{"schema_version":1,"research_id":"eoa_inverse_innovation_exp05_external_evaluation_20260803","source_assessment_id":"computability_boundary_mapping__film_media_production:P5:v0","cell_id":"computability_boundary_mapping__film_media_production","search_queries":["site:screenskills.com script supervisor continuity production role continuity errors","film television story bible continuity software canon checker recursive rules theorem prover","W3C OWL 2 profiles decidable reasoning finite rules recursion official","Datalog recursion termination finite domain official documentation","formal story world knowledge graph continuity checking film television canon research paper","narrative consistency checking story bible knowledge graph screenplay research","automated story consistency checker temporal logic narrative research paper","script supervisor continuity AI expressed need film production official","site:screenskills.com/job-profiles script supervisor continuity film TV role","site:screenskills.com \"script supervisor\" \"continuity\"","site:careers.bfi.org.uk script supervisor continuity","official film commission script supervisor continuity job role","script supervisor continuity role film official guild site","script supervisor continuity responsibilities film commission site","site:sagaftra.org script supervisor continuity","site:iatse.net script supervisor continuity"],"sources":[{"source_id":"S1","title":"Script Supervisor’s Department: Screen Guilds of Ireland Competency Framework","publisher":"Fís Éireann/Screen Ireland and Screen Guilds of Ireland","url":"https://www.screenireland.ie/images/uploads/general/Booklet_14_-_Script_Supervisors_Department_2024.pdf","source_class":"OFFICIAL_GUIDANCE","publication_date":"2022","accessed_at":"2026-08-03","claims_supported":["Script supervisors are responsible for script integrity and continuity and are hired by the production manager or line producer with director approval.","Their documented tasks include story-day and timeline breakdowns, identifying continuity elements, consulting the director about interpretation, recording script changes, and organizing continuity evidence.","This establishes a real workflow, a credible user, and identifiable authorizers, but it does not establish demand for formal recursive-canon reasoning."]},{"source_id":"S2","title":"Los Angeles Script Supervisors Network","publisher":"Los Angeles Script Supervisors Network","url":"https://lassn.org/","source_class":"OFFICIAL_ORGANIZATION_DATA","publication_date":"2026","accessed_at":"2026-08-03","claims_supported":["Script supervisors report to producers, work with directors, and supervise cumulative narrative developments including dialogue and plot points.","The network says continuity supervision saves production time, money, and quality and that supervisors flag narrative-continuity issues as scripts change.","The organization describes scripting as a specialized skill rather than merely software, supporting human authority and caution against automated displacement."]},{"source_id":"S3","title":"STAGE: A Full-Screenplay Benchmark for Reasoning over Evolving Stories","publisher":"arXiv, submitted by the research authors","url":"https://arxiv.org/abs/2601.08510","source_class":"PRIMARY_RESEARCH","publication_date":"2026-07-12","accessed_at":"2026-08-03","claims_supported":["Full screenplays contain complex relationships and temporally ordered events that require coherent story-world representations.","STAGE supplies curated knowledge graphs and annotations for 150 films and tests knowledge-graph construction, event verification, screenplay question answering, and character-consistent responses.","This is adjacent research infrastructure, not evidence of a production service promising total exact entailment over recursive canon rules."]},{"source_id":"S4","title":"Lost in Stories: Consistency Bugs in Long Story Generation by LLMs","publisher":"Association for Computational Linguistics","url":"https://aclanthology.org/2026.findings-acl.410.pdf","source_class":"PRIMARY_RESEARCH","publication_date":"2026-07-02","accessed_at":"2026-08-03","claims_supported":["The study reports factual, temporal, character, and world-rule inconsistencies in long-form generated narratives.","ConStory-Checker grounds contradiction judgments in textual evidence and achieved 0.678 F1 on an injected-error diagnostic dataset, showing both feasibility and substantial residual error.","The experiment concerns probabilistic contradiction detection in generated stories, not sound and complete formal canon entailment or production-script approval."]},{"source_id":"S5","title":"Story Bible","publisher":"The Story Distillery Pty Ltd","url":"https://storydistiller.com/help/bible/","source_class":"COMMERCIAL_FIRST_PARTY","publication_date":"2023","accessed_at":"2026-08-03","claims_supported":["A commercial screenwriting product already lets users create, store, categorize, and retrieve project-specific story rules and automatically adds some character or relationship details as bible entries.","The documented product is a structured rule repository; the page does not claim formal entailment, model checking, recursive inference, proof certificates, or exact three-way answers."]},{"source_id":"S6","title":"Telling Non-linear Stories with Interval Temporal Logic","publisher":"Bath Spa University ResearchSPAce; published in ICIDS 2015 proceedings","url":"https://researchspace.bathspa.ac.uk/8218/","source_class":"PRIMARY_RESEARCH","publication_date":"2015","accessed_at":"2026-08-03","claims_supported":["Prior research explicitly models narrative worlds as Kripke structures with interval temporal logic and model-checks possible tellings for consistency.","The authors identify exhaustive specification of narrative deviations as difficult.","This closely anticipates finite formal narrative-consistency checking, although not serialized-production canon governance with proof, bounded-countermodel, and UNKNOWN labels."]