLM multiple-choice selection.","m":"agent draft (gemini)"},{"id":"e1419","i":"pham2025rcaeval--OPERATIONALIZES-->lee2023eadro_icse","s":"pham2025rcaeval","t":"lee2023eadro_icse","r":"OPERATIONALIZES","w":2,"why":"RCAEval provides multi-source (metrics, logs, traces) benchmark datasets and baselines on which multi-source methods such as Eadro are evaluated.","m":"agent draft (claude)"},{"id":"e1420","i":"phatak2021dirty--COMPLEMENTS-->wright2003professional","s":"phatak2021dirty","t":"wright2003professional","r":"COMPLEMENTS","w":2,"why":"Compares foul-to-card conversion ratios across European leagues, showing how lenient booking thresholds create systematic incentives for tactical fouling.","m":"agent draft (gemini)"},{"id":"e1421","i":"pika2017resourceprofiles--CITES-->vanderaalst2016processmining","s":"pika2017resourceprofiles","t":"vanderaalst2016processmining","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3041218 -> 10.1007/978-3-662-49851-4)","m":"reference-list extractor"},{"id":"e1422","i":"pika2017resourceprofiles--EXTENDS-->nakatumba2010resourcebehavior","s":"pika2017resourceprofiles","t":"nakatumba2010resourcebehavior","r":"EXTENDS","w":2,"why":"Pika et al. (with van der Aalst) generalise resource-behaviour analysis from workload-speed effects to a full indicator set of resource profiles mined from event logs.","m":"agent draft (claude)"},{"id":"e1432","i":"ports2012ssipg--OPERATIONALIZES-->cahill2008ssi","s":"ports2012ssipg","t":"cahill2008ssi","r":"OPERATIONALIZES","w":2,"why":"PostgreSQL implementation of SSI","m":"agent draft (gemini)"},{"id":"e1440","i":"psaraftis-2021--COMPLEMENTS-->johansson-2017","s":"psaraftis-2021","t":"johansson-2017","r":"COMPLEMENTS","w":2,"why":"Pairs bottom-up vessel emission inventories with EU ETS, FuelEU Maritime, and IMO carbon-pricing market-based compliance mechanisms.","m":"agent draft (gemini)"},{"id":"e1441","i":"pullenayegum-2016--CITES-->rubin1976missing","s":"pullenayegum-2016","t":"rubin1976missing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1177/0962280214536537 -> 10.1093/biomet/63.3.581)","m":"reference-list extractor"},{"id":"e1443","i":"putnick-2016--CITES-->meredith-1993","s":"putnick-2016","t":"meredith-1993","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.dr.2016.06.004 -> 10.1007/bf02294825)","m":"reference-list extractor"},{"id":"e1444","i":"putnick-2016--OPERATIONALIZES-->vandenberg-2000","s":"putnick-2016","t":"vandenberg-2000","r":"OPERATIONALIZES","w":2,"why":"Putnick and Bornstein turn the Vandenberg-Lance invariance sequence into reporting conventions and a state-of-practice review.","m":"agent draft (claude-sonnet-5-5)"},{"id":"e1446","i":"qiu2013schemacoevolution--CITES-->curino2008prism","s":"qiu2013schemacoevolution","t":"curino2008prism","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2491411.2491431 -> 10.14778/1453856.1453939)","m":"reference-list extractor"},{"id":"e1447","i":"qiu2013schemacoevolution--COMPLEMENTS-->curino2008wikipedia","s":"qiu2013schemacoevolution","t":"curino2008wikipedia","r":"COMPLEMENTS","w":2,"why":"Adds schema-code co-change analysis to schema change taxonomy","m":"agent draft (claude)"},{"id":"e1448","i":"raasveldt2017hostage--CITES-->neumann-2011","s":"raasveldt2017hostage","t":"neumann-2011","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3115404.3115408 -> 10.14778/2002938.2002940)","m":"reference-list extractor"},{"id":"e1449","i":"raasveldt2019duckdb--CITES-->neumann-2011","s":"raasveldt2019duckdb","t":"neumann-2011","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3299869.3320212 -> 10.14778/2002938.2002940)","m":"reference-list extractor"},{"id":"e1450","i":"raasveldt2019duckdb--COMPLEMENTS-->langdale2019simdjson","s":"raasveldt2019duckdb","t":"langdale2019simdjson","r":"COMPLEMENTS","w":3,"why":"Pairs DuckDB's vectorized execution and Parquet zone-map pushdown with simdjson's branchless structural JSON indexing for high-throughput analytical ingestion.","m":"agent draft (gemini)"},{"id":"e1451","i":"raasveldt2019duckdb--EXTENDS-->raasveldt2017hostage","s":"raasveldt2019duckdb","t":"raasveldt2017hostage","r":"EXTENDS","w":2,"why":"DuckDB's embedded zero-copy Arrow/DataFrame scan/export directly operationalizes the zero-serialization protocol principles of Raasveldt & Mühleisen (2017).","m":"agent draft (gemini)"},{"id":"e1456","i":"raji2020closing--CITES-->breck2017mltestscore","s":"raji2020closing","t":"breck2017mltestscore","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3351095.3372873 -> 10.1109/bigdata.2017.8258038)","m":"reference-list extractor"},{"id":"e1457","i":"raji2020closing--COMPLEMENTS-->iso-42001","s":"raji2020closing","t":"iso-42001","r":"COMPLEMENTS","w":2,"why":"SMACTR internal audit trail complements the management-system audit of cl. 9.2 with artifact-level evidence","m":"agent draft (gemini)"},{"id":"e1458","i":"raji2020closing--USES_METHOD-->mitchell2019modelcards","s":"raji2020closing","t":"mitchell2019modelcards","r":"USES_METHOD","w":2,"why":"SMACTR Artifact Collection stage requires model cards and datasheets as audit artifacts.","m":"agent draft (claude)"},{"id":"e1459","i":"raju-2017--CITES-->bloom1970space","s":"raju-2017","t":"bloom1970space","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3132747.3132765 -> 10.1145/362686.362692)","m":"reference-list extractor"},{"id":"e1460","i":"raju-2017--CITES-->o’neil-1996","s":"raju-2017","t":"o’neil-1996","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3132747.3132765 -> 10.1007/s002360050048)","m":"reference-list extractor"},{"id":"e1461","i":"raju-2017--CONTRASTS_WITH-->lu-2017","s":"raju-2017","t":"lu-2017","r":"CONTRASTS_WITH","w":2,"why":"Reduces LSM WAF via guard-based Fragmented LSM (FLSM) without separating values into a vLog","m":"agent draft (gemini)"},{"id":"e1462","i":"rambachan2023honest--CITES-->callaway2021did","s":"rambachan2023honest","t":"callaway2021did","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1093/restud/rdad018 -> 10.1016/j.jeconom.2020.12.001)","m":"reference-list extractor"},{"id":"e1463","i":"rambachan2023honest--CITES-->goodmanbacon2021decomp","s":"rambachan2023honest","t":"goodmanbacon2021decomp","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1093/restud/rdad018 -> 10.1016/j.jeconom.2021.03.014)","m":"reference-list extractor"},{"id":"e1464","i":"rambachan2023honest--CITES-->sun2021event","s":"rambachan2023honest","t":"sun2021event","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1093/restud/rdad018 -> 10.1016/j.jeconom.2020.09.006)","m":"reference-list extractor"},{"id":"e1465","i":"rambachan2023honest--EXTENDS-->roth2023trending","s":"rambachan2023honest","t":"roth2023trending","r":"COMPLEMENTS","w":2,"why":"honest DiD sensitivity is one of the tools synthesised in the review","m":"agent draft (gemini)"},{"id":"e1467","i":"redbench2026--CITES-->leis-2015","s":"redbench2026","t":"leis-2015","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3819518.3819538 -> 10.14778/2850583.2850594)","m":"reference-list extractor"},{"id":"e1468","i":"redbench2026--CITES-->szlang2025snowflakeworkload","s":"redbench2026","t":"szlang2025snowflakeworkload","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3819518.3819538 -> 10.14778/3750601.3750632)","m":"reference-list extractor"},{"id":"e1469","i":"redbench2026--USES_METHOD-->vanrenen2024redshiftfleet","s":"redbench2026","t":"vanrenen2024redshiftfleet","r":"USES_METHOD","w":2,"why":"Synthesizes workloads from Redset cloud traces","m":"agent draft (claude)"},{"id":"e1470","i":"redman1998dataquality--COMPLEMENTS-->sambasivan2021cascades","s":"redman1998dataquality","t":"sambasivan2021cascades","r":"COMPLEMENTS","w":2,"why":"Both: downstream cost of data-quality debt; Redman enterprise cost, Sambasivan compounding cascades in AI","m":"agent draft (gemini)"},{"id":"e1472","i":"reise-2012--CITES-->takane-1987","s":"reise-2012","t":"takane-1987","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1080/00273171.2012.715555 -> 10.1007/bf02294363)","m":"reference-list extractor"},{"id":"e1473","i":"reise-2012--COMPLEMENTS-->campbell-1959-mtmm","s":"reise-2012","t":"campbell-1959-mtmm","r":"COMPLEMENTS","w":2,"why":"Bifactor models with method factors are the latent-variable form of MTMM method variance.","m":"agent draft (claude)"},{"id":"e1474","i":"reitsma-2009--CITES-->begg1983verification","s":"reitsma-2009","t":"begg1983verification","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.jclinepi.2009.02.005 -> 10.2307/2530820)","m":"reference-list extractor"},{"id":"e1476","i":"reitsma-2009--CITES-->ransohoff1978spectrum","s":"reitsma-2009","t":"ransohoff1978spectrum","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.jclinepi.2009.02.005 -> 10.1056/nejm197810262991705)","m":"reference-list extractor"},{"id":"e1477","i":"reitsma-2009--COMPLEMENTS-->dawid1979maximum","s":"reitsma-2009","t":"dawid1979maximum","r":"COMPLEMENTS","w":2,"why":"Reitsma et al. list latent-class models (of which Dawid-Skene is the multi-rater EM form) among the solutions for imperfect reference standards in accuracy studies.","m":"agent draft (claude)"},{"id":"e1478","i":"remil2021subgroupsql--CITES-->chaudhuri2002compressing","s":"remil2021subgroupsql","t":"chaudhuri2002compressing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/ase51524.2021.9678915 -> 10.1145/564691.564747)","m":"reference-list extractor"},{"id":"e1479","i":"remil2021subgroupsql--CITES-->chen2014ormantipatterns","s":"remil2021subgroupsql","t":"chen2014ormantipatterns","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/ase51524.2021.9678915 -> 10.1145/2568225.2568259)","m":"reference-list extractor"},{"id":"e1480","i":"remil2021subgroupsql--CITES-->deep2020workloadcompression","s":"remil2021subgroupsql","t":"deep2020workloadcompression","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/ase51524.2021.9678915 -> 10.14778/3430915.3430931)","m":"reference-list extractor"},{"id":"e1481","i":"remil2021subgroupsql--COMPLEMENTS-->kul2018sqlclustering","s":"remil2021subgroupsql","t":"kul2018sqlclustering","r":"COMPLEMENTS","w":2,"why":"Outcome-aware subgroup discovery vs unsupervised SQL clustering","m":"agent draft (claude)"},{"id":"e1482","i":"ren-2010--COMPLEMENTS-->gregg-2016","s":"ren-2010","t":"gregg-2016","r":"COMPLEMENTS","w":2,"why":"GWP gathers low-overhead fleet-wide samples continuously; flame graphs are the standard way to visualize such stack samples. Modern eBPF continuous profilers (Parca, Pyroscope) combine both.","m":"agent draft (claude)"},{"id":"e1486","i":"rizun-2021--CITES-->wood-1986","s":"rizun-2021","t":"wood-1986","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1108/bpmj-04-2021-0217 -> 10.1016/0749-5978(86)90044-0)","m":"reference-list extractor"},{"id":"e1490","i":"rocher2019reidentification--CITES-->sweeney2002kanonymity","s":"rocher2019reidentification","t":"sweeney2002kanonymity","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1038/s41467-019-10933-3 -> 10.1142/s0218488502001648)","m":"reference-list extractor"},{"id":"e1492","i":"rodrigues2021clp--EXTENDS-->lin2015cowic","s":"rodrigues2021clp","t":"lin2015cowic","r":"EXTENDS","w":2,"why":"CLP advances Cowic's column-wise log stream decomposition by separating static logtypes, deduplicated variable dictionaries (for compound tokens like UUIDs/paths), and raw numeric streams for direct compressed search.","m":"agent draft (gemini)"},{"id":"e1493","i":"rodrigues2022tracesim--CITES-->brodie2005symptom","s":"rodrigues2022tracesim","t":"brodie2005symptom","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s10664-021-10070-w -> 10.1109/icac.2005.49)","m":"reference-list extractor"},{"id":"e1494","i":"rodrigues2022tracesim--CITES-->dhaliwal2011crash","s":"rodrigues2022tracesim","t":"dhaliwal2011crash","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s10664-021-10070-w -> 10.1109/icsm.2011.6080800)","m":"reference-list extractor"},{"id":"e1495","i":"rodrigues2022tracesim--CITES-->glerum2009wer","s":"rodrigues2022tracesim","t":"glerum2009wer","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s10664-021-10070-w -> 10.1145/1629575.1629586)","m":"reference-list extractor"},{"id":"e1496","i":"rodrigues2022tracesim--CITES-->kim2011crashgraphs","s":"rodrigues2022tracesim","t":"kim2011crashgraphs","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s10664-021-10070-w -> 10.1109/dsn.2011.5958261)","m":"reference-list extractor"},{"id":"e1497","i":"rodrigues2022tracesim--EXTENDS-->dang2012rebucket","s":"rodrigues2022tracesim","t":"dang2012rebucket","r":"EXTENDS","w":2,"why":"Unifies ReBucket's position-dependent depth weighting with CrashLocator/Brodie's inverse document frequency frame specificity in a global alignment algorithm.","m":"agent draft (gemini)"},{"id":"e1498","i":"rodrigues2022tracesim--USES_METHOD-->wu2014crashlocator","s":"rodrigues2022tracesim","t":"wu2014crashlocator","r":"USES_METHOD","w":2,"why":"Incorporates inverse-document-frequency (IDF) frame specificity weights into stack trace alignment.","m":"agent draft (gemini)"},{"id":"e1501","i":"roth2023trending--CITES-->abadie2010sc","s":"roth2023trending","t":"abadie2010sc","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.jeconom.2023.03.008 -> 10.1198/jasa.2009.ap08746)","m":"reference-list extractor"},{"id":"e1503","i":"roth2023trending--CITES-->callaway2021did","s":"roth2023trending","t":"callaway2021did","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.jeconom.2023.03.008 -> 10.1016/j.jeconom.2020.12.001)","m":"reference-list extractor"},{"id":"e1504","i":"roth2023trending--CITES-->cameron2008cluster","s":"roth2023trending","t":"cameron2008cluster","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.jeconom.2023.03.008 -> 10.1162/rest.90.3.414)","m":"reference-list extractor"},{"id":"e1506","i":"roth2023trending--CITES-->goodmanbacon2021decomp","s":"roth2023trending","t":"goodmanbacon2021decomp","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.jeconom.2023.03.008 -> 10.1016/j.jeconom.2021.03.014)","m":"reference-list extractor"},{"id":"e1508","i":"roth2023trending--CITES-->sun2021event","s":"roth2023trending","t":"sun2021event","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.jeconom.2023.03.008 -> 10.1016/j.jeconom.2020.09.006)","m":"reference-list extractor"},{"id":"e1511","i":"rule2020auditlogreview--CITES-->arndt2017tethered","s":"rule2020auditlogreview","t":"arndt2017tethered","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1093/jamia/ocz196 -> 10.1370/afm.2121)","m":"reference-list extractor"},{"id":"e1513","i":"rust-1995--CITES-->cronin-1992","s":"rust-1995","t":"cronin-1992","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1177/002224299505900205 -> 10.1177/002224299205600304)","m":"reference-list extractor"},{"id":"e1514","i":"rust-1995--COMPLEMENTS-->anderson2004satisfaction","s":"rust-1995","t":"anderson2004satisfaction","r":"COMPLEMENTS","w":2,"why":"Quality-to-retention-to-revenue chain; satisfaction to shareholder value.","m":"agent draft (gemini)"},{"id":"e1516","i":"saadfalcon2024ares--EXTENDS-->es2024ragas","s":"saadfalcon2024ares","t":"es2024ragas","r":"EXTENDS","w":2,"why":"ARES targets the same RAG dimensions as RAGAS (context relevance, faithfulness, answer relevance) but replaces prompted judges with fine-tuned judges plus PPI-calibrated intervals; compares against RAGAS.","m":"agent draft (claude)"},{"id":"e1517","i":"saadfalcon2024ares--USES_METHOD-->angelopoulos2023ppi","s":"saadfalcon2024ares","t":"angelopoulos2023ppi","r":"USES_METHOD","w":2,"why":"ARES uses prediction-powered inference with a small human-annotated set to give confidence intervals on judge-scored RAG metrics.","m":"agent draft (claude)"},{"id":"e1518","i":"saerens2002prior--COMPLEMENTS-->tversky1974judgment","s":"saerens2002prior","t":"tversky1974judgment","r":"COMPLEMENTS","w":2,"why":"Saerens et al. (2002) provide the exact log-odds prior-shift correction that eliminates the base-rate neglect and numeric anchoring documented by Tversky & Kahneman (1974).","m":"agent draft (gemini)"},{"id":"e1519","i":"safford-2007--CITES-->charlson-1987","s":"safford-2007","t":"charlson-1987","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s11606-007-0307-0 -> 10.1016/0021-9681(87)90171-8)","m":"reference-list extractor"},{"id":"e1521","i":"saha-2022--COMPLEMENTS-->ding-2014","s":"saha-2022","t":"ding-2014","r":"COMPLEMENTS","w":2,"why":"both retrieve past incidents for a new one; ICA uses a knowledge graph, Ding uses signature+healing action","m":"agent draft (gemini)"},{"id":"e1522","i":"saisana-2005--OPERATIONALIZES-->author-2008","s":"saisana-2005","t":"author-2008","r":"COMPLEMENTS","w":2,"why":"Saisana et al. give the uncertainty and sensitivity analysis procedure that the OECD/JRC handbook prescribes as step 7.","m":"agent draft (claude)"},{"id":"e1526","i":"sambasivan2021cascades--CITES-->buneman-2001","s":"sambasivan2021cascades","t":"buneman-2001","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3411764.3445518 -> 10.1007/3-540-44503-x_20)","m":"reference-list extractor"},{"id":"e1529","i":"sambasivan2021cascades--COMPLEMENTS-->sculley2015hiddendebt","s":"sambasivan2021cascades","t":"sculley2015hiddendebt","r":"COMPLEMENTS","w":3,"why":"Sambasivan et al. (2021) document the organizational and upstream data-collection cascades (92% prevalence) that compound Sculley et al.'s (2015) ML system-level pipeline jungles and unstable data dependencies.","m":"agent draft (gemini)"},{"id":"e1533","i":"sanders2019leastprivabac--EXTENDS-->vaidya2007rolemining","s":"sanders2019leastprivabac","t":"vaidya2007rolemining","r":"EXTENDS","w":2,"why":"Least-privilege policy mining from audit logs","m":"agent draft (claude)"},{"id":"e1535","i":"sandve2013tenrules--OPERATIONALIZES-->iso-42001","s":"sandve2013tenrules","t":"iso-42001","r":"COMPLEMENTS","w":2,"why":"Reproducibility rules give testable criteria for 'methods that ensure valid results' (cl. 9.1) and controlled documented information (cl. 7.5) when notebooks are management evidence","m":"agent draft (gemini)"},{"id":"e1536","i":"sanyam-2024-arxiv--CONTRASTS_WITH-->tian2023justask","s":"sanyam-2024-arxiv","t":"tian2023justask","r":"CONTRASTS_WITH","w":2,"why":"Tian et al. find verbalized confidence can be better calibrated than RLHF logits; Kapoor et al. find prompting-only uncertainty is unreliable and requires fine-tuning on ~1k examples.","m":"agent draft (claude)"},{"id":"e1537","i":"satyanarayan-2017--OPERATIONALIZES-->shneiderman1996eyes","s":"satyanarayan-2017","t":"shneiderman1996eyes","r":"OPERATIONALIZES","w":2,"why":"Operationalizes Shneiderman's overview-first, zoom-and-filter, details-on-demand taxonomy as composable declarative selection predicates.","m":"agent draft (gemini)"},{"id":"e1546","i":"schad-2010--COMPLEMENTS-->hu2026cbstail","s":"schad-2010","t":"hu2026cbstail","r":"COMPLEMENTS","w":2,"why":"Guest-side variance measurement vs provider-side root cause of block-store tails.","m":"agent draft (claude)"},{"id":"e1547","i":"schelter2018deequ--CITES-->breck2017mltestscore","s":"schelter2018deequ","t":"breck2017mltestscore","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3229863.3229867 -> 10.1109/bigdata.2017.8258038)","m":"reference-list extractor"},{"id":"e1548","i":"schelter2018deequ--COMPLEMENTS-->breck2019datavalidation","s":"schelter2018deequ","t":"breck2019datavalidation","r":"COMPLEMENTS","w":2,"why":"Pairs declarative unit tests for large-scale dataset constraints (Deequ) with continuous schema and distribution-skew validation in ML pipelines.","m":"agent draft (gemini)"},{"id":"e1549","i":"schelter2018deequ--COMPLEMENTS-->caveness2020tfdv","s":"schelter2018deequ","t":"caveness2020tfdv","r":"COMPLEMENTS","w":2,"why":"Two production data-unit-test systems: declarative constraints vs inferred schema","m":"agent draft (gemini)"},{"id":"e1550","i":"schelter2018deequ--COMPLEMENTS-->shankar2023brokendata","s":"schelter2018deequ","t":"shankar2023brokendata","r":"COMPLEMENTS","w":2,"why":"declarative DQ tests vs validation under continuous ingest","m":"agent draft (gemini)"},{"id":"e1551","i":"schelter2018deequ--OPERATIONALIZES-->batini2009methodologies","s":"schelter2018deequ","t":"batini2009methodologies","r":"OPERATIONALIZES","w":2,"why":"Deequ turns survey-level DQ dimensions (completeness, uniqueness, consistency) into executable constraints and metric histories.","m":"agent draft (claude)"},{"id":"e1553","i":"schipper-2019--CITES-->beschastnikh2011synoptic","s":"schipper-2019","t":"beschastnikh2011synoptic","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/msr.2019.00081 -> 10.1145/2025113.2025151)","m":"reference-list extractor"},{"id":"e1554","i":"schipper-2019--CITES-->du2016spell","s":"schipper-2019","t":"du2016spell","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/msr.2019.00081 -> 10.1109/icdm.2016.0103)","m":"reference-list extractor"},{"id":"e1555","i":"schipper-2019--CITES-->fu2009execution","s":"schipper-2019","t":"fu2009execution","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/msr.2019.00081 -> 10.1109/icdm.2009.60)","m":"reference-list extractor"},{"id":"e1556","i":"schipper-2019--CITES-->he2017drain","s":"schipper-2019","t":"he2017drain","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/msr.2019.00081 -> 10.1109/icws.2017.13)","m":"reference-list extractor"},{"id":"e1557","i":"schipper-2019--CITES-->nagaraj2012distalyzer","s":"schipper-2019","t":"nagaraj2012distalyzer","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/msr.2019.00081 -> nagaraj2012distalyzer)","m":"reference-list extractor"},{"id":"e1558","i":"schipper-2019--COMPLEMENTS-->yuan2012characterizing","s":"schipper-2019","t":"yuan2012characterizing","r":"COMPLEMENTS","w":2,"why":"Validates how developer logging idioms (wrappers, string concatenation, custom formatters) identified by Yuan et al. affect static AST log template extractors in production.","m":"agent draft (gemini)"},{"id":"e1560","i":"schloss-2019--CITES-->gramazio-2017","s":"schloss-2019","t":"gramazio-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2018.2865147 -> 10.1109/tvcg.2016.2598918)","m":"reference-list extractor"},{"id":"e1561","i":"schloss-2019--CITES-->szafir-2018","s":"schloss-2019","t":"szafir-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2018.2865147 -> 10.1109/tvcg.2017.2744359)","m":"reference-list extractor"},{"id":"e1562","i":"schloss-2019--COMPLEMENTS-->liu-2018","s":"schloss-2019","t":"liu-2018","r":"COMPLEMENTS","w":2,"why":"direction of colour-quantity mapping vs background for colormaps","m":"agent draft (claude)"},{"id":"e1564","i":"schroeder-2006--COMPLEMENTS-->oliner2007supercomputers","s":"schroeder-2006","t":"oliner2007supercomputers","r":"COMPLEMENTS","w":2,"why":"Both mine HPC field logs; Schroeder classifies root causes, Oliner exposes log pitfalls","m":"agent draft (gemini)"},{"id":"e1565","i":"schroeder-2006--EXTENDS-->oliner2007supercomputers","s":"oliner2007supercomputers","t":"schroeder-2006","r":"EXTENDS","w":2,"why":"Both mine HPC field logs; Schroeder classifies root causes (env/hw/sw), Oliner shows log pitfalls (timestamps, redundancy)","m":"agent draft (gemini)"},{"id":"e1570","i":"sculley2015hiddendebt--COMPLEMENTS-->breck2017mltestscore","s":"sculley2015hiddendebt","t":"breck2017mltestscore","r":"COMPLEMENTS","w":2,"why":"debt taxonomy vs readiness tests","m":"agent draft (gemini)"},{"id":"e1571","i":"sculley2015hiddendebt--COMPLEMENTS-->iso-42001","s":"sculley2015hiddendebt","t":"iso-42001","r":"COMPLEMENTS","w":2,"why":"CACE entanglement and unstable data dependencies explain why ISO/IEC 42001 Clause 6.3/8.2 change management and A.6.2.4 V&V must gate prompts, schemas, and data together.","m":"agent draft (gemini)"},{"id":"e1572","i":"segel2010narrative--COMPLEMENTS-->kosara-2013","s":"segel2010narrative","t":"kosara-2013","r":"COMPLEMENTS","w":2,"why":"narrative genres/structures for presentation vs storytelling agenda","m":"agent draft (claude)"},{"id":"e1573","i":"segura2016metamorphicsurvey--CITES-->jia2011mutation","s":"segura2016metamorphicsurvey","t":"jia2011mutation","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2016.2532875 -> 10.1109/tse.2010.62)","m":"reference-list extractor"},{"id":"e1574","i":"segura2016metamorphicsurvey--EXTENDS-->chen1998metamorphic","s":"segura2016metamorphicsurvey","t":"chen1998metamorphic","r":"EXTENDS","w":2,"why":"Survey of the MT technique Chen et al. proposed","m":"agent draft (gemini)"},{"id":"e1576","i":"shah2026logjack--CONTRASTS_WITH-->jiahao-2023-arxiv","s":"shah2026logjack","t":"jiahao-2023-arxiv","r":"CONTRASTS_WITH","w":2,"why":"Data-plane injection via log payloads vs prompt-plane adversarial fuzzing: two attack surfaces of LLM tool stacks","m":"agent draft (gemini)"},{"id":"e1577","i":"shankar2023brokendata--EXTENDS-->schelter2018deequ","s":"shankar2023brokendata","t":"schelter2018deequ","r":"EXTENDS","w":2,"why":"GATE addresses the false-positive problem of static Deequ/TFDV-style constraints on continuously ingested partitions.","m":"agent draft (claude)"},{"id":"e1579","i":"shankar2024validators--EXTENDS-->zheng2023judging","s":"shankar2024validators","t":"zheng2023judging","r":"EXTENDS","w":2,"why":"Zheng et al. establish LLM judges agree with humans on aggregate; Shankar et al. show the judge criteria themselves must be aligned to human grades and drift during grading.","m":"agent draft (claude)"},{"id":"e1581","i":"shao2020dbaccessantipatterns--CITES-->chen2014ormantipatterns","s":"shao2020dbaccessantipatterns","t":"chen2014ormantipatterns","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icsme46990.2020.00016 -> 10.1145/2568225.2568259)","m":"reference-list extractor"},{"id":"e1583","i":"shao2020dbaccessantipatterns--CITES-->chen2016redundantdataaccess","s":"shao2020dbaccessantipatterns","t":"chen2016redundantdataaccess","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icsme46990.2020.00016 -> 10.1109/tse.2016.2553039)","m":"reference-list extractor"},{"id":"e1584","i":"shao2020dbaccessantipatterns--CITES-->muse2020sqlsmells","s":"shao2020dbaccessantipatterns","t":"muse2020sqlsmells","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icsme46990.2020.00016 -> muse2020sqlsmells)","m":"reference-list extractor"},{"id":"e1585","i":"shao2020dbaccessantipatterns--CITES-->qiu2013schemacoevolution","s":"shao2020dbaccessantipatterns","t":"qiu2013schemacoevolution","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icsme46990.2020.00016 -> qiu2013schemacoevolution)","m":"reference-list extractor"},{"id":"e1586","i":"shao2020dbaccessantipatterns--CITES-->yang2018webappperf","s":"shao2020dbaccessantipatterns","t":"yang2018webappperf","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icsme46990.2020.00016 -> 10.1145/3180155.3180194)","m":"reference-list extractor"},{"id":"e1587","i":"shao2020dbaccessantipatterns--EXTENDS-->dintyala2020sqlcheck","s":"shao2020dbaccessantipatterns","t":"dintyala2020sqlcheck","r":"EXTENDS","w":2,"why":"broader