{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:UIQL5Z24HBMU6G5GWGGDZIUBNF","short_pith_number":"pith:UIQL5Z24","schema_version":"1.0","canonical_sha256":"a220bee75c38594f1ba6b18c3ca28169707037268d9907da628c1eace31207a9","source":{"kind":"arxiv","id":"2504.16777","version":1},"attestation_state":"computed","paper":{"title":"Systemic Flakiness: An Empirical Analysis of Co-Occurring Flaky Test Failures","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.SE","authors_text":"Gregory Kapfhammer, Michael Hilton, Owain Parry, Phil McMinn","submitted_at":"2025-04-23T14:51:23Z","abstract_excerpt":"Flaky tests produce inconsistent outcomes without code changes, creating major challenges for software developers. An industrial case study reported that developers spend 1.28% of their time repairing flaky tests at a monthly cost of $2,250. We discovered that flaky tests often exist in clusters, with co-occurring failures that share the same root causes, which we call systemic flakiness. This suggests that developers can reduce repair costs by addressing shared root causes, enabling them to fix multiple flaky tests at once rather than tackling them individually. This study represents an infle"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2504.16777","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2025-04-23T14:51:23Z","cross_cats_sorted":[],"title_canon_sha256":"a63444d499303ebbd7acc087503cd915aea65ba6516381dfb88a1ba9e0eaa39a","abstract_canon_sha256":"c03f8f4c26fa950d68e3d97a8ea872fd5034d516bbb97af49d284f8d548eaca3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:53:04.420399Z","signature_b64":"BmA0e0iKSaAcm3hsUrCLbRaBUm5mnxTfN35+oBMqc5kDHNvBUz7w6+lOC26pFXSZeQc4aktEgxBix7vKLJvrCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a220bee75c38594f1ba6b18c3ca28169707037268d9907da628c1eace31207a9","last_reissued_at":"2026-07-05T10:53:04.419928Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:53:04.419928Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Systemic Flakiness: An Empirical Analysis of Co-Occurring Flaky Test Failures","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.SE","authors_text":"Gregory Kapfhammer, Michael Hilton, Owain Parry, Phil McMinn","submitted_at":"2025-04-23T14:51:23Z","abstract_excerpt":"Flaky tests produce inconsistent outcomes without code changes, creating major challenges for software developers. An industrial case study reported that developers spend 1.28% of their time repairing flaky tests at a monthly cost of $2,250. We discovered that flaky tests often exist in clusters, with co-occurring failures that share the same root causes, which we call systemic flakiness. This suggests that developers can reduce repair costs by addressing shared root causes, enabling them to fix multiple flaky tests at once rather than tackling them individually. This study represents an infle"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.16777","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2504.16777/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2504.16777","created_at":"2026-07-05T10:53:04.419985+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.16777v1","created_at":"2026-07-05T10:53:04.419985+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.16777","created_at":"2026-07-05T10:53:04.419985+00:00"},{"alias_kind":"pith_short_12","alias_value":"UIQL5Z24HBMU","created_at":"2026-07-05T10:53:04.419985+00:00"},{"alias_kind":"pith_short_16","alias_value":"UIQL5Z24HBMU6G5G","created_at":"2026-07-05T10:53:04.419985+00:00"},{"alias_kind":"pith_short_8","alias_value":"UIQL5Z24","created_at":"2026-07-05T10:53:04.419985+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UIQL5Z24HBMU6G5GWGGDZIUBNF","json":"https://pith.science/pith/UIQL5Z24HBMU6G5GWGGDZIUBNF.json","graph_json":"https://pith.science/api/pith-number/UIQL5Z24HBMU6G5GWGGDZIUBNF/graph.json","events_json":"https://pith.science/api/pith-number/UIQL5Z24HBMU6G5GWGGDZIUBNF/events.json","paper":"https://pith.science/paper/UIQL5Z24"},"agent_actions":{"view_html":"https://pith.science/pith/UIQL5Z24HBMU6G5GWGGDZIUBNF","download_json":"https://pith.science/pith/UIQL5Z24HBMU6G5GWGGDZIUBNF.json","view_paper":"https://pith.science/paper/UIQL5Z24","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.16777&json=true","fetch_graph":"https://pith.science/api/pith-number/UIQL5Z24HBMU6G5GWGGDZIUBNF/graph.json","fetch_events":"https://pith.science/api/pith-number/UIQL5Z24HBMU6G5GWGGDZIUBNF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UIQL5Z24HBMU6G5GWGGDZIUBNF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UIQL5Z24HBMU6G5GWGGDZIUBNF/action/storage_attestation","attest_author":"https://pith.science/pith/UIQL5Z24HBMU6G5GWGGDZIUBNF/action/author_attestation","sign_citation":"https://pith.science/pith/UIQL5Z24HBMU6G5GWGGDZIUBNF/action/citation_signature","submit_replication":"https://pith.science/pith/UIQL5Z24HBMU6G5GWGGDZIUBNF/action/replication_record"}},"created_at":"2026-07-05T10:53:04.419985+00:00","updated_at":"2026-07-05T10:53:04.419985+00:00"}