{"schema_version":1,"research_id":"eoa_inverse_innovation_exp05_external_evaluation_20260803","source_assessment_id":"layer_decay_and_expiration_management__physics:P3:v0","cell_id":"layer_decay_and_expiration_management__physics","search_queries":["CERN ATLAS conditions database calibration validity interval global tag COOL official","CMS conditions database interval of validity global tag calibration official","Belle II conditions database payload validity interval global tag documentation","particle physics data preservation conditions database reproducibility calibration payload archive restore","site:cds.cern.ch conditions database calibration payload validity gaps overlaps global tag","site:software.belle2.org conditions database invalidated globaltag validation state payload official","HEP conditions database calibration payload reproducibility long term preservation software validation","CREST conditions database payload interval validity global tag official documentation","site:bls.gov OEWS database architects physicists median annual wage 2025","site:bls.gov Occupational Outlook Handbook physicists database administrators architects pay"],"sources":[{"source_id":"S1","title":"Conditions Database — basf2 Development Documentation","publisher":"Belle II Collaboration","url":"https://software.belle2.org/development/sphinx/framework/doc/conditions-database.html","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"n.d.","accessed_at":"2026-08-03","claims_supported":["Belle II stores immutable calibration payload revisions with run-based intervals of validity and global tags.","Global tags may contain gaps or overlaps and may be invalidated.","Only fixed global tags may be used for analysis; global-tag replay preserves the conditions used to create existing files.","Multiple selected global tags are searched by priority, demonstrating operational fallback/resolution semantics."]},{"source_id":"S2","title":"b2conditionsdb: Conditions DB Interface — basf2 Release 05-01-25","publisher":"Belle II Collaboration","url":"https://software.belle2.org/sphinx/release-05-01-25/framework/doc/tools/06-b2conditionsdb.html","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"n.d.","accessed_at":"2026-08-03","claims_supported":["Existing tooling inventories intervals of validity, downloads payload metadata and files, and supports local SQLite representations.","Invalidating a global tag disqualifies it from user analysis without implying destruction of its payload bytes.","Global-tag merge tooling detects overlaps, fills gaps by priority, and supports dry runs.","The tag-state workflow already separates OPEN, TESTING, VALIDATED, PUBLISHED, RUNNING, and terminal INVALID states."]},{"source_id":"S3","title":"HEP Software Foundation Community White Paper Working Group — Conditions Data","publisher":"HEP Software Foundation","url":"https://arxiv.org/abs/1901.05429","source_class":"PRIMARY_RESEARCH","publication_date":"2019-01-16","accessed_at":"2026-08-03","claims_supported":["Calibration and alignment conditions are required to reconstruct detector events correctly.","Versioned conditions databases are established practice for permitting improvements while preserving reproducibility.","Conditions access is critical, complex, and required at large distributed-computing scale.","Intervals of validity and global-tag-like version selection are established conditions-database concepts rather than novel elements."]},{"source_id":"S4","title":"Evolving ATLAS Conditions Data Architecture for LHC Runs 3 and 4","publisher":"ATLAS Collaboration, CERN","url":"https://atlas.cern/Updates/Briefing/Conditions-Data","source_class":"OFFICIAL_ORGANIZATION_DATA","publication_date":"2023-06-22","accessed_at":"2026-08-03","claims_supported":["ATLAS identifies conditions data as essential to collision reconstruction.","ATLAS reported cumbersome management across more than 10,000 tables, software-support risk, replication complexity, and degraded service for some workflows.","ATLAS committed substantial effort to reorganize its conditions architecture and is developing CREST with CMS.","Filesystem portability and backend independence are explicitly connected to preservation and long-term evolution."]},{"source_id":"S5","title":"Data Preservation in High Energy Physics","publisher":"DPHEP Collaboration / European Physical Journal C","url":"https://arxiv.org/abs/2302.03583","source_class":"PRIMARY_RESEARCH","publication_date":"2023-09-09","accessed_at":"2026-08-03","claims_supported":["HEP data preservation encompasses data, metadata, software, databases, and documentation, not bytes alone.","Long-term analysis capability and reproducibility are expressed community needs.","Preservation requires inventory, governance, identified resources, access policies, validation, and decisions about what can be dismissed.","Preserving reconstruction software and raw data is an established high-complexity preservation level."]},{"source_id":"S6","title":"A Validation Framework for the Long Term Preservation of High Energy Physics Data","publisher":"DPHEP","url":"https://arxiv.org/abs/1310.7814","source_class":"PRIMARY_RESEARCH","publication_date":"2013-10-29","accessed_at":"2026-08-03","claims_supported":["Automated testing of preserved data and software against evolving environments is established prior art.","Executable validation using controlled computing environments and external dependencies predates the proposal.","Archive presence alone is insufficient to establish continued usability; compatibility must be exercised."]