anti-pattern taxonomy from issue trackers vs SQLCheck's static detector","m":"agent draft (gemini)"},{"id":"e1589","i":"shen-2023--CITES-->corbett2013spanner","s":"shen-2023","t":"corbett2012spanner","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3575693.3575727 -> 10.1145/2491245)","m":"reference-list extractor"},{"id":"e1590","i":"shen-2023--CITES-->kanev-2015","s":"shen-2023","t":"kanev-2015","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3575693.3575727 -> 10.1145/2749469.2750392)","m":"reference-list extractor"},{"id":"e1591","i":"shen-2023--CITES-->yasin-2014","s":"shen-2023","t":"yasin-2014","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3575693.3575727 -> 10.1109/ispass.2014.6844459)","m":"reference-list extractor"},{"id":"e1592","i":"shen-2023--EXTENDS-->panchenko-2019","s":"shen-2023","t":"panchenko-2019","r":"EXTENDS","w":2,"why":"Propeller moves BOLT-style post-link layout optimization into a relinking flow for warehouse-scale binaries","m":"agent draft (gemini)"},{"id":"e1598","i":"shippee-2012--COMPLEMENTS-->safford-2007","s":"shippee-2012","t":"safford-2007","r":"COMPLEMENTS","w":2,"why":"Two patient-complexity frameworks: demand/capacity balance vs multi-axis vector; both separate complexity from care volume.","m":"agent draft (claude-sonnet-5-5)"},{"id":"e1599","i":"siddiqui2022isum--CITES-->deep2020workloadcompression","s":"siddiqui2022isum","t":"deep2020workloadcompression","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3514221.3526152 -> 10.14778/3430915.3430931)","m":"reference-list extractor"},{"id":"e1600","i":"siddiqui2022isum--EXTENDS-->chaudhuri2002compressing","s":"siddiqui2022isum","t":"chaudhuri2002compressing","r":"EXTENDS","w":2,"why":"Workload compression for scalable index tuning","m":"agent draft (claude)"},{"id":"e1601","i":"siegel-2022--COMPLEMENTS-->ravid-2019","s":"siegel-2022","t":"ravid-2019","r":"COMPLEMENTS","w":2,"why":"Meta-analytic estimate of EPM effects complements Ravid et al.'s EPM framework (purpose, transparency, invasiveness).","m":"agent draft (gemini)"},{"id":"e1603","i":"sigelman2010dapper--COMPLEMENTS-->ren-2010","s":"sigelman2010dapper","t":"ren-2010","r":"COMPLEMENTS","w":2,"why":"Both Google always-on low-overhead sampling infrastructures: Dapper samples request traces, GWP samples machine/process profiles; together they attribute latency vs CPU.","m":"agent draft (gemini)"},{"id":"e1605","i":"silberzahn2018manyanalysts--CITES-->simmons2011false","s":"silberzahn2018manyanalysts","t":"simmons2011false","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1177/2515245917747646 -> 10.1177/0956797611417632)","m":"reference-list extractor"},{"id":"e1606","i":"sim-2005--CITES-->cohen1960kappa","s":"sim-2005","t":"cohen1960kappa","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1093/ptj/85.3.257 -> 10.1177/001316446002000104)","m":"reference-list extractor"},{"id":"e1608","i":"sim-2005--CITES-->landis-1977","s":"sim-2005","t":"landis-1977","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1093/ptj/85.3.257 -> 10.2307/2529310)","m":"reference-list extractor"},{"id":"e1609","i":"sim-2005--COMPLEMENTS-->feinstein1990kappa","s":"sim-2005","t":"feinstein1990kappa","r":"COMPLEMENTS","w":2,"why":"Both stress kappa's dependence on prevalence. Sim & Wright add sample-size guidance for estimating kappa to a given precision.","m":"agent draft (claude)"},{"id":"e1610","i":"simonsohn-2020--CITES-->simmons2011false","s":"simonsohn-2020","t":"simmons2011false","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1038/s41562-020-0912-z -> 10.1177/0956797611417632)","m":"reference-list extractor"},{"id":"e1611","i":"simonsohn-2020--COMPLEMENTS-->steegen-2016","s":"simonsohn-2020","t":"steegen-2016","r":"COMPLEMENTS","w":2,"why":"Specification curve and multiverse: two framings of analyst-choice sensitivity.","m":"agent draft (gemini)"},{"id":"e1612","i":"simonsohn-2020--EXTENDS-->steegen-2016","s":"simonsohn-2020","t":"steegen-2016","r":"EXTENDS","w":2,"why":"adds joint inference over the multiverse","m":"agent draft (gemini)"},{"id":"e1615","i":"sirin-2016--CITES-->neumann-2011","s":"sirin-2016","t":"neumann-2011","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2882903.2882916 -> 10.14778/2002938.2002940)","m":"reference-list extractor"},{"id":"e1616","i":"sirin-2016--USES_METHOD-->yasin-2014","s":"sirin-2016","t":"yasin-2014","r":"USES_METHOD","w":2,"why":"Applies Intel Top-Down Microarchitecture Analysis Method (TMAM) to in-memory and disk OLTP engines","m":"agent draft (gemini)"},{"id":"e1617","i":"sirin-2017--CITES-->sirin-2016","s":"sirin-2017","t":"sirin-2016","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3076113.3076116 -> 10.1145/2882903.2882916)","m":"reference-list extractor"},{"id":"e1618","i":"sirin-2017--EXTENDS-->yasin-2014","s":"sirin-2017","t":"yasin-2014","r":"EXTENDS","w":2,"why":"Corrects TMAM's underestimation of memory stalls on pointer-chasing OLTP workloads","m":"agent draft (gemini)"},{"id":"e1625","i":"sirin-2021--CITES-->chen-2016","s":"sirin-2021","t":"chen-2016","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s00778-021-00663-8 -> 10.1145/2854038.2854044)","m":"reference-list extractor"},{"id":"e1627","i":"sirin-2021--CITES-->neumann-2011","s":"sirin-2021","t":"neumann-2011","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s00778-021-00663-8 -> 10.14778/2002938.2002940)","m":"reference-list extractor"},{"id":"e1628","i":"sirin-2021--CITES-->yasin-2014","s":"sirin-2021","t":"yasin-2014","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s00778-021-00663-8 -> 10.1109/ispass.2014.6844459)","m":"reference-list extractor"},{"id":"e1629","i":"sirin-2021--EXTENDS-->sirin-2016","s":"sirin-2021","t":"sirin-2016","r":"EXTENDS","w":2,"why":"Revisits OLTP microarchitectural stalls across multi-socket architectures and memory bandwidth ceilings","m":"agent draft (gemini)"},{"id":"e1631","i":"skau-2016--CITES-->correll-2014","s":"skau-2016","t":"correll-2014","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/cgf.12888 -> 10.1109/tvcg.2014.2346298)","m":"reference-list extractor"},{"id":"e1632","i":"skau-2016--CITES-->heer2010crowdsourcing","s":"skau-2016","t":"heer2010crowdsourcing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/cgf.12888 -> 10.1145/1753326.1753357)","m":"reference-list extractor"},{"id":"e1633","i":"skau-2016--CITES-->munzner2014vad","s":"skau-2016","t":"munzner2014vad","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/cgf.12888 -> 10.1201/b17511)","m":"reference-list extractor"},{"id":"e1634","i":"skau-2016--CONTRASTS_WITH-->cleveland1984graphical","s":"skau-2016","t":"cleveland1984graphical","r":"CONTRASTS_WITH","w":2,"why":"angle is least used cue in pie/donut; donut as accurate as pie","m":"agent draft (claude)"},{"id":"e1635","i":"skoulis2015growingup--CITES-->curino2008prism","s":"skoulis2015growingup","t":"curino2008prism","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.is.2015.03.009 -> 10.14778/1453856.1453939)","m":"reference-list extractor"},{"id":"e1636","i":"skoulis2015growingup--CITES-->qiu2013schemacoevolution","s":"skoulis2015growingup","t":"qiu2013schemacoevolution","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.is.2015.03.009 -> 10.1145/2491411.2491431)","m":"reference-list extractor"},{"id":"e1637","i":"skoulis2015growingup--EXTENDS-->curino2008wikipedia","s":"skoulis2015growingup","t":"curino2008wikipedia","r":"EXTENDS","w":2,"why":"Extends Wikipedia schema-evolution study to multiple open-source DB histories","m":"agent draft (claude)"},{"id":"e1638","i":"slagell2005sharing--CITES-->xu2002prefix","s":"slagell2005sharing","t":"xu2002prefix","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/seccmw.2005.1588299 -> xu2002prefix)","m":"reference-list extractor"},{"id":"e1639","i":"slagell2005sharing--COMPLEMENTS-->pang2006devil","s":"slagell2005sharing","t":"pang2006devil","r":"COMPLEMENTS","w":2,"why":"Establishes multi-level log anonymization accounting and motivates reconciling rule-level vs artifact-level scrubbing with fail-closed verification.","m":"agent draft (gemini)"},{"id":"e1644","i":"sors2021silence--EXTENDS-->dohmen2008social","s":"sors2021silence","t":"dohmen2008social","r":"EXTENDS","w":3,"why":"Uses behind-closed-doors COVID-19 matches across European top-flight leagues as a natural experiment, proving that crowd absence eliminates referee foul and card asymmetry against away teams.","m":"agent draft (gemini)"},{"id":"e1646","i":"souihli2021bullfrog--CITES-->dejong2017zerodowntime","s":"souihli2021bullfrog","t":"dejong2017zerodowntime","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3448016.3452842 -> 10.1109/icse-seip.2017.5)","m":"reference-list extractor"},{"id":"e1647","i":"souihli2021bullfrog--CITES-->qiu2013schemacoevolution","s":"souihli2021bullfrog","t":"qiu2013schemacoevolution","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3448016.3452842 -> 10.1145/2491411.2491431)","m":"reference-list extractor"},{"id":"e1648","i":"souihli2021bullfrog--COMPLEMENTS-->rae2013f1","s":"souihli2021bullfrog","t":"rae2013f1","r":"COMPLEMENTS","w":2,"why":"Lazy online schema evolution vs F1 asynchronous multi-state schema change","m":"agent draft (claude)"},{"id":"e1649","i":"sovrano2026mitigating--COMPLEMENTS-->chaparro2017missing","s":"sovrano2026mitigating","t":"chaparro2017missing","r":"COMPLEMENTS","w":2,"why":"Sovrano et al. (FSE 2026) shows how to elicit missing software engineering context without inducing prompt-driven confirmation or framing bias.","m":"agent draft (gemini)"},{"id":"e1651","i":"spiegelhalter2005funnel--COMPLEMENTS-->bevan2006targets","s":"spiegelhalter2005funnel","t":"bevan2006targets","r":"COMPLEMENTS","w":2,"why":"Funnel plots counter the misreading of small-n league tables, one of the target-regime pathologies Bevan & Hood describe.","m":"agent draft (claude)"},{"id":"e1652","i":"spiegelhalter2005funnel--COMPLEMENTS-->franconeri2021science","s":"spiegelhalter2005funnel","t":"franconeri2021science","r":"COMPLEMENTS","w":2,"why":"funnel plots as the honest alternative to ranked league tables","m":"agent draft (gemini)"},{"id":"e1653","i":"spiegelhalter2005funnel--COMPLEMENTS-->gelman-1999","s":"spiegelhalter2005funnel","t":"gelman-1999","r":"COMPLEMENTS","w":2,"why":"Both address small-n unit rates: funnel limits (frequentist) vs partial pooling (Bayesian) for per-customer SLA tables.","m":"agent draft (claude)"},{"id":"e1654","i":"spiegelhalter2005overdispersion--EXTENDS-->spiegelhalter2005funnel","s":"spiegelhalter2005overdispersion","t":"spiegelhalter2005funnel","r":"EXTENDS","w":2,"why":"adds over-dispersion (multiplicative phi or random effects) to binomial funnel limits","m":"agent draft (claude)"},{"id":"e1655","i":"spiegelhalter2012regulation--CITES-->gelman-1999","s":"spiegelhalter2012regulation","t":"gelman-1999","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1467-985x.2011.01010.x -> 10.1002/(sici)1097-0258(19991215)18:23<3221::aid-sim312>3.0.co;2-m)","m":"reference-list extractor"},{"id":"e1656","i":"spiegelhalter2012regulation--CITES-->spiegelhalter2005overdispersion","s":"spiegelhalter2012regulation","t":"spiegelhalter2005overdispersion","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1467-985x.2011.01010.x -> 10.1136/qshc.2005.013755)","m":"reference-list extractor"},{"id":"e1657","i":"spiegelhalter2012regulation--CITES-->woodall-2006","s":"spiegelhalter2012regulation","t":"woodall-2006","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1467-985x.2011.01010.x -> 10.1080/00224065.2006.11918593)","m":"reference-list extractor"},{"id":"e1658","i":"spiegelhalter2012regulation--EXTENDS-->spiegelhalter2005funnel","s":"spiegelhalter2012regulation","t":"spiegelhalter2005funnel","r":"EXTENDS","w":2,"why":"screening many units: funnel plots combined with FDR and over-dispersion adjustment","m":"agent draft (claude)"},{"id":"e1659","i":"spiegelhalter2012regulation--USES_METHOD-->benjamini1995fdr","s":"spiegelhalter2012regulation","t":"benjamini1995fdr","r":"USES_METHOD","w":2,"why":"FDR control when screening many providers","m":"agent draft (claude)"},{"id":"e1661","i":"srinivasan-2023--CITES-->jung-2022","s":"srinivasan-2023","t":"jung-2022","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3597638.3608405 -> 10.1109/tvcg.2021.3114846)","m":"reference-list extractor"},{"id":"e1662","i":"srinivasan-2023--CITES-->lundgard-2022","s":"srinivasan-2023","t":"lundgard-2022","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3597638.3608405 -> lundgard-2022)","m":"reference-list extractor"},{"id":"e1663","i":"srinivasan-2023--CITES-->satyanarayan-2017","s":"srinivasan-2023","t":"satyanarayan-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3597638.3608405 -> satyanarayan-2017)","m":"reference-list extractor"},{"id":"e1664","i":"srinivasan-2023--CITES-->shneiderman1996eyes","s":"srinivasan-2023","t":"shneiderman1996eyes","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3597638.3608405 -> 10.1109/vl.1996.545307)","m":"reference-list extractor"},{"id":"e1665","i":"srinivasan-2023--CITES-->tory2023datavoice","s":"srinivasan-2023","t":"tory2023datavoice","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3597638.3608405 -> tory2023datavoice)","m":"reference-list extractor"},{"id":"e1666","i":"srinivasan-2023--OPERATIONALIZES-->w3c2024wcag22","s":"srinivasan-2023","t":"w3c2024wcag22","r":"OPERATIONALIZES","w":2,"why":"dashboard-level