{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp04_retrieval_first_paired20_20260802","cell_id":"layer_decay_and_expiration_management__film_media_production","arm":"RETRIEVAL_FIRST","round_index":0,"hypotheses":[{"hypothesis_id":"H1","title":"Expiring editorial decision layers","problem":"Obsolete notes, review links, and cut instructions remain visible after a sequence changes.","affected_stakeholder":"Picture editors and assistant editors","workflow_boundary":"Director and producer review through picture lock","failure_mode":"A stale instruction is mistaken for current intent, causing contradictory edits or rework.","unit_of_analysis":"Versioned editorial note linked to a sequence and cut","causal_lever":"Assign each note a lifecycle state and expiry trigger, with renewal or supersession at every cut revision.","archetype_mapping":"Treat accumulated review artifacts as layers that decay when their referenced cut changes; demote expired notes while preserving decision lineage.","expected_value":"Less time reconciling contradictory feedback and fewer edits made against superseded decisions.","falsifiable_claim":"Compared with an ungoverned notes workflow, lifecycle-labeled notes reduce edits attributable to superseded feedback by at least 25% across three productions.","diversity_rationale":"Targets semantic staleness at a creative-decision boundary, using a note as the unit and supersession rather than storage pressure as the trigger.","mechanism_slugs":["stale_layer_detection_dashboard","time_to_live_ttl_policy","tombstone_or_deletion_marker"],"search_questions":["Do editing or review platforms already expire or supersede notes when their referenced cut changes?","How often does superseded feedback cause measurable editorial rework?","Which note dependencies must remain visible after expiration?"]},{"hypothesis_id":"H2","title":"Dependency-gated VFX version cleanup","problem":"VFX intermediates accumulate because teams cannot tell which versions still feed approved shots, conforms, or downstream renders.","affected_stakeholder":"VFX producers and pipeline technical directors","workflow_boundary":"Vendor delivery through final conform and mastering","failure_mode":"Cleanup is deferred indefinitely, or a seemingly obsolete asset is deleted while still referenced downstream.","unit_of_analysis":"VFX asset or render version and its inbound dependency set","causal_lever":"Permit quarantine only after live-reference tracing proves a version orphaned; size the rollback window by deletion impact.","archetype_mapping":"Apply orphan detection and reversible deletion to sequential VFX versions while retaining an auditable disposition record.","expected_value":"Lower active-storage burden without increasing broken renders or emergency asset reconstruction.","falsifiable_claim":"A dependency-gated quarantine process can remove at least 20% more inactive VFX versions than manual cleanup with no increase in broken downstream jobs over one project.","diversity_rationale":"Targets dependency risk in a computational handoff, with graph nodes as units and proven orphanhood—not age—as the causal trigger.","mechanism_slugs":["dependency_safe_delete_check","soft_delete_quarantine_window","tombstone_or_deletion_marker"],"search_questions":["Do major VFX pipelines expose complete inbound references across vendors and conform systems?","What proportion of retained intermediates are true orphans?","Which dynamic or external dependencies evade automated tracing?"]},{"hypothesis_id":"H3","title":"Access-driven camera-media tiering","problem":"Original camera files, proxies, audio, and transcodes compete for limited high-speed storage throughout production.","affected_stakeholder":"Digital imaging technicians and post-production supervisors","workflow_boundary":"On-set ingest through offline edit and online relink","failure_mode":"Fast storage fills, prompting ad hoc moves that delay editorial access or separate media from verification metadata.","unit_of_analysis":"Verified media object’s residency in a storage tier","causal_lever":"Demote least-recently or least-frequently accessed verified media while pinning the current working set and preserving automatic recall.","archetype_mapping":"Translate cache eviction and lifecycle tiering into reversible hot-to-cold movement of deposited production media rather than deletion.","expected_value":"More usable high-speed capacity with bounded retrieval delay and preserved originals.","falsifiable_claim":"Access-driven tiering reduces peak fast-storage occupancy by at least 30% while keeping 95% of editorial media requests within the agreed retrieval target.","diversity_rationale":"Targets capacity pressure during ingest and editing, using tier residency as the unit and observed access—not validity or dependency—as the lever.","mechanism_slugs":["cache_eviction_rule","lifecycle_storage_tiering_policy","archive_restore_test"],"search_questions":["Which production storage products already tier camera media by access state?","What access patterns predict the editorial working set without cache pollution?","What retrieval latency is tolerable at offline, conform, and finishing stages?"]},{"hypothesis_id":"H4","title":"Rights-window lifecycle controls","problem":"Licenses, releases, and territorial permissions remain attached to media after their authorized use changes or expires.","affected_stakeholder":"Clearance producers and distribution operations teams","workflow_boundary":"Asset acquisition through platform and territory delivery","failure_mode":"An expired permission is treated as current, or evidence is destroyed despite a dispute or preservation duty.","unit_of_analysis":"Rights grant for one asset, use, territory, platform, and time window","causal_lever":"Map rights classes to activation and expiration dates, block unauthorized reuse, and override disposal for holds or continuing evidentiary value.","archetype_mapping":"Manage rights records as time-bounded layers whose operational authority expires separately from archival evidence.","expected_value":"Fewer unauthorized deliveries and faster verification of whether a planned use remains permitted.","falsifiable_claim":"Lifecycle-controlled rights records reduce deliveries later halted for expired or mismatched permissions by at least 40% relative to spreadsheet tracking.","diversity_rationale":"Targets legal validity at distribution, with a permission tuple as the unit and contractual dates plus holds as governing levers.","mechanism_slugs":["retention_schedule","time_to_live_ttl_policy","tombstone_or_deletion_marker"],"search_questions":["Do rights-management systems distinguish expiration of authority from deletion of evidence?","How often do expired or territory-mismatched grants disrupt distribution?","Which record classes require preservation after operational rights expire?"]},{"hypothesis_id":"H5","title":"Tested rehydratability for finished-program archives","problem":"Archived masters and project packages may become unreadable or irreconnectable as formats, credentials, and tools change.","affected_stakeholder":"Studio archivists and remastering producers","workflow_boundary":"Final delivery through long-term preservation and later re-release","failure_mode":"An archive appears intact until a restoration request reveals missing media, obsolete codecs, or broken project relationships.","unit_of_analysis":"Self-contained archived title package sampled by age and format cohort","causal_lever":"Periodically restore, parse, validate, and reconnect real packages; escalate failures into migration or preservation exceptions.","archetype_mapping":"Treat cold archival layers as retained but decaying assets whose recoverability must be revalidated rather than assumed.","expected_value":"Earlier detection of preservation failures and lower cost or delay when titles are remastered or reissued.","falsifiable_claim":"Scheduled end-to-end restore tests identify actionable archive defects at least one release cycle earlier than checksum-only monitoring in a one-year pilot.","diversity_rationale":"Targets latent recoverability failure after production, using whole title packages as units and exercised restoration—not cleanup—as the intervention.","mechanism_slugs":["archive_restore_test","age_weighted_value_score","retention_schedule"],"search_questions":["Which film archives already perform end-to-end project reconstruction rather than fixity checks alone?","What defect rate varies by package age, codec, storage medium, or authoring application?","How should sampling prioritize cultural value, replacement cost, and format obsolescence?"]}]}