},{"source_id":"S7","title":"The HEASARC Calibration Database","publisher":"NASA Goddard Space Flight Center, HEASARC","url":"https://heasarc.gsfc.nasa.gov/docs/heasarc/caldb/","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"2025-04-21","accessed_at":"2026-08-03","claims_supported":["NASA operates an identifiable calibration-data service and names a CALDB manager.","The service stores and indexes instrument-calibration datasets and tells users and software which calibration data should be used for reduction and analysis.","Multiple active and past missions use CALDB, demonstrating cross-mission adoptability and recurring calibration updates."]},{"source_id":"S8","title":"Database Administrators and Architects — Occupational Outlook Handbook","publisher":"U.S. Bureau of Labor Statistics","url":"https://www.bls.gov/ooh/computer-and-information-technology/database-administrators.htm","source_class":"GOVERNMENT_OR_REGULATOR","publication_date":"2025-08-28","accessed_at":"2026-08-03","claims_supported":["The 2024 U.S. median annual wage was $104,620 for database administrators and $135,980 for database architects.","Relevant duties include database design, integration, security, backup, restore, testing modifications, and maintaining permissions.","These wages provide an external labor anchor for broad 2026 resource-equivalent estimates, but not an experiment-specific quote."]}],"problem_evidence":{"support":"MODERATE","rationale":"Conditions data are demonstrably essential to correct event reconstruction, version accumulation is intrinsic, Belle II documentation explicitly permits gaps and overlaps, and ATLAS reports substantial management and long-term-support burdens. Belle II also documents that earlier software did not prevent use of OPEN global tags. However, no relied-upon source measures the frequency or scientific effect of obsolete payloads silently resolving outside their validated detector context, so the proposal's specific failure prevalence remains unverified.","source_ids":["S1","S3","S4","S5"]},"stakeholder_evidence":{"support":"MODERATE","rationale":"Belle II and ATLAS operate conditions infrastructure and expose concrete tag-management workflows; NASA HEASARC names a CALDB manager and supports mission users; DPHEP expresses community demand for preservation, validation, governance, and resources. These are credible adopters and authorizers, but none publicly requests calendar-style authority leases or the complete proposed lifecycle controller.","source_ids":["S1","S2","S4","S5","S7"]},"prior_art":{"proximity":"SUBSTANTIAL_COLLISION","closest_analogues":[{"name":"Belle II global-tag and interval-of-validity lifecycle","similarity":"Very high: immutable payload revisions, run-bounded validity, fixed analysis tags, replay, gap/overlap handling, validation states, invalidation, prioritized resolution, metadata download, and dry-run tooling already exist.","remaining_difference":"No verified periodic calendar lease that independently expires default-resolution authority, no demonstrated dependency graph spanning derived datasets and analyses, and no documented end-to-end archive reconstruction drill tied to payload disposition.","source_ids":["S1","S2"]},{"name":"HEP conditions-database archetype","similarity":"High: conditions versioning, intervals of validity, global selection, scalable access, and reproducibility are established community practice.","remaining_difference":"The proposal adds a review-date lease and joins resolver authority to retention tiering, dependency-gated disposition, quarantine, and tombstones.","source_ids":["S3"]},{"name":"DPHEP preservation and executable validation","similarity":"High on the preservation side: inventory, governance, software and database preservation, controlled validation, and long-term usability tests are established.","remaining_difference":"These practices do not themselves expire a calibration payload's eligibility for default resolution or compare lease behavior against ordinary validity intervals.","source_ids":["S5","S6"]},{"name":"NASA HEASARC CALDB","similarity":"Moderate: an operational cross-mission calibration repository indexes revisions and supports software selection of calibration datasets for analysis.","remaining_difference":"The reviewed page does not establish authority leases, dependency-aware retirement, quarantine, tombstones, or archived-stack reconstruction tests.","source_ids":["S7"]}],"distinctive_claim_remaining":"Relative to strict intervals of validity plus immutable and invalidatable global tags, adding a calendar review-date lease on default-resolution authority—without expiring byte retention—will detect additional stale selections while causing zero false uncovered intervals, preserving exact manifest replay, and leaving dependency-bearing historical payloads reconstructable. This is contrastive and falsifiable, but its advantage has not been demonstrated.","confidence":"HIGH"},"implementation_evidence":{"support":"STRONG","rationale":"The constituent mechanisms are technically routine or already deployed: versioned payloads, validity intervals, tag-state transitions, invalidation, priority resolution, local metadata snapshots, dry runs, archive validation environments, and backup/restore. The main uncertainty is workflow integration and completeness of cross-system dependency manifests, not basic computability. No public source establishes statutory retention requirements; production authority would remain experiment-governed and must use existing authenticated approval