screen-reader patterns that implement WCAG 2.2 for multi-chart pages","m":"agent draft (gemini)"},{"id":"e1667","i":"starmer-2014--COMPLEMENTS-->carmel2010followsun","s":"starmer-2014","t":"carmel2010followsun","r":"COMPLEMENTS","w":2,"why":"Clinical structured handoff evidence complements FTS handoff-cost model for support shift handovers","m":"agent draft (gemini)"},{"id":"e1668","i":"starmer-2014--COMPLEMENTS-->kcs2016v6","s":"starmer-2014","t":"kcs2016v6","r":"COMPLEMENTS","w":2,"why":"Structured handoff (I-PASS) and KCS solve-loop both treat structured, reusable case knowledge as the unit of quality","m":"agent draft (gemini)"},{"id":"e1670","i":"steegen-2016--EXTENDS-->simmons2011false","s":"steegen-2016","t":"simmons2011false","r":"EXTENDS","w":2,"why":"multiverse as a systematic answer to researcher degrees of freedom","m":"agent draft (gemini)"},{"id":"e1671","i":"steegen-2016--OPERATIONALIZES-->silberzahn2018manyanalysts","s":"silberzahn2018manyanalysts","t":"steegen-2016","r":"OPERATIONALIZES","w":2,"why":"Many-analysts variation operationalised as a single-team multiverse.","m":"agent draft (gemini)"},{"id":"e1674","i":"stokes-2022--CITES-->kong2018titles","s":"stokes-2022","t":"kong2018titles","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2022.3209383 -> 10.1145/3173574.3174012)","m":"reference-list extractor"},{"id":"e1675","i":"stokes-2022--COMPLEMENTS-->kim2021captions","s":"stokes-2022","t":"kim2021captions","r":"COMPLEMENTS","w":2,"why":"text-chart integration and reader takeaways","m":"agent draft (claude)"},{"id":"e1677","i":"sun2021event--COMPLEMENTS-->callaway2021did","s":"sun2021event","t":"callaway2021did","r":"COMPLEMENTS","w":2,"why":"Interaction-weighted event study; alternative heterogeneity-robust estimator for staggered adoption.","m":"agent draft (gemini)"},{"id":"e1678","i":"swaminathan-rogers-1990--OPERATIONALIZES-->meredith-1993","s":"meredith-1993","t":"swaminathan-rogers-1990","r":"OPERATIONALIZES","w":2,"why":"Logistic-regression DIF is a practical item-level test of the invariance property Meredith formalises.","m":"agent draft (claude-sonnet-5-5)"},{"id":"e1679","i":"szafir-2018--CITES-->satyanarayan-2017","s":"szafir-2018","t":"satyanarayan-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tvcg.2017.2744359 -> 10.1109/tvcg.2016.2599030)","m":"reference-list extractor"},{"id":"e1681","i":"szlang2025snowflakeworkload--CITES-->jain2016sqlshare","s":"szlang2025snowflakeworkload","t":"jain2016sqlshare","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3750601.3750632 -> 10.1145/2882903.2882957)","m":"reference-list extractor"},{"id":"e1682","i":"szlang2025snowflakeworkload--CITES-->ma2018queryforecast","s":"szlang2025snowflakeworkload","t":"ma2018queryforecast","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3750601.3750632 -> 10.1145/3183713.3196908)","m":"reference-list extractor"},{"id":"e1683","i":"szlang2025snowflakeworkload--CITES-->melnik-2010","s":"szlang2025snowflakeworkload","t":"melnik-2010","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3750601.3750632 -> 10.14778/1920841.1920886)","m":"reference-list extractor"},{"id":"e1684","i":"szlang2025snowflakeworkload--COMPLEMENTS-->vanrenen2024redshiftfleet","s":"szlang2025snowflakeworkload","t":"vanrenen2024redshiftfleet","r":"COMPLEMENTS","w":2,"why":"Snowflake SQL-text fleet analysis complements Redshift fleet data-centric analysis","m":"agent draft (claude)"},{"id":"e1686","i":"taft2020cockroachdb--CITES-->rae2013f1","s":"taft2020cockroachdb","t":"rae2013f1","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3318464.3386134 -> rae2013f1)","m":"reference-list extractor"},{"id":"e1687","i":"taft2020cockroachdb--CITES-->thomson-2012","s":"taft2020cockroachdb","t":"thomson-2012","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3318464.3386134 -> 10.1145/2213836.2213838)","m":"reference-list extractor"},{"id":"e1689","i":"taft2020cockroachdb--COMPLEMENTS-->corbett2012spanner","s":"taft2020cockroachdb","t":"corbett2012spanner","r":"COMPLEMENTS","w":2,"why":"CockroachDB adapts Spanner design with HLC instead of TrueTime.","m":"agent draft (gemini)"},{"id":"e1690","i":"taft2020cockroachdb--EXTENDS-->corbett2012spanner","s":"taft2020cockroachdb","t":"corbett2012spanner","r":"EXTENDS","w":2,"why":"Adapts Spanner's distributed transactional architecture to commodity clocks and Raft/LSM (RocksDB/Pebble)","m":"agent draft (gemini)"},{"id":"e1691","i":"taipalus2023sqlerrorframework--EXTENDS-->becker2019compilererrors","s":"taipalus2023sqlerrorframework","t":"becker2019compilererrors","r":"EXTENDS","w":2,"why":"Carries general error-message guideline research into the SQL/DBMS domain.","m":"agent draft (gemini)"},{"id":"e1692","i":"taipalus2023sqlerrorframework--EXTENDS-->taipalus2021rdbmserrors","s":"taipalus2023sqlerrorframework","t":"taipalus2021rdbmserrors","r":"EXTENDS","w":2,"why":"From comparing four DBMS message sets to deriving design guidelines.","m":"agent draft (gemini)"},{"id":"e1693","i":"taipalus2025enhancedsql--CITES-->taipalus2021rdbmserrors","s":"taipalus2025enhancedsql","t":"taipalus2021rdbmserrors","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s10664-025-10695-1 -> 10.1016/j.jss.2021.111034)","m":"reference-list extractor"},{"id":"e1694","i":"taipalus2025enhancedsql--OPERATIONALIZES-->taipalus2023sqlerrorframework","s":"taipalus2025enhancedsql","t":"taipalus2023sqlerrorframework","r":"OPERATIONALIZES","w":2,"why":"Experimental test of the nine-guideline framework: enhanced messages faster to fix.","m":"agent draft (gemini)"},{"id":"e1698","i":"tang2015holistic--CITES-->xu2015csur","s":"tang2015holistic","t":"xu2015csur","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2815400.2815401 -> 10.1145/2791577)","m":"reference-list extractor"},{"id":"e1700","i":"tang2015holistic--OPERATIONALIZES-->yin2011configerrors","s":"tang2015holistic","t":"yin2011configerrors","r":"OPERATIONALIZES","w":2,"why":"Config-as-code validation and canaried rollout as the engineering answer to configuration errors","m":"agent draft (gemini)"},{"id":"e1701","i":"tang2024minicheck--COMPLEMENTS-->angelopoulos2022ltt","s":"tang2024minicheck","t":"angelopoulos2022ltt","r":"COMPLEMENTS","w":3,"why":"Pairs sentence-level fact-checking entailment verification (MiniCheck) with distribution-free risk-controlling abstention gates (Learn Then Test).","m":"agent draft (gemini)"},{"id":"e1702","i":"tang2025cola--CITES-->le2022logad","s":"tang2025cola","t":"le2022logad","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3749646.3749668 -> 10.1145/3510003.3510155)","m":"reference-list extractor"},{"id":"e1703","i":"tang2025cola--CITES-->wei2023loggrep","s":"tang2025cola","t":"wei2023loggrep","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3749646.3749668 -> 10.1145/3552326.3567484)","m":"reference-list extractor"},{"id":"e1704","i":"tang2025cola--COMPLEMENTS-->aggarwal2024automix","s":"tang2025cola","t":"aggarwal2024automix","r":"COMPLEMENTS","w":2,"why":"Small-local-then-large-cloud escalation for log anomaly detection mirrors AutoMix's verify-then-escalate cascade","m":"agent draft (qwen)"},{"id":"e1707","i":"tawosi-2023--CITES-->fu-2023","s":"tawosi-2023","t":"fu-2023","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2022.3228739 -> 10.1109/tse.2022.3158252)","m":"reference-list extractor"},{"id":"e1708","i":"tawosi-2023--CITES-->shepperd-2012","s":"tawosi-2023","t":"shepperd-2012","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tse.2022.3228739 -> 10.1016/j.infsof.2011.12.008)","m":"reference-list extractor"},{"id":"e1709","i":"tawosi-2023--REFUTES-->choetkiertikul-2019","s":"tawosi-2023","t":"choetkiertikul-2019","r":"REFUTES","w":2,"why":"Replication finds Deep-SE story-point estimates do not reliably beat naive baselines.","m":"agent draft (claude)","tr":1},{"id":"e1710","i":"tene2015latency--COMPLEMENTS-->dean-2013","s":"tene2015latency","t":"dean-2013","r":"COMPLEMENTS","w":2,"why":"measurement pitfalls (coordinated omission, percentile averaging) vs causes of tail latency","m":"agent draft (gemini)"},{"id":"e1711","i":"tezcan2014chat--CITES-->aksin2007moderncallcenter","s":"tezcan2014chat","t":"aksin2007moderncallcenter","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1287/opre.2014.1284 -> 10.1111/j.1937-5956.2007.tb00288.x)","m":"reference-list extractor"},{"id":"e1712","i":"tezcan2014chat--CITES-->gans-2003","s":"tezcan2014chat","t":"gans-2003","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1287/opre.2014.1284 -> 10.1287/msom.5.2.79.16071)","m":"reference-list extractor"},{"id":"e1713","i":"tezcan2014chat--EXTENDS-->garnett2002erlanga","s":"tezcan2014chat","t":"garnett2002erlanga","r":"EXTENDS","w":2,"why":"Concurrent-session staffing with impatient customers","m":"agent draft (gemini)"},{"id":"e1715","i":"thomson-2012--CITES-->brewer-2000","s":"thomson-2012","t":"brewer-2000","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2213836.2213838 -> 10.1145/564585.564601)","m":"reference-list extractor"},{"id":"e1716","i":"thomson-2012--CONTRASTS_WITH-->corbett2012spanner","s":"thomson-2012","t":"corbett2012spanner","r":"CONTRASTS_WITH","w":2,"why":"Deterministic pre-ordering vs 2PC+TrueTime.","m":"agent draft (gemini)"},{"id":"e1719","i":"tian-2016--CITES-->zhang-2013","s":"tian-2016","t":"zhang-2013","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s10664-015-9409-1 -> 10.1109/icse.2013.6606654)","m":"reference-list extractor"},{"id":"e1724","i":"tikv2026slowstore--OPERATIONALIZES-->panda2019iaso","s":"tikv2026slowstore","t":"panda2019iaso","r":"OPERATIONALIZES","w":2,"why":"Productized score-with-decay slow-node detection and single-node quarantine (leader eviction).","m":"agent draft (claude)"},{"id":"e1728","i":"tory2023datavoice--COMPLEMENTS-->sarikaya2019dashboards","s":"tory2023datavoice","t":"sarikaya2019dashboards","r":"COMPLEMENTS","w":2,"why":"Empirical dashboard-user practices (exporting, adding narrative) add to Sarikaya's genre typology.","m":"agent draft (claude)"},{"id":"e1730","i":"traini-2022--CITES-->laaber2019cloud","s":"traini-2022","t":"laaber2019cloud","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s10664-022-10247-x -> 10.1007/s10664-019-09681-1)","m":"reference-list extractor"},{"id":"e1731","i":"traini-2022--CITES-->mytkowicz2009wrong","s":"traini-2022","t":"mytkowicz2009wrong","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1007/s10664-022-10247-x -> 10.1145/1508244.1508275)","m":"reference-list extractor"},{"id":"e1732","i":"traini-2022--EXTENDS-->georges2007statistically","s":"traini-2022","t":"georges2007statistically","r":"EXTENDS","w":2,"why":"Empirically shows steady-state assumptions behind warmup heuristics often fail","m":"agent draft (gemini)"},{"id":"e1733","i":"treynor-2017--CITES-->beyer2016sre","s":"treynor-2017","t":"beyer2016sre","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3080202 -> beyer2016sre)","m":"reference-list extractor"},{"id":"e1734","i":"treynor-2017--COMPLEMENTS-->beyer2018workbook","s":"treynor-2017","t":"beyer2018workbook","r":"COMPLEMENTS","w":2,"why":"Dependency-aware availability budgeting complements SLO practice","m":"agent draft (gemini)"},{"id":"e1735","i":"turnbull1976censored--CITES-->kaplan-1958","s":"turnbull1976censored","t":"kaplan-1958","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.2517-6161.1976.tb01597.x -> 10.1080/01621459.1958.10501452)","m":"reference-list extractor"},{"id":"e1737","i":"turner-2021--CITES-->cohen1960kappa","s":"turner-2021","t":"cohen1960kappa","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1186/s12874-021-01306-w -> 10.1177/001316446002000104)","m":"reference-list extractor"},{"id":"e1738","i":"turner-2021--CITES-->lopezbernal2017its","s":"turner-2021","t":"lopezbernal2017its","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1186/s12874-021-01306-w -> 10.1093/ije/dyw098)","m":"reference-list extractor"},{"id":"e1739","i":"turner-2021--CITES-->newey1987hac","s":"turner-2021","t":"newey1987hac","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1186/s12874-021-01306-w -> 10.2307/1913610)","m":"reference-list extractor"},{"id":"e1741","i":"turner-2021--COMPLEMENTS-->wagner2002segmented","s":"turner-2021","t":"wagner2002segmented","r":"COMPLEMENTS","w":2,"why":"Turner et al. empirically compare ITS estimators and show that the choice of method changes the conclusions of the segmented regression that Wagner et al. popularised.","m":"agent draft (claude)"},{"id":"e1742","i":"unkel-2011--CITES-->benjamini1995fdr","s":"unkel-2011","t":"benjamini1995fdr","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1467-985x.2011.00714.x -> 10.1111/j.2517-6161.1995.tb02031.x)","m":"reference-list extractor"},{"id":"e1743","i":"unkel-2011--CITES-->farrington-1996","s":"unkel-2011","t":"farrington-1996","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1467-985x.2011.00714.x -> 10.2307/2983331)","m":"reference-list extractor"},{"id":"e1744","i":"unkel-2011--CITES-->page-1954","s":"unkel-2011","t":"page-1954","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1467-985x.2011.00714.x -> 10.1093/biomet/41.1-2.100)","m":"reference-list extractor"},{"id":"e1745","i":"unkel-2011--CITES-->roberts1959ewma","s":"unkel-2011","t":"roberts1959ewma","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1467-985x.2011.00714.x -> 10.1080/00401706.1959.10489860)","m":"reference-list