controls.","source_ids":["S1","S2","S3","S5","S6","S7"]},"scores":{"meaningful_impact":{"score":4,"rationale":"Applying an incorrect calibration can undermine reconstruction, while loss of historical conditions can prevent reproducibility; incidence and effect size are not measured.","source_ids":["S3","S4","S5"]},"stakeholder_pull":{"score":3,"rationale":"Named operating organizations invest in conditions management and preservation, but expressed pull is for those broader capabilities rather than for authority leases specifically.","source_ids":["S4","S5","S7"]},"incremental_advantage":{"score":2,"rationale":"Validity intervals, tag invalidation, immutable replay, lifecycle states, dry runs, and preservation validation already cover most proposed functions; only the periodic authority lease and unified dependency-aware lifecycle remain incremental.","source_ids":["S1","S2","S3","S5","S6"]},"distinctiveness_plausibility":{"score":2,"rationale":"The separation of authority from physical retention is already approximated by invalidating a global tag while retaining immutable payloads. Calendar-triggered review is a plausible narrower distinction, not a demonstrated new system class.","source_ids":["S1","S2"]},"technical_implementability":{"score":4,"rationale":"Existing tools expose the needed metadata and state transitions, and preservation validation can run in isolated environments; dependency discovery across private analyses remains difficult.","source_ids":["S2","S6"]},"adoption_authority_feasibility":{"score":4,"rationale":"Conditions managers, collaboration coordinators, data stewards, and named calibration-service managers are identifiable, and tag-state changes already follow controlled workflows. Joint production approval would still need local policy.","source_ids":["S2","S4","S7"]},"evidence_readiness":{"score":3,"rationale":"A read-only replay is bounded and technically ready, but decisive evidence requires proprietary manifests, archived payloads, known anomalies, and experiment-controlled reconstruction environments.","source_ids":["S1","S2","S6"]},"safety_net_benefit":{"score":4,"rationale":"Shadow resolution, immutable historical tags, quarantine, and isolated restoration can prevent irreversible production changes, although quarantine may conflict with an experiment-specific destruction rule not found publicly.","source_ids":["S1","S2","S5","S6"]},"scalability":{"score":4,"rationale":"Conditions services already operate across distributed experiments and missions, but mapping analysis dependencies and preservation holds is organizationally intensive and may not scale automatically.","source_ids":["S3","S4","S7"]}},"score_confidence":"MODERATE","costs":{"first_evidence":{"band_2026_usd":"10K_TO_50K","scope":"Read-only shadow registry for one retired subsystem, at most two global tags and 500 payloads; resolver replay, seeded gap/overlap cases, audit of 30 dependency verdicts, and one reference-stack reconstruction.","confidence":"MODERATE","assumptions":["Existing database exports, manifests, archived stack, reconstruction release, and compute allocation are available without procurement.","Approximately 1.5–3 combined person-months are required across a calibration physicist, software maintainer, and database specialist.","A 2026 loaded labor equivalent of roughly $12,000–$18,000 per person-month is used, informed by BLS database-administrator and architect wages plus an assumed benefits/overhead allowance."],"source_ids":["S2","S6","S8"]},"initial_deployment_startup":{"band_2026_usd":"50K_TO_250K","scope":"Production-grade registry schema, authenticated read integrations, lease and successor metadata, resolver shadow service, dashboards, audit logging, preservation-hold workflow, and security review for one experiment.","confidence":"LOW","assumptions":["Approximately 4–12 combined person-months are required.","Existing conditions, identity, monitoring, and archival platforms are reused.","No bulk payload migration or new archival facility is included."],"source_ids":["S1","S2","S4","S8"]},"operational_launch":{"band_2026_usd":"50K_TO_250K","scope":"Historical inventory reconciliation, policy approval, staged backfill, validation against known campaigns, operator training, runbooks, launch monitoring, and first scheduled restore drill.","confidence":"LOW","assumptions":["Approximately 6–15 combined person-months are required across scientific, software, database, preservation, and governance roles.","Launch initially covers one detector subsystem and expands only after resolver parity is demonstrated.","Existing compute and archival storage absorb the bounded validation workload."],"source_ids":["S4","S5","S6","S8"]},"annual_recurring":{"band_2026_usd":"50K_TO_250K","scope":"Lease review, exception revalidation, dependency reconciliation, quarterly sampled restore tests, software maintenance, incident response, and audit reporting.","confidence":"LOW","assumptions":["Recurring effort is approximately 0.3–1.0 full-time-equivalent plus bounded reconstruction compute.","Payload bytes are small relative to event data, so specialist labor dominates resource-equivalent cost.","No experiment-specific staffing or storage quotation was publicly verified."],"source_ids":["S4","S5","S6","S8"]}},"verified_pipeline_gates":{"externally_supported_problem":{"status":"YES","reason":"Authoritative and primary sources establish that calibration conditions are essential, versioned, operationally complex, capable of gaps or overlaps, and preservation-sensitive. Specific