extractor"},{"id":"e1746","i":"unkel-2011--CITES-->woodall-2006","s":"unkel-2011","t":"woodall-2006","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/j.1467-985x.2011.00714.x -> 10.1080/00224065.2006.11918593)","m":"reference-list extractor"},{"id":"e1749","i":"ustun-2016--EXTENDS-->charlson-1987","s":"ustun-2016","t":"charlson-1987","r":"EXTENDS","w":2,"why":"SLIM optimises integer-point scoring systems that Charlson-type indices built by rounding regression coefficients.","m":"agent draft (claude)"},{"id":"e1751","i":"vaidya2007rolemining--COMPLEMENTS-->kamra2008anomalousaccess","s":"vaidya2007rolemining","t":"kamra2008anomalousaccess","r":"COMPLEMENTS","w":2,"why":"Role mining vs role profiling from SQL logs; both derive roles from observed access","m":"agent draft (claude)"},{"id":"e1752","i":"vakkari-1999-task--COMPLEMENTS-->campbell-1988","s":"vakkari-1999-task","t":"campbell-1988","r":"COMPLEMENTS","w":2,"why":"Both define complexity on the task side (a priori determinability, information cues) separate from performer actions.","m":"agent draft (claude-sonnet-5-5)"},{"id":"e1753","i":"vakkari-1999-task--EXTENDS-->bystrom-1995","s":"vakkari-1999-task","t":"bystrom-1995","r":"EXTENDS","w":2,"why":"Vakkari extends Bystrom and Jarvelin task complexity/information-seeking link with problem structure; information actions are consequences, not definition.","m":"agent draft (claude-sonnet-5-5)"},{"id":"e1755","i":"van-houwelingen-2007--EXTENDS-->cox-1972","s":"van-houwelingen-2007","t":"cox-1972","r":"EXTENDS","w":2,"why":"Landmark dynamic prediction refits Cox-type models at landmark times on subjects still at risk.","m":"agent draft (claude)"},{"id":"e1756","i":"vandenberg-2000--CITES-->campbell-1959-mtmm","s":"vandenberg-2000","t":"campbell-1959-mtmm","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1177/109442810031002 -> 10.1037/h0046016)","m":"reference-list extractor"},{"id":"e1758","i":"vandenberg-2000--EXTENDS-->meredith-1993","s":"vandenberg-2000","t":"meredith-1993","r":"EXTENDS","w":2,"why":"Operationalises Meredith's invariance hierarchy into a testing sequence for applied research.","m":"agent draft (claude)"},{"id":"e1763","i":"vankempen2024green--CONTRASTS_WITH-->pereira2021ranking","s":"vankempen2024green","t":"pereira2021ranking","r":"CONTRASTS_WITH","w":2,"why":"van Kempen et al. (2024) control for algorithmic/library differences in Pereira et al. (2017, 2021), showing why hybrid Python+Rust/C native kernels approach pure Rust performance when FFI copies are eliminated.","m":"agent draft (gemini)"},{"id":"e1764","i":"vanrenen2024redshiftfleet--COMPLEMENTS-->crdb_diagreporting","s":"vanrenen2024redshiftfleet","t":"crdb_diagreporting","r":"COMPLEMENTS","w":2,"why":"Vendor installed-base telemetry for product decisions.","m":"agent draft (gemini)"},{"id":"e1765","i":"vanrenen2024redshiftfleet--CONTRASTS_WITH-->jain2016sqlshare","s":"vanrenen2024redshiftfleet","t":"jain2016sqlshare","r":"CONTRASTS_WITH","w":2,"why":"Provider-side near-census fleet telemetry vs opt-in user query logs for characterising SQL feature use.","m":"agent draft (gemini)"},{"id":"e1767","i":"vasconcelos-2023--CITES-->buçinca-2021","s":"vasconcelos-2023","t":"buçinca-2021","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3579605 -> 10.1145/3449287)","m":"reference-list extractor"},{"id":"e1769","i":"vassiliadis2017rigidity--CITES-->qiu2013schemacoevolution","s":"vassiliadis2017rigidity","t":"qiu2013schemacoevolution","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1016/j.is.2016.06.010 -> 10.1145/2491411.2491431)","m":"reference-list extractor"},{"id":"e1770","i":"vassiliadis2017rigidity--EXTENDS-->skoulis2015growingup","s":"vassiliadis2017rigidity","t":"skoulis2015growingup","r":"EXTENDS","w":2,"why":"Same lineage; table-level lifetime patterns (rigidity)","m":"agent draft (claude)"},{"id":"e1774","i":"vidigal-2026-q1--COMPLEMENTS-->vidigal2026acl","s":"vidigal-2026-q1","t":"vidigal2026acl","r":"COMPLEMENTS","w":2,"why":"Pairs calibrated B2B customer-support classification (ACL 2026) with empirical failure-mode auditing of autonomous incident-queue triage.","m":"agent draft (gemini)"},{"id":"e1775","i":"vidigal-2026-q1--CONTRASTS_WITH-->chen-rcacopilot-2024","s":"vidigal-2026-q1","t":"chen2024rcacopilot","r":"CONTRASTS_WITH","w":3,"why":"Evaluates autonomous RCA across 544 multi-model runs with planted decoys rather than single-model top-1 similarity","m":"seed curator"},{"id":"e1776","i":"vidigal-2026-q1--EXTENDS-->vidigal2026acl","s":"vidigal-2026-q1","t":"vidigal2026acl","r":"EXTENDS","w":3,"why":"Extends outcome-grounded calibration from sentiment (N=533) to autonomous multi-turn RCA agents (n=102)","m":"seed curator"},{"id":"e1777","i":"vidigal-2026-q1--OPERATIONALIZES-->gebru2021datasheets","s":"vidigal-2026-q1","t":"gebru2021datasheets","r":"OPERATIONALIZES","w":2,"why":"Operationalizes dataset provenance, collection-process disclosure, and failure-mode documentation on live incident-queue telemetry under ISO/IEC 42001 and 5259 controls.","m":"agent draft (gemini)"},{"id":"e1778","i":"vidigal-2026-q1--REFUTES-->xiong2024uncertainty","s":"vidigal-2026-q1","t":"xiong2024uncertainty","r":"REFUTES","w":3,"why":"Across n=368 incidents and 787 production runs, self-reported verbalized LLM confidence carries no usable discriminative signal (AUROC 0.460, p=0.315; and 0.518, p=0.661, 95% CI 0.447-0.580), refuting reliance on verbalized confidence for operational gating.","m":"agent draft (gemini)","tr":1},{"id":"e1779","i":"vidigal-2026-q1--USES_METHOD-->murphy1973new","s":"vidigal-2026-q1","t":"murphy1973new","r":"USES_METHOD","w":3,"why":"Applies Murphy (1973) exact Brier partition (MCB - DSC + UNC) to prove verbalized LLM confidence has DSC=0.0038","m":"seed curator"},{"id":"e1780","i":"vidigal-2026-q2--EXTENDS-->vidigal-2026-q1","s":"vidigal-2026-q2","t":"vidigal-2026-q1","r":"EXTENDS","w":3,"why":"Stratifies 4 frontier models across routine vs engineering-escalated incidents (n=89) and introduces Evidence Recovery","m":"seed curator"},{"id":"e1781","i":"vidigal-2026-q2--OPERATIONALIZES-->xu2002prefix","s":"vidigal-2026-q2","t":"xu2002prefix","r":"OPERATIONALIZES","w":2,"why":"Enforces deterministic source-side redaction and monotonic local-model privacy filtering before support-queue payloads reach external frontier model APIs.","m":"agent draft (gemini)"},{"id":"e1782","i":"vidigal2026acl--OPERATIONALIZES-->saerens2002prior","s":"vidigal2026acl","t":"saerens2002prior","r":"OPERATIONALIZES","w":3,"why":"Applies exact prior-shift log-odds calibration (-2.9704) from Saerens et al. (2002) to production B2B support sentiment classification with verified CSAT ground truth.","m":"agent draft (gemini)"},{"id":"e1783","i":"vidigal2026acl--USES_METHOD-->mcc2025imbalance","s":"vidigal2026acl","t":"mcc2025imbalance","r":"USES_METHOD","w":3,"why":"Uses Matthews Correlation Coefficient (MCC) to evaluate 20:1 imbalanced B2B CSAT queues","m":"seed curator"},{"id":"e1784","i":"vidigal2026acl--USES_METHOD-->saerens2002prior","s":"vidigal2026acl","t":"saerens2002prior","r":"USES_METHOD","w":1,"why":"Applies Saerens prior-odds log-shift (-2.9704) to restore rare-event posterior calibration.","m":"agent draft (gemini)"},{"id":"e1785","i":"vidigal2026gorust--OPERATIONALIZES-->jung2021safe","s":"vidigal2026gorust","t":"jung2021safe","r":"OPERATIONALIZES","w":2,"why":"Empirically validates safe Rust (`#![forbid(unsafe_code)]`) performance and memory bounds on YugabyteDB telemetry collection workloads.","m":"agent draft (gemini)"},{"id":"e1786","i":"vidigal2026rulepacks--COMPLEMENTS-->jeyakumar2019explainit","s":"vidigal2026rulepacks","t":"jeyakumar2019explainit","r":"COMPLEMENTS","w":2,"why":"Population-calibrated three-layer thresholds for 212 database diagnostic rules complement declarative RCA engines","m":"agent draft (gemini)"},{"id":"e1787","i":"vidigal2026rulepacks--OPERATIONALIZES-->jordan2016souffle","s":"vidigal2026rulepacks","t":"jordan2016souffle","r":"OPERATIONALIZES","w":2,"why":"Applies Soufflé's compiled-Datalog synthesis pattern to population-calibrated fleet diagnostic rulepacks.","m":"agent draft (gemini)"},{"id":"e1788","i":"vidigal2026wtlformal--COMPLEMENTS-->vikram2023pbt","s":"vidigal2026wtlformal","t":"vikram2023pbt","r":"COMPLEMENTS","w":2,"why":"Combines proptest C-ABI differential property-based testing and cargo-mutants with zero-axiom Lean 4 formal proofs.","m":"agent draft (gemini)"},{"id":"e1789","i":"vidigal2026wtlformal--OPERATIONALIZES-->claessen2000quickcheck","s":"vidigal2026wtlformal","t":"claessen2000quickcheck","r":"OPERATIONALIZES","w":2,"why":"Uses proptest random generators over C-ABI FFI to differentially test Rust glog parsing against Lean 4.","m":"agent draft (gemini)"},{"id":"e1790","i":"vidigal2026wtlformal--OPERATIONALIZES-->jia2011mutation","s":"vidigal2026wtlformal","t":"jia2011mutation","r":"OPERATIONALIZES","w":2,"why":"Applies cargo-mutants across 121 workspace mutants (killing 72/72 semantic data-plane mutants) alongside Lean 4 proofs.","m":"agent draft (gemini)"},{"id":"e1791","i":"vidigal2026wtllakehouse--CONTRASTS_WITH-->jiang2024lilac","s":"vidigal2026wtllakehouse","t":"jiang2024lilac","r":"CONTRASTS_WITH","w":2,"why":"Replaces LILAC's runtime LLM-populated template cache with compile-time source-AST template extraction and deterministic fallback shapes.","m":"agent draft (gemini)"},{"id":"e1792","i":"vidigal2026wtllakehouse--EXTENDS-->akidau2015dataflow","s":"vidigal2026wtllakehouse","t":"akidau2015dataflow","r":"EXTENDS","w":2,"why":"Extends Dataflow session windowing with Two-Tier Statistical Temporal RLE (AutoEwma log-space IAT threshold + 60s span cap + Welford/quantile/outlier reservoir).","m":"agent draft (gemini)"},{"id":"e1794","i":"vidigal2026wtllakehouse--OPERATIONALIZES-->khan2022guidelines","s":"vidigal2026wtllakehouse","t":"khan2022guidelines","r":"OPERATIONALIZES","w":2,"why":"Implements Khan et al.'s canonical wildcard post-processing and token-normalization rules to prevent template over-splitting.","m":"agent draft (gemini)"},{"id":"e1795","i":"vidigal2026wtllakehouse--OPERATIONALIZES-->zeng2023columnar","s":"vidigal2026wtllakehouse","t":"zeng2023columnar","r":"OPERATIONALIZES","w":2,"why":"Calibrates Parquet row-group sizing (256k rows), time-primary k-way merge sorting, and W4 per-column SBBF NDV sizing (18.96x Bloom byte reduction) over log lakehouses.","m":"agent draft (gemini)"},{"id":"e1796","i":"vidigal2026wtllakehouse--USES_METHOD-->aho1975efficient","s":"vidigal2026wtllakehouse","t":"aho1975efficient","r":"USES_METHOD","w":2,"why":"Uses multi-pattern Aho-Corasick finite-state automata ordered by literal length for single-pass log template prefix matching.","m":"agent draft (gemini)"},{"id":"e1797","i":"vidigal2026wtllakehouse--USES_METHOD-->hoefler-2015","s":"vidigal2026wtllakehouse","t":"hoefler-2015","r":"USES_METHOD","w":2,"why":"Applies Hoefler & Belli's reproducible systems-benchmarking rules to compile-time AST log template mining and local-vs-cloud telemetry processing.","m":"agent draft (gemini)"},{"id":"e1798","i":"vidigal2026wtllakehouse--USES_METHOD-->morrison-1968","s":"vidigal2026wtllakehouse","t":"morrison-1968","r":"USES_METHOD","w":2,"why":"Verifies candidate log templates by walking compressed PATRICIA radix-tree literal and wildcard segments with SIMD substring search.","m":"agent draft (gemini)"},{"id":"e1799","i":"vidigal2026wtlspeedup--OPERATIONALIZES-->shypula2024pie","s":"vidigal2026wtlspeedup","t":"shypula2024pie","r":"OPERATIONALIZES","w":2,"why":"Operationalizes profiler-in-the-loop iterative code optimization with strict multiset-equivalence and pre-registered benchmark gates on a 27.6k-LOC Rust log engine.","m":"agent draft (gemini)"},{"id":"e1800","i":"vidigal2027mcpfirewall--COMPLEMENTS-->alexander-2023-arxiv","s":"vidigal2026mcpfirewall","t":"alexander-2023-arxiv","r":"COMPLEMENTS","w":2,"why":"Pairs runtime monotonic PII/secret egress redaction with tamper-evident cryptographic audit logging.","m":"agent draft (gemini)"},{"id":"e1801","i":"vidigal2027pv2--USES_METHOD-->lipsitch-2010","s":"vidigal2027pv2","t":"lipsitch-2010","r":"USES_METHOD","w":2,"why":"account-concentration masquerading as regression is detected with negative-control style comparisons","m":"agent draft (gemini)"},{"id":"e1806","i":"voynov-2026--COMPLEMENTS-->satyanarayan-2017","s":"voynov-2026","t":"satyanarayan-2017","r":"COMPLEMENTS","w":2,"why":"Pairs Vega-Lite's declarative grammar of interactive graphics with deterministic vector compilation for automated technical reports.","m":"agent draft (gemini)"},{"id":"e1809","i":"wagner2002segmented--COMPLEMENTS-->lopezbernal2017its","s":"wagner2002segmented","t":"lopezbernal2017its","r":"COMPLEMENTS","w":2,"why":"segment-length and autocorrelation guidance for segmented ITS","m":"agent draft (gemini)"},{"id":"e1810","i":"wallace-2005--CITES-->brown-2005","s":"wallace-2005","t":"brown-2005","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1287/msom.1050.0086 -> 10.1198/016214504000001808)","m":"reference-list extractor"},{"id":"e1811","i":"wallace-2005--CITES-->garnett2002erlanga","s":"wallace-2005","t":"garnett2002erlanga","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1287/msom.1050.0086 -> 10.1287/msom.4.3.208.7753)","m":"reference-list extractor"},{"id":"e1812","i":"wallace-2005--CITES-->halfin-1981","s":"wallace-2005","t":"halfin-1981","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1287/msom.1050.0086 -> 10.1287/opre.29.3.567)","m":"reference-list extractor"},{"id":"e1813","i":"wallace-2005--USES_METHOD-->gans-2003","s":"wallace-2005","t":"gans-2003","r":"USES_METHOD","w":2,"why":"skill-based routing staffing within call-centre queueing framework","m":"agent draft (gemini)"},{"id":"e1815","i":"wang2019hyperscan--COMPLEMENTS-->he2017drain","s":"wang2019hyperscan","t":"he2017drain","r":"COMPLEMENTS","w":3,"why":"Pairs SIMD literal-prefiltered regular-expression matching for known diagnostic signatures with fixed-depth parse-tree clustering (Drain) for unstructured log templates.","m":"agent draft (gemini)"},{"id":"e1816","i":"wang2019hyperscan--USES_METHOD-->horspool1980practical","s":"wang2019hyperscan","t":"horspool1980practical","r":"USES_METHOD","w":3,"why":"Extracts mandatory literal substrings from regular expressions and scans buffers via shift-based literal prefiltering before invoking NFA/DFA automata.","m":"agent draft (gemini)"},{"id":"e1817","i":"wang2021outage--CITES-->chen2019incidenttriage","s":"wang2021outage","t":"chen2019incidenttriage","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse43902.2021.00085 -> 10.1109/icse-seip.2019.00020)","m":"reference-list extractor"},{"id":"e1819","i":"wang2021outage--COMPLEMENTS-->chen2020lidar","s":"wang2021outage","t":"chen2020lidar","r":"COMPLEMENTS","w":2,"why":"Outage aggregation vs pairwise incident linking","m":"agent draft (gemini)"},{"id":"e1820","i":"wang2023trustcodegen--EXTENDS-->lee-2004","s":"wang2023trustcodegen","t":"lee-2004","r":"EXTENDS","w":2,"why":"Applies trust-calibration to AI code tools via community and in-tool reliability cues","m":"agent draft (gemini)"},{"id":"e1821","i":"wang2024humanllmannot--CITES-->bansal-2021","s":"wang2024humanllmannot","t":"bansal-2021","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3613904.3641960 -> 10.1145/3411764.3445717)","m":"reference-list extractor"},{"id":"e1822","i":"wang2024humanllmannot--CITES-->gilardi2023chatgpt","s":"wang2024humanllmannot","t":"gilardi2023chatgpt","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3613904.3641960 -> 10.1073/pnas.2305016120)","m":"reference-list extractor"},{"id":"e1823","i":"wang2024humanllmannot--CITES-->lingren2014preannotation","s":"wang2024humanllmannot","t":"lingren2014preannotation","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3613904.3641960 -> 10.1136/amiajnl-2013-001837)","m":"reference-list extractor"},{"id":"e1824","i":"wang2024humanllmannot--CITES-->zhang2020confidenceexplanation","s":"wang2024humanllmannot","t":"zhang2020confidenceexplanation","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3613904.3641960 -> 10.1145/3351095.3372852)","m":"reference-list extractor"},{"id":"e1825","i":"wang2024humanllmannot--CITES-->ziems-2024","s":"wang2024humanllmannot","t":"ziems-2024","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3613904.3641960 -> 10.1162/coli_a_00502)","m":"reference-list extractor"},{"id":"e1826","i":"wang2024humanllmannot--COMPLEMENTS-->fisch2024stratppi","s":"wang2024humanllmannot","t":"fisch2024stratppi","r":"COMPLEMENTS","w":2,"why":"Human verification of LLM labels vs statistical rectification of LLM labels","m":"agent draft (gemini)"},{"id":"e1830","i":"wei-1992--CITES-->andersen-1982","s":"wei-1992","t":"andersen-1982","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1002/sim.4780111409 -> 10.1214/aos/1176345976)","m":"reference-list extractor"},{"id":"e1831","i":"wei-1992--CONTRASTS_WITH-->cox-1972","s":"wei-1992","t":"cox-1972","r":"CONTRASTS_WITH","w":2,"why":"Argues the AFT model as a directly interpretable alternative to proportional hazards.","m":"agent draft (claude)"},{"id":"e1832","i":"wei-2023--CITES-->boyer1977fast","s":"wei2023loggrep","t":"boyer1977fast","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3552326.3567484 -> 10.1145/359842.359859)","m":"reference-list extractor"},{"id":"e1833","i":"wei-2023--CITES-->lin2015cowic","s":"wei2023loggrep","t":"lin2015cowic","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3552326.3567484 -> 10.1109/ccgrid.2015.45)","m":"reference-list extractor"},{"id":"e1835","i":"wei-2023--CITES-->yuan2010sherlog","s":"wei2023loggrep","t":"yuan2010sherlog","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3552326.3567484 -> 10.1145/1736020.1736038)","m":"reference-list extractor"},{"id":"e1836","i":"wei-2023--CITES-->zhu2019tools","s":"wei2023loggrep","t":"zhu2019tools","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3552326.3567484 -> 10.1109/icse-seip.2019.00021)","m":"reference-list extractor"},{"id":"e1837","i":"wei-2023--EXTENDS-->he2017drain","s":"wei2023loggrep","t":"he2017drain","r":"EXTENDS","w":2,"why":"LogGrep exploits static template decomposition and runtime variable sub-patterns for capsule-level compressed log querying without full decompression.","m":"agent draft (gemini)"},{"id":"e1838","i":"wei2021logreducer--USES_METHOD-->he2017drain","s":"wei2021logreducer","t":"he2017drain","r":"USES_METHOD","w":2,"why":"LogReducer builds its template mining front-end on Drain's fixed-depth prefix tree before applying delta timestamp and numeric correlation encoding.","m":"agent draft (gemini)"},{"id":"e1839","i":"wei2023loggrep--COMPLEMENTS-->rodrigues2021clp","s":"wei2023loggrep","t":"rodrigues2021clp","r":"COMPLEMENTS","w":2,"why":"Both CLP and LogGrep enable direct search over compressed logs without full decompression: CLP via logtype + variable dictionary indexes, and LogGrep via fine-grained runtime capsule value stamps.","m":"agent draft (gemini)"},{"id":"e1841","i":"wei2023loggrep--EXTENDS-->rodrigues2021clp","s":"wei2023loggrep","t":"rodrigues2021clp","r":"EXTENDS","w":2,"why":"Decomposes CLP-style coarse variable slots ('<*>') into fine-grained runtime sub-pattern 'capsules' with min/max and bitmap value stamps for sub-block query skipping.","m":"agent draft (gemini)"},{"id":"e1842","i":"wei2023loggrep--EXTENDS-->wei2021logreducer","s":"wei2023loggrep","t":"wei2021logreducer","r":"EXTENDS","w":2,"why":"LogGrep extends LogReducer's static template decomposition by further splitting variable slots into fine-grained runtime sub-patterns ('capsules') with value stamps for block skipping.","m":"agent draft (gemini)"},{"id":"e1843","i":"weiss-2007--EXTENDS-->panjer-2007","s":"weiss-2007","t":"panjer-2007","r":"EXTENDS","w":2,"why":"both predict bug lifetime/effort from early report attributes; Weiss uses nearest-neighbour text similarity","m":"agent draft (gemini)"},{"id":"e1844","i":"weston-2025-arxiv--CONTRASTS_WITH-->zeng2023columnar","s":"weston-2025-arxiv","t":"zeng2023columnar","r":"CONTRASTS_WITH","w":2,"why":"Lance targets random access; Zeng targets scans","m":"agent draft (gemini)"},{"id":"e1845","i":"weston-2025-arxiv--EXTENDS-->melnik-2010","s":"weston-2025-arxiv","t":"melnik-2010","r":"EXTENDS","w":2,"why":"Lance revisits Dremel rep/def structural encoding for random access","m":"agent draft (gemini)"},{"id":"e1854","i":"won2018barrier--COMPLEMENTS-->chidambaram-2013","s":"won2018barrier","t":"chidambaram-2013","r":"COMPLEMENTS","w":2,"why":"Both locate fsync cost in cache flushes and ordering; different remedies (barrier command vs optimistic consistency).","m":"agent draft (claude)"},{"id":"e1855","i":"woodall-2006--CITES-->benneyan-2003","s":"woodall-2006","t":"benneyan-2003","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1080/00224065.2006.11918593 -> 10.1136/qhc.12.6.458)","m":"reference-list extractor"},{"id":"e1856","i":"woodall-2006--COMPLEMENTS-->roberts1959ewma","s":"woodall-2006","t":"roberts1959ewma","r":"COMPLEMENTS","w":2,"why":"Woodall recommends EWMA/CUSUM for small sustained shifts in surveillance; Roberts introduced the EWMA chart.","m":"agent draft (claude)"},{"id":"e1857","i":"wooldridge2023nonlineardid--CITES-->borusyak2024imputation","s":"borusyak2024imputation","t":"wooldridge2023nonlineardid","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1093/ectj/utad016 -> borusyak2024imputation)","m":"reference-list extractor"},{"id":"e1858","i":"wooldridge2023nonlineardid--CITES-->callaway2021did","s":"wooldridge2023nonlineardid","t":"callaway2021did","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1093/ectj/utad016 -> 10.1016/j.jeconom.2020.12.001)","m":"reference-list extractor"},{"id":"e1859","i":"wooldridge2023nonlineardid--CITES-->goodmanbacon2021decomp","s":"wooldridge2023nonlineardid","t":"goodmanbacon2021decomp","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1093/ectj/utad016 -> 10.1016/j.jeconom.2021.03.014)","m":"reference-list extractor"},{"id":"e1860","i":"wooldridge2023nonlineardid--CITES-->sun2021event","s":"wooldridge2023nonlineardid","t":"sun2021event","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1093/ectj/utad016 -> 10.1016/j.jeconom.2020.09.006)","m":"reference-list extractor"},{"id":"e1861","i":"wu2014crashlocator--CITES-->dhaliwal2011crash","s":"wu2014crashlocator","t":"dhaliwal2011crash","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2610384.2610386 -> 10.1109/icsm.2011.6080800)","m":"reference-list extractor"},{"id":"e1862","i":"wu2014crashlocator--CITES-->glerum2009wer","s":"wu2014crashlocator","t":"glerum2009wer","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2610384.2610386 -> 10.1145/1629575.1629586)","m":"reference-list extractor"},{"id":"e1863","i":"wu2014crashlocator--EXTENDS-->kim2011crashgraphs","s":"wu2014crashlocator","t":"kim2011crashgraphs","r":"EXTENDS","w":2,"why":"Extends crash graphs with TF-IDF-style inverse bucket frequency weighting and inverse stack depth to rank the specific faulty function and suppress ubiquitous dispatch frames.","m":"agent draft (gemini)"},{"id":"e1864","i":"wu2023changeinduced--COMPLEMENTS-->li2020gandalf","s":"wu2023changeinduced","t":"li2020gandalf","r":"COMPLEMENTS","w":2,"why":"Empirical incident-change study vs deployment-time change-safety system","m":"agent draft (gemini)"},{"id":"e1865","i":"wu2023changeinduced--COMPLEMENTS-->maclure1991casecrossover","s":"wu2023changeinduced","t":"maclure1991casecrossover","r":"COMPLEMENTS","w":2,"why":"Change-before-incident prevalence needs a within-subject referent window (case-crossover) to become a lift.","m":"agent draft (gemini)"},{"id":"e1869","i":"xia2018comprehension--USES_METHOD-->halfaker2015sessions","s":"xia2018comprehension","t":"halfaker2015sessions","r":"USES_METHOD","w":2,"why":"Xia et al. split interaction logs into effective working sessions with a 1-hour idle threshold, the same gap-based sessionization family (not a citation claim; methodological match).","m":"agent draft (claude)"},{"id":"e1876","i":"xiong2024uncertainty--CONTRASTS_WITH-->saerens2002prior","s":"xiong2024uncertainty","t":"saerens2002prior","r":"CONTRASTS_WITH","w":2,"why":"Contrasts uncalibrated verbalized self-confidence elicited from LLMs against likelihood-based EM prior-shift correction on class-imbalanced distributions.","m":"agent draft (gemini)"},{"id":"e1877","i":"xu-2025--EXTENDS-->zhang2024ebsglory","s":"xu-2025","t":"zhang2024ebsglory","r":"EXTENDS","w":2,"why":"TOS extension of the FAST'24 EBS paper.","m":"agent draft (claude)"},{"id":"e1880","i":"xu2015knobs--COMPLEMENTS-->yin2011configerrors","s":"xu2015knobs","t":"yin2011configerrors","r":"COMPLEMENTS","w":2,"why":"Over-designed knobs and empirical config errors: together justify hazard-ranked gflag review over popularity","m":"agent draft (gemini)"},{"id":"e1881","i":"xu2016pcheck--EXTENDS-->yin2011configerrors","s":"xu2016pcheck","t":"yin2011configerrors","r":"EXTENDS","w":2,"why":"PCheck acts on Yin et al.'s finding that configuration errors are frequent and severe by detecting latent configuration errors early.","m":"agent draft (claude)"},{"id":"e1882","i":"xu2016pcheck--OPERATIONALIZES-->xu2015csur","s":"xu2016pcheck","t":"xu2015csur","r":"OPERATIONALIZES","w":2,"why":"Early config checks as the systems remedy surveyed in CSUR","m":"agent draft (gemini)"},{"id":"e1884","i":"xu2024divlog--CITES-->he2017drain","s":"xu2024divlog","t":"he2017drain","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3597503.3639155 -> 10.1109/icws.2017.13)","m":"reference-list extractor"},{"id":"e1885","i":"xu2024divlog--CITES-->zhu2019tools","s":"xu2024divlog","t":"zhu2019tools","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3597503.3639155 -> 10.1109/icse-seip.2019.00021)","m":"reference-list extractor"},{"id":"e1886","i":"xu2024divlog--EXTENDS-->le-2023","s":"xu2024divlog","t":"le-2023","r":"EXTENDS","w":2,"why":"Replaces LogPPT's gradient-based prompt tuning with training-free Determinantal Point Process (DPP) candidate sampling and k-NN in-context LLM prompting (DivLog).","m":"agent draft (gemini)"},{"id":"e1888","i":"xu2024ragkgcs--EXTENDS-->brodie2005symptom","s":"xu2024ragkgcs","t":"brodie2005symptom","r":"EXTENDS","w":2,"why":"Historical-issue retrieval for new tickets: curated symptom DB generalized to KG plus RAG over past tickets.","m":"agent draft (claude)"},{"id":"e1889","i":"yang2018hownot--CITES-->chen2016redundantdataaccess","s":"yang2018webappperf","t":"chen2016redundantdataaccess","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3180155.3180194 -> 10.1109/tse.2016.2553039)","m":"reference-list extractor"},{"id":"e1890","i":"yang2018hownot--COMPLEMENTS-->chen2014ormantipatterns","s":"yang2018webappperf","t":"chen2014ormantipatterns","r":"COMPLEMENTS","w":2,"why":"Empirical Rails ORM anti-patterns complement Java/Hibernate detection","m":"agent draft (claude)"},{"id":"e1893","i":"yang2024sweagent--EXTENDS-->jimenez2024swebench","s":"yang2024sweagent","t":"jimenez2024swebench","r":"EXTENDS","w":2,"why":"SWE-agent introduces the agent-computer interface evaluated on SWE-bench","m":"agent draft (qwen)"},{"id":"e1898","i":"yao2024taubench--COMPLEMENTS-->dellacqua2023jagged","s":"yao2024taubench","t":"dellacqua2023jagged","r":"COMPLEMENTS","w":2,"why":"Reliability across reruns (pass^k) and frontier-dependent correctness: both say one-shot speed/quality metrics overstate AI usefulness","m":"agent draft (gemini)"},{"id":"e1906","i":"yiwei-2026-arxiv--EXTENDS-->ni2026genaialibaba","s":"yiwei-2026-arxiv","t":"ni2026genaialibaba","r":"EXTENDS","w":2,"why":"Same Alibaba customer-service setting; moves from GenAI copilot to agentic AI with human-in-the-loop escalation.","m":"agent draft (gemini)"},{"id":"e1907","i":"young2017modeluncertainty--CITES-->ioannidis2005why","s":"young2017modeluncertainty","t":"ioannidis2005why","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1177/0049124115610347 -> 10.1371/journal.pmed.0020124)","m":"reference-list extractor"},{"id":"e1908","i":"young2017modeluncertainty--CITES-->simmons2011false","s":"young2017modeluncertainty","t":"simmons2011false","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1177/0049124115610347 -> 10.1177/0956797611417632)","m":"reference-list extractor"},{"id":"e1909","i":"yu-2023--CITES-->dai2020logram","s":"yu-2023","t":"dai2020logram","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tsc.2023.3270566 -> 10.1109/tse.2020.3007554)","m":"reference-list extractor"},{"id":"e1910","i":"yu-2023--CITES-->du2016spell","s":"yu-2023","t":"du2016spell","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tsc.2023.3270566 -> 10.1109/icdm.2016.0103)","m":"reference-list extractor"},{"id":"e1911","i":"yu-2023--CITES-->fu2009execution","s":"yu-2023","t":"fu2009execution","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tsc.2023.3270566 -> 10.1109/icdm.2009.60)","m":"reference-list extractor"},{"id":"e1913","i":"yu-2023--CITES-->keiichi-2016-arxiv","s":"yu-2023","t":"keiichi-2016-arxiv","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tsc.2023.3270566 -> keiichi-2016-arxiv)","m":"reference-list extractor"},{"id":"e1914","i":"yu-2023--CITES-->khan-2022","s":"yu-2023","t":"khan2022guidelines","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tsc.2023.3270566 -> 10.1145/3510003.3510101)","m":"reference-list extractor"},{"id":"e1915","i":"yu-2023--CITES-->le2022logad","s":"yu-2023","t":"le2022logad","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3611643.3616355 -> 10.1145/3510003.3510155)","m":"reference-list extractor"},{"id":"e1917","i":"yu-2023--CITES-->zhu2019tools","s":"yu-2023","t":"zhu2019tools","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/tsc.2023.3270566 -> 10.1109/icse-seip.2019.00021)","m":"reference-list extractor"},{"id":"e1918","i":"yu-2023--CONTRASTS_WITH-->chu-2021","s":"yu-2023","t":"chu-2021","r":"CONTRASTS_WITH","w":2,"why":"Solves prefix divergence via a deterministic bidirectional parallel tree around the most frequent co-occurring word combination rather than iterative probabilistic graph clique cutting.","m":"agent draft (gemini)"},{"id":"e1919","i":"yu-2023--EXTENDS-->he2017drain","s":"yu-2023","t":"he2017drain","r":"EXTENDS","w":2,"why":"Generalizes Drain's unidirectional root-to-leaf prefix tree into a bidirectional parallel tree (Brain) rooted at the longest common word combination to handle variable prefixes.","m":"agent draft (gemini)"},{"id":"e1920","i":"yu2023nezha_fse--CITES-->he-2018","s":"yu2023nezha_fse","t":"he-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3611643.3616249 -> 10.1145/3236024.3236083)","m":"reference-list extractor"},{"id":"e1921","i":"yu2023nezha_fse--CITES-->he2017drain","s":"yu2023nezha_fse","t":"he2017drain","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3611643.3616249 -> 10.1109/icws.2017.13)","m":"reference-list extractor"},{"id":"e1922","i":"yu2023nezha_fse--CITES-->lin2016logcluster","s":"yu2023nezha_fse","t":"lin2016logcluster","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3611643.3616249 -> 10.1145/2889160.2889232)","m":"reference-list extractor"},{"id":"e1925","i":"yu2023nezha_fse--COMPLEMENTS-->lee2023eadro_icse","s":"yu2023nezha_fse","t":"lee2023eadro_icse","r":"COMPLEMENTS","w":2,"why":"Both fuse logs+metrics+traces for RCA; Eadro is supervised end-to-end fusion, Nezha is unsupervised interpretable fault-free vs faulty phase contrast.","m":"agent draft (claude)"},{"id":"e1926","i":"yu2024denum--CITES-->he2017drain","s":"yu2024denum","t":"he2017drain","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3691620.3695474 -> 10.1109/icws.2017.13)","m":"reference-list extractor"},{"id":"e1929","i":"yu2024denum--CITES-->wei2023loggrep","s":"yu2024denum","t":"wei2023loggrep","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3691620.3695474 -> 10.1145/3552326.3567484)","m":"reference-list extractor"},{"id":"e1931","i":"yuan2010sherlog--CITES-->glerum2009wer","s":"yuan2010sherlog","t":"glerum2009wer","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1736020.1736038 -> 10.1145/1629575.1629586)","m":"reference-list extractor"},{"id":"e1933","i":"yuan2011logenhancer--CITES-->castro2008bugreporting","s":"yuan2011logenhancer","t":"castro2008bugreporting","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1950365.1950369 -> 10.1145/1346281.1346322)","m":"reference-list extractor"},{"id":"e1934","i":"yuan2011logenhancer--CITES-->glerum2009wer","s":"yuan2011logenhancer","t":"glerum2009wer","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/1950365.1950369 -> 10.1145/1629575.1629586)","m":"reference-list extractor"},{"id":"e1935","i":"yuan2011logenhancer--EXTENDS-->yuan2010sherlog","s":"yuan2011logenhancer","t":"yuan2010sherlog","r":"EXTENDS","w":2,"why":"Uses SherLog's post-mortem symbolic constraint model to identify and automatically inject missing causally relevant variables into source-code log statements.","m":"agent draft (gemini)"},{"id":"e1936","i":"yuan2012characterizing--COMPLEMENTS-->yuan2011logenhancer","s":"yuan2012characterizing","t":"yuan2011logenhancer","r":"COMPLEMENTS","w":2,"why":"Empirically characterizes real-world logging density (1 per 30 LOC) and patch churn (18% of commits modify logs, 39% alter static template text) across Apache, OpenSSH, PostgreSQL, and Squid.","m":"agent draft (gemini)"},{"id":"e1939","i":"zaharia2013dstreams--CITES-->akidau2013millwheel","s":"zaharia2013dstreams","t":"akidau2013millwheel","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/2517349.2522737 -> 10.14778/2536222.2536229)","m":"reference-list extractor"},{"id":"e1944","i":"zeng-2023--CITES-->armbrust-2020","s":"zeng2023columnar","t":"armbrust-2020","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3626292.3626298 -> 10.14778/3415478.3415560)","m":"reference-list extractor"},{"id":"e1945","i":"zeng-2023--CITES-->melnik-2010","s":"zeng2023columnar","t":"melnik-2010","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3626292.3626298 -> 10.14778/1920841.1920886)","m":"reference-list extractor"},{"id":"e1946","i":"zeng-2023--CITES-->pelkonen2015gorilla","s":"zeng2023columnar","t":"pelkonen2015gorilla","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.14778/3626292.3626298 -> 10.14778/2824032.2824078)","m":"reference-list extractor"},{"id":"e1948","i":"zeng-2023--COMPLEMENTS-->raasveldt2019duckdb","s":"zeng2023columnar","t":"raasveldt2019duckdb","r":"COMPLEMENTS","w":2,"why":"Empirically evaluates Parquet and ORC columnar encoding, row-group sizing, and zone-map predicate pushdown exploited by vectorized embedded engines like DuckDB.","m":"agent draft (gemini)"},{"id":"e1949","i":"zeng2023columnar--CITES-->cleveland1984graphical","s":"zeng2023columnar","t":"cleveland1984graphical","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3544548.3581349 -> 10.1080/01621459.1984.10478080)","m":"reference-list extractor"},{"id":"e1950","i":"zeng2023columnar--CITES-->heer2010crowdsourcing","s":"zeng2023columnar","t":"heer2010crowdsourcing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3544548.3581349 -> 10.1145/1753326.1753357)","m":"reference-list extractor"},{"id":"e1951","i":"zeng2023columnar--CITES-->kim-2018","s":"zeng2023columnar","t":"kim-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3544548.3581349 -> 10.1111/cgf.13409)","m":"reference-list extractor"},{"id":"e1952","i":"zeng2023columnar--CITES-->satyanarayan-2017","s":"zeng2023columnar","t":"satyanarayan-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3544548.3581349 -> 10.1109/tvcg.2016.2599030)","m":"reference-list extractor"},{"id":"e1953","i":"zeng2023columnar--CITES-->shneiderman1996eyes","s":"zeng2023columnar","t":"shneiderman1996eyes","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3544548.3581349 -> 10.1109/vl.1996.545307)","m":"reference-list extractor"},{"id":"e1954","i":"zeng2023columnar--CITES-->szafir-2018","s":"zeng2023columnar","t":"szafir-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3544548.3581349 -> 10.1109/tvcg.2017.2744359)","m":"reference-list extractor"},{"id":"e1955","i":"zeng2023columnar--COMPLEMENTS-->graefe2006sorting","s":"zeng2023columnar","t":"graefe2006sorting","r":"COMPLEMENTS","w":2,"why":"Zeng et al.'s empirical quantification of Parquet/ORC decoding CPU costs (LZ4 vs ZSTD, row-group sizing, and page indexes) informs the codec and row-group trade-offs for Graefe-style external sort spill runs vs final archival output.","m":"agent draft (gemini)"},{"id":"e1956","i":"zeng2023graphicalperception--CITES-->cleveland1984graphical","s":"zeng2023graphicalperception","t":"cleveland1984graphical","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3544548.3581349 -> 10.1080/01621459.1984.10478080)","m":"reference-list extractor"},{"id":"e1957","i":"zeng2023graphicalperception--CITES-->correll-2020","s":"zeng2023graphicalperception","t":"correll-2020","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3544548.3581349 -> 10.1145/3313831.3376222)","m":"reference-list extractor"},{"id":"e1958","i":"zeng2023graphicalperception--CITES-->gramazio-2017","s":"zeng2023graphicalperception","t":"gramazio-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3544548.3581349 -> gramazio-2017)","m":"reference-list extractor"},{"id":"e1959","i":"zeng2023graphicalperception--CITES-->heer2010crowdsourcing","s":"zeng2023graphicalperception","t":"heer2010crowdsourcing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3544548.3581349 -> 10.1145/1753326.1753357)","m":"reference-list extractor"},{"id":"e1960","i":"zeng2023graphicalperception--CITES-->liu-2018","s":"zeng2023graphicalperception","t":"liu-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3544548.3581349 -> 10.1145/3173574.3174172)","m":"reference-list extractor"},{"id":"e1961","i":"zeng2023graphicalperception--CITES-->satyanarayan-2017","s":"zeng2023graphicalperception","t":"satyanarayan-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3544548.3581349 -> 10.1109/tvcg.2016.2599030)","m":"reference-list extractor"},{"id":"e1962","i":"zeng2023graphicalperception--CITES-->schloss-2019","s":"zeng2023graphicalperception","t":"schloss-2019","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3544548.3581349 -> 10.1109/tvcg.2018.2865147)","m":"reference-list extractor"},{"id":"e1963","i":"zeng2023graphicalperception--CITES-->shneiderman1996eyes","s":"zeng2023graphicalperception","t":"shneiderman1996eyes","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3544548.3581349 -> 10.1109/vl.1996.545307)","m":"reference-list extractor"},{"id":"e1964","i":"zeng2023graphicalperception--CITES-->skau-2016","s":"zeng2023graphicalperception","t":"skau-2016","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3544548.3581349 -> 10.1111/cgf.12888)","m":"reference-list extractor"},{"id":"e1965","i":"zeng2023graphicalperception--CITES-->szafir-2018","s":"zeng2023graphicalperception","t":"szafir-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3544548.3581349 -> 10.1109/tvcg.2017.2744359)","m":"reference-list extractor"},{"id":"e1966","i":"zeng2023graphicalperception--COMPLEMENTS-->kim-2018","s":"zeng2023graphicalperception","t":"kim-2018","r":"COMPLEMENTS","w":2,"why":"collates graphical perception results into task-conditioned rules for recommendation","m":"agent draft (claude)"},{"id":"e1970","i":"zerrow2025arrow--COMPLEMENTS-->lamb2024datafusion","s":"zerrow2025arrow","t":"lamb2024datafusion","r":"COMPLEMENTS","w":2,"why":"Zerrow's zero-copy Arrow pipeline design complements DataFusion/arrow-rs in-memory RecordBatch execution by eliminating cross-boundary serialization.","m":"agent draft (gemini)"},{"id":"e1971","i":"zerrow2025arrow--COMPLEMENTS-->zeng2023columnar","s":"zerrow2025arrow","t":"zeng2023columnar","r":"COMPLEMENTS","w":2,"why":"Zerrow complements Zeng et al.'s (PVLDB 2023) columnar storage evaluation by eliminating in-memory serialization copies between Rust Arrow/Parquet producers and Python/DuckDB consumers.","m":"agent draft (gemini)"},{"id":"e1972","i":"zhang-2013--CITES-->giger-2010","s":"zhang-2013","t":"giger-2010","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse.2013.6606654 -> 10.1145/1808920.1808933)","m":"reference-list extractor"},{"id":"e1974","i":"zhang-2013--CITES-->kim2006bugfixtime","s":"zhang-2013","t":"kim2006bugfixtime","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse.2013.6606654 -> 10.1145/1137983.1138027)","m":"reference-list extractor"},{"id":"e1975","i":"zhang-2013--COMPLEMENTS-->panjer-2007","s":"zhang-2013","t":"panjer-2007","r":"COMPLEMENTS","w":2,"why":"Industrial vs OSS prediction of bug resolution time from early report attributes","m":"agent draft (claude)"},{"id":"e1979","i":"zhang2020confidenceexplanation--CITES-->dietvorst2015aversion","s":"zhang2020confidenceexplanation","t":"dietvorst2015aversion","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3351095.3372852 -> 10.1037/xge0000033)","m":"reference-list extractor"},{"id":"e1980","i":"zhang2020confidenceexplanation--CITES-->lee-2004","s":"zhang2020confidenceexplanation","t":"lee-2004","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3351095.3372852 -> 10.1518/hfes.46.1.50_30392)","m":"reference-list extractor"},{"id":"e1982","i":"zhang2021upgrade--COMPLEMENTS-->gunawi2014cloudbugs","s":"zhang2021upgrade","t":"gunawi2014cloudbugs","r":"COMPLEMENTS","w":2,"why":"upgrade failures are a focused study of a bug class in the cloud-bug taxonomy","m":"agent draft (gemini)"},{"id":"e1983","i":"zhang2024ebsglory--COMPLEMENTS-->gunawi2018failslow","s":"zhang2024ebsglory","t":"gunawi2018failslow","r":"COMPLEMENTS","w":2,"why":"cloud block-store perspective on slow-storage faults","m":"agent draft (gemini)"},{"id":"e1984","i":"zhang2024genverifiers--COMPLEMENTS-->wang2023selfconsistency","s":"zhang2024genverifiers","t":"wang2023selfconsistency","r":"COMPLEMENTS","w":2,"why":"VRM scoring of sampled candidates is the learned complement to unsupervised majority-vote self-consistency","m":"agent draft (qwen)"},{"id":"e1985","i":"zhang2024genverifiers--EXTENDS-->lightman2023verify","s":"zhang2024genverifiers","t":"lightman2023verify","r":"EXTENDS","w":2,"why":"Generative verifiers recast the reward/verifier model as next-token prediction, scaling the discriminative PRM line","m":"agent draft (qwen)"},{"id":"e1986","i":"zhang2026piitrace--COMPLEMENTS-->xu2002prefix","s":"zhang2026piitrace","t":"xu2002prefix","r":"COMPLEMENTS","w":2,"why":"Combines structured PII and secret entity detection in telemetry text with Crypto-PAn prefix-preserving IPv4/IPv6 network address anonymization.","m":"agent draft (gemini)"},{"id":"e1987","i":"zhao-2020--CITES-->lin2016logcluster","s":"zhao-2020","t":"lin2016logcluster","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/infocom41043.2020.9155219 -> 10.1145/2889160.2889232)","m":"reference-list extractor"},{"id":"e1988","i":"zhao-2020--CITES-->rousseeuw1987silhouettes","s":"zhao-2020","t":"rousseeuw1987silhouettes","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/infocom41043.2020.9155219 -> 10.1016/0377-0427(87)90125-7)","m":"reference-list extractor"},{"id":"e1989","i":"zhao-2020--EXTENDS-->zhao2020alertstorm","s":"zhao-2020","t":"zhao2020alertstorm","r":"EXTENDS","w":2,"why":"Severity ranking of individual alerts after alert-storm handling","m":"agent draft (gemini)"},{"id":"e1990","i":"zhao-2021--CITES-->chen2019incidenttriage","s":"zhao-2021","t":"chen2019incidenttriage","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3468264.3468543 -> 10.1109/icse-seip.2019.00020)","m":"reference-list extractor"},{"id":"e1991","i":"zhao-2021--CITES-->chen2020incidental","s":"zhao-2021","t":"chen2020incidental","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3468264.3468543 -> 10.1145/3324884.3416624)","m":"reference-list extractor"},{"id":"e1992","i":"zhao-2021--CITES-->chen2020lidar","s":"zhao-2021","t":"chen2020lidar","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3468264.3468543 -> 10.1145/3368089.3417055)","m":"reference-list extractor"},{"id":"e1993","i":"zhao-2021--CITES-->fu2009execution","s":"zhao-2021","t":"fu2009execution","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3468264.3468543 -> 10.1109/icdm.2009.60)","m":"reference-list extractor"},{"id":"e1994","i":"zhao-2021--CITES-->gunawi2016cloudstop","s":"zhao-2021","t":"gunawi2016cloudstop","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3468264.3468543 -> 10.1145/2987550.2987583)","m":"reference-list extractor"},{"id":"e1995","i":"zhao-2021--CITES-->he2016experience","s":"zhao-2021","t":"he2016experience","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3468264.3468543 -> 10.1109/issre.2016.21)","m":"reference-list extractor"},{"id":"e1996","i":"zhao-2021--CITES-->he2017drain","s":"zhao-2021","t":"he2017drain","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3468264.3468543 -> 10.1109/icws.2017.13)","m":"reference-list extractor"},{"id":"e1997","i":"zhao-2021--CITES-->lin2016logcluster","s":"zhao-2021","t":"lin2016logcluster","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3468264.3468543 -> 10.1145/2889160.2889232)","m":"reference-list extractor"},{"id":"e1999","i":"zhao-2021--CITES-->zhao-2020","s":"zhao-2021","t":"zhao-2020","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3468264.3468543 -> 10.1109/infocom41043.2020.9155219)","m":"reference-list extractor"},{"id":"e2001","i":"zhao-2021--CITES-->zhu2019tools","s":"zhao-2021","t":"zhu2019tools","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3468264.3468543 -> 10.1109/icse-seip.2019.00021)","m":"reference-list extractor"},{"id":"e2002","i":"zhao-2021--EXTENDS-->wu2023changeinduced","s":"wu2023changeinduced","t":"zhao-2021","r":"EXTENDS","w":2,"why":"Operationalises change-induced incident detection with multimodal anomaly detection per change","m":"agent draft (gemini)"},{"id":"e2005","i":"zhao2020alertstorm--OPERATIONALIZES-->beyer2018workbook","s":"zhao2020alertstorm","t":"beyer2018workbook","r":"OPERATIONALIZES","w":2,"why":"An empirical alert-storm study that motivates the workbook's advice to alert on symptoms and SLO burn rather than on every cause.","m":"agent draft (claude)"},{"id":"e2010","i":"zhou-2024--COMPLEMENTS-->tang2024minicheck","s":"zhou2024dbot","t":"tang2024minicheck","r":"COMPLEMENTS","w":2,"why":"Pairs D-Bot's tree-search database root-cause diagnosis with MiniCheck's grounded claim-entailment verification.","m":"agent draft (gemini)"},{"id":"e2012","i":"zhou2024dbot--CITES-->dayan-2018","s":"zhou2024dbot","t":"dayan-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3643851 -> 10.1145/3183713.3196927)","m":"reference-list extractor"},{"id":"e2013","i":"zhou2024dbot--CITES-->huang2020tidb","s":"zhou2024dbot","t":"huang2020tidb","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3643851 -> 10.14778/3415478.3415535)","m":"reference-list extractor"},{"id":"e2014","i":"zhou2024dbot--CITES-->kanev-2015","s":"zhou2024dbot","t":"kanev-2015","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3620666.3651350 -> 10.1145/2749469.2750392)","m":"reference-list extractor"},{"id":"e2015","i":"zhou2024dbot--CITES-->panchenko-2019","s":"zhou2024dbot","t":"panchenko-2019","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3620666.3651350 -> 10.1109/cgo.2019.8661201)","m":"reference-list extractor"},{"id":"e2016","i":"zhou2024dbot--CITES-->raju-2017","s":"zhou2024dbot","t":"raju-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3643851 -> 10.1145/3132747.3132765)","m":"reference-list extractor"},{"id":"e2017","i":"zhou2024dbot--CITES-->shen-2023","s":"zhou2024dbot","t":"shen-2023","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3620666.3651350 -> 10.1145/3575693.3575727)","m":"reference-list extractor"},{"id":"e2018","i":"zhou2024dbot--CITES-->yasin-2014","s":"zhou2024dbot","t":"yasin-2014","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3620666.3651350 -> 10.1109/ispass.2014.6844459)","m":"reference-list extractor"},{"id":"e2019","i":"zhou2024dbot--COMPLEMENTS-->leijen2019mimalloc_aplas","s":"zhou2024dbot","t":"leijen2019mimalloc_aplas","r":"COMPLEMENTS","w":2,"why":"Characterizes TCMalloc span/hugepage fragmentation and dTLB stalls at warehouse scale across Spanner/Bigtable","m":"agent draft (gemini)"},{"id":"e2020","i":"zhou2024dbot--EXTENDS-->zhou2023calcspar","s":"zhou2024dbot","t":"zhou2023calcspar","r":"EXTENDS","w":2,"why":"Journal extension of Calcspar.","m":"agent draft (claude)"},{"id":"e2021","i":"zhu-2019--CITES-->ding-2014","s":"zhu2019tools","t":"ding-2014","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse-seip.2019.00021 -> 10.1109/dsn.2014.39)","m":"reference-list extractor"},{"id":"e2022","i":"zhu-2019--CITES-->du2016spell","s":"zhu2019tools","t":"du2016spell","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse-seip.2019.00021 -> 10.1109/icdm.2016.0103)","m":"reference-list extractor"},{"id":"e2023","i":"zhu-2019--CITES-->fu2009execution","s":"zhu2019tools","t":"fu2009execution","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse-seip.2019.00021 -> 10.1109/icdm.2009.60)","m":"reference-list extractor"},{"id":"e2025","i":"zhu-2019--CITES-->he-2018","s":"zhu2019tools","t":"he-2018","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse-seip.2019.00021 -> 10.1145/3236024.3236083)","m":"reference-list extractor"},{"id":"e2026","i":"zhu-2019--CITES-->he2016experience","s":"zhu2019tools","t":"he2016experience","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse-seip.2019.00021 -> 10.1109/issre.2016.21)","m":"reference-list extractor"},{"id":"e2027","i":"zhu-2019--CITES-->lin2016logcluster","s":"zhu2019tools","t":"lin2016logcluster","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse-seip.2019.00021 -> 10.1145/2889160.2889232)","m":"reference-list extractor"},{"id":"e2028","i":"zhu-2019--CITES-->nagaraj2012distalyzer","s":"zhu2019tools","t":"nagaraj2012distalyzer","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse-seip.2019.00021 -> nagaraj2012distalyzer)","m":"reference-list extractor"},{"id":"e2029","i":"zhu-2019--CITES-->oliner2007supercomputers","s":"zhu2019tools","t":"oliner2007supercomputers","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse-seip.2019.00021 -> 10.1109/dsn.2007.103)","m":"reference-list extractor"},{"id":"e2031","i":"zhu-2019--CITES-->yuan2010sherlog","s":"zhu2019tools","t":"yuan2010sherlog","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse-seip.2019.00021 -> 10.1145/1736020.1736038)","m":"reference-list extractor"},{"id":"e2032","i":"zhu-2019--CITES-->yuan2011logenhancer","s":"zhu2019tools","t":"yuan2011logenhancer","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse-seip.2019.00021 -> 10.1145/1950365.1950369)","m":"reference-list extractor"},{"id":"e2033","i":"zhu-2019--CITES-->yuan2012characterizing","s":"zhu2019tools","t":"yuan2012characterizing","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1109/icse-seip.2019.00021 -> 10.1109/icse.2012.6227202)","m":"reference-list extractor"},{"id":"e2034","i":"zhu2019tools--SHARES_BENCHMARK-->he2017drain","s":"zhu2019tools","t":"he2017drain","r":"SHARES_BENCHMARK","w":2,"why":"Establishes the 16-dataset Loghub-2k benchmark and Grouping Accuracy (GA) evaluation suite where Drain achieved top overall accuracy among 13 unsupervised parsers.","m":"agent draft (gemini)"},{"id":"e2035","i":"zhu2023loghub--EXTENDS-->oliner2007supercomputers","s":"zhu2023loghub","t":"oliner2007supercomputers","r":"EXTENDS","w":2,"why":"Loghub curates and redistributes the classic BGL and Thunderbird supercomputer logs from Oliner & Stearley (DSN 2007) alongside 15 additional distributed, OS, mobile, and server log datasets.","m":"agent draft (gemini)"},{"id":"e2037","i":"ziegler-2024--CITES-->forsgren-2021","s":"ziegler-2024","t":"forsgren-2021","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3633453 -> 10.1145/3454122.3454124)","m":"reference-list extractor"},{"id":"e2038","i":"ziems-2024--CITES-->gilardi2023chatgpt","s":"ziems-2024","t":"gilardi2023chatgpt","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1162/coli_a_00502 -> 10.1073/pnas.2305016120)","m":"reference-list extractor"},{"id":"e2041","i":"ziems-2024--CITES-->zhao2021calibrate","s":"ziems-2024","t":"zhao2021calibrate","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1162/coli_a_00502 -> zhao2021calibrate)","m":"reference-list extractor"},{"id":"e2043","i":"zong-2022--CITES-->satyanarayan-2017","s":"zong-2022","t":"satyanarayan-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1111/cgf.14519 -> 10.1109/tvcg.2016.2599030)","m":"reference-list extractor"},{"id":"e2044","i":"zong-2024--CITES-->munzner2014vad","s":"zong-2024","t":"munzner2014vad","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3613904.3641996 -> 10.1201/b17511)","m":"reference-list extractor"},{"id":"e2045","i":"zong-2024--CITES-->satyanarayan-2017","s":"zong-2024","t":"satyanarayan-2017","r":"CITES","w":2.5,"why":"Extracted outgoing reference (10.1145/3613904.3641996 -> 10.1109/tvcg.2016.2599030)","m":"reference-list extractor"},{"id":"e2046","i":"zong-2024--EXTENDS-->zong-2022","s":"zong-2024","t":"zong-2022","r":"EXTENDS","w":2,"why":"multi-modal (visual/text/sonification) structured editing","m":"agent draft (claude)"}],"limits":{"cites_rule":"The drawn dataset also holds 661 mechanically extracted citation edges (from 1413 publisher reference-list links in the vault); they are hidden by default and can be switched on in paper views.","authored_note":"13 of 23 canonical authored records have no curated relation yet; the authored index lists all of them, drawn or not.","edge_note":"Relation notes are agent-drafted (drafting model on every edge) and reviewed in the study pipelines. Nothing on this page is labelled human-curated. Peer-reviewed status is only shown where a resolvable external identifier exists."}}