stale-resolution prevalence remains a gap but does not negate the underlying problem.","source_ids":["S1","S3","S4","S5"]},"externally_credible_adopter_or_authorizer":{"status":"YES","reason":"Belle II and ATLAS operate relevant conditions systems, while NASA HEASARC operates a named calibration service with a named manager; these organizations contain credible technical and scientific authorizers.","source_ids":["S2","S4","S7"]},"distinct_testable_incremental_claim":{"status":"YES","reason":"The remaining claim compares calendar authority leases against strict validity intervals and invalidatable immutable tags on measurable stale catches, false expirations, replay parity, dependency recall, and restore success.","source_ids":["S1","S2","S3"]},"bounded_next_evidence_step":{"status":"YES","reason":"A read-only test on one retired subsystem, bounded tags, payloads, manifests, seeded faults, and one reference reconstruction has explicit comparators and falsifiers and makes no production changes.","source_ids":["S2","S6"]},"no_unresolved_safety_or_authority_stop":{"status":"YES","reason":"The first step can use exports, dry-run or shadow resolution, immutable reference manifests, and an isolated restoration environment. Production revocation or deletion remains excluded and would require local authorization.","source_ids":["S1","S2","S6"]},"credible_cost_scope_and_range":{"status":"YES","reason":"All four estimates state bounded scope, person-month assumptions, reused infrastructure, and a government wage anchor. Confidence remains low to moderate because local staffing, compute, and storage prices are unavailable.","source_ids":["S8"]}},"next_evidence_step":"With an experiment partner, select one retired subsystem, one bounded historical run interval, no more than two authoritative global tags, and at most 500 payloads. Export metadata and replay the resolver under three comparators: A) current strict interval-of-validity and immutable/invalidatable-tag behavior; B) the same baseline plus alert-only review dates; and C) proposed fail-closed authority leases. Seed known gaps, overlaps, superseded payloads, and stale-priority cases. Compare all selections with frozen reconstruction manifests; independently inspect 30 dependency verdicts; restore one archived stack into isolation and process the existing reference event set. Measure incremental stale selections detected, false expirations, false uncovered intervals, resolver parity, known-dependency recall, restoration fidelity, latency, and specialist hours. Falsify the intervention if the baseline has no actionable stale resolution and leases add no distinction, if any authoritative payload is falsely rejected, if a known dependency is missed, if frozen-manifest replay changes, or if the documented archived stack cannot be restored. Make no production tag, retention, migration, or deletion changes.","blocking_evidence":["No public measurement establishes how often stale calibration payloads currently resolve or alter reconstructed quantities.","No relied-upon source demonstrates that a calendar authority lease outperforms strict intervals of validity plus existing invalidation and immutable replay.","Complete global-tag, derived-dataset, private-analysis, external-copy, and preservation-package dependency manifests are not publicly available.","Experiment-specific retention, legal-hold, destruction, and joint-approval policies were not verified.","Archive-restoration success for an exact historical calibration stack was not observed.","Local staffing, compute, archival-storage, integration, and governance costs are unavailable; provided bands are resource-equivalent estimates rather than quotations."],"research_disposition":"PARTNERED_RESEARCH_PROGRAM","world_novelty_boundary":"Only bounded open-web prior-art proximity was assessed. World novelty, patentability, freedom to operate, market size, realized impact, unpublished internal practice, proprietary systems, and unsearched patent literature remain unmeasured.","arm":"COMPLETE_PROPOSAL_PORTFOLIO","candidate_version":0,"controller_recommendation":{"action":"STOP_EMPIRICAL_RESEARCH_NEEDED","repairable":false,"material_progress_observed":true,"progress_targets":["Obtain an experiment partner with authority to provide a retired-subsystem metadata export, frozen manifests, known anomaly cases, and an isolated reconstruction environment.","Measure baseline stale-resolution prevalence and scientific relevance before attributing value to authority leases.","Demonstrate incremental detection over strict intervals of validity and tag invalidation with zero false uncovered intervals and exact frozen-manifest replay.","Measure dependency-verdict recall against registered consumers and characterize untraceable private or external analyses.","Complete one end-to-end archived-stack reconstruction and record fidelity, latency, failures, and specialist effort.","Replace labor-only estimates with observed person-hours, compute use, storage use, and local governance effort."],"reason":"Open-book research verifies a meaningful conditions-management and preservation problem, credible operators, and strong technical feasibility, but also finds substantial collision with existing validity intervals, immutable global tags, tag invalidation, dry-run tooling, and preservation validation. The remaining value proposition depends on proprietary manifests and live shadow replay in an experiment-controlled environment; bounded web research cannot establish prevalence or incremental effect."},"proposal_index":3}