},{"source_id":"S7","title":"OWL 2 Web Ontology Language Profiles (Second Edition)","publisher":"World Wide Web Consortium","url":"https://www.w3.org/TR/owl2-profiles/","source_class":"STANDARD","publication_date":"2012-12-11","accessed_at":"2026-08-03","claims_supported":["OWL 2 profiles are syntactically restricted sublanguages that trade expressiveness for efficient reasoning.","OWL 2 RL supports rule-based implementations and specified reasoning tasks with polynomial-time bounds, while the standard reports undecidability for corresponding OWL 2 RDF-based reasoning problems.","This is established standards-level prior art for enforceable fragments and guarantee-sensitive reasoning, and it shows that merely requiring acyclic rules is only one possible design."]},{"source_id":"S8","title":"Acyclicity Notions for Existential Rules and Their Application to Query Answering in Ontologies","publisher":"arXiv, submitted by the research authors","url":"https://arxiv.org/abs/1406.4110","source_class":"PRIMARY_RESEARCH","publication_date":"2014-06-16","accessed_at":"2026-08-03","claims_supported":["For existential-rule reasoning, chase execution need not terminate and deciding chase termination for a given rule-and-fact set is undecidable.","The paper develops enforceable sufficient acyclicity conditions and reports practical feasibility for restricted ontology reasoning.","This substantially overlaps the proposal's restriction-and-fallback logic but does not validate its specific computation-to-canon reduction or story semantics."]}],"problem_evidence":{"support":"MODERATE","rationale":"Continuity and evolving narrative consistency are visible professional problems: official industry guidance assigns story timelines, script interpretation, and continuity evidence to a senior department, while recent research documents factual and temporal failures in long narratives. However, no opened source shows an actual serialized-production service accepting unrestricted recursive canon rules, promising a total exact REQUIRED/FORBIDDEN/OPEN answer, or converting timeout into OPEN. The broad problem exists and matters; the proposal's decisive computability-pathology premise remains unobserved.","source_ids":["S1","S2","S3","S4"]},"stakeholder_evidence":{"support":"MODERATE","rationale":"Script supervisors are identifiable users, production managers or line producers are hiring authorities, and directors approve the role and control interpretation. Industry organizations expressly connect continuity work to time, money, editorial coherence, and narrative integrity. No source expresses demand for formal proof certificates, computability-boundary review, or an UNKNOWN-bearing canon reasoner, and no production company or showrunner is documented as willing to fund or pilot it.","source_ids":["S1","S2"]},"prior_art":{"proximity":"SUBSTANTIAL_COLLISION","closest_analogues":[{"name":"Interval-temporal model checking of narrative worlds","similarity":"Directly formalizes a story world and model-checks possible narrative paths for consistency, covering the proposal's central finite-model narrative reasoning concept.","remaining_difference":"It does not document serialized canon operations, a halting reduction, independently checkable derivations, bounded-countermodel labels, operational UNKNOWN, or creative-approval governance.","source_ids":["S6"]},{"name":"OWL 2 profiles and rule-based reasoning","similarity":"Established standard practice restricts ontology syntax to obtain defined computational properties and sound/complete answers for specified query classes.","remaining_difference":"It is domain-general and does not provide the proposed canon grammar, story-specific semantics, routing labels, or production authority workflow.","source_ids":["S7"]},{"name":"Acyclic existential-rule reasoning","similarity":"Established research distinguishes potentially nonterminating rule reasoning from acyclic fragments with termination guarantees.","remaining_difference":"The restrictions and results concern ontology chase procedures, not the proposal's exact canon language or claimed program-to-canon preservation proof.","source_ids":["S8"]},{"name":"Story Distiller Story Bible","similarity":"A commercial screenwriting product already stores project-specific story rules and structured character or relationship entries.","remaining_difference":"The documented feature is retrieval and editing, not formal inference or guarantee-aware decision routing.","source_ids":["S5"]},{"name":"ConStory-Checker and STAGE","similarity":"Current systems and benchmarks construct story-world representations, reason over screenplays, and detect evidence-grounded narrative inconsistencies.","remaining_difference":"They provide empirical model outputs rather than sound, complete, terminating formal classifications and do not govern production authority.","source_ids":["S3","S4"]}],"distinctive_claim_remaining":"Conditional on a separately validated canon grammar and reduction, a serialized-production workflow can enforce exact three-way classification only inside a mechanically recognized finite fragment, while recursive theories return proof-backed REQUIRED or FORBIDDEN results when found and otherwise retain UNKNOWN and bounded-countermodel states without converting them into script permission. The falsifiable increment is operational label preservation and routing under stated semantics, not the novelty of model checking, language restriction, proof checking, or story bibles.","confidence":"HIGH"},"implementation_evidence":{"support":"MODERATE","rationale":"Standards and research support restricted rule languages, finite/model-relative reasoning, acyclicity checks, and narrative model checking. Recent screenplay and consistency datasets also support construction of synthetic tests. Major feasibility gaps remain: no independently reviewed computation-to-canon reduction exists; finite, total, and acyclic alone do not fully specify negation, contradiction, or open-world semantics; formalization fidelity and authoring effort are unmeasured; no production-system integration, confidentiality, rights, or data-retention assessment was found; and no evidence shows users will preserve UNKNOWN rather than treating it as permission.","source_ids":["S3","S4","S6","S7","S8"]},"scores":{"meaningful_impact":{"score":3,"rationale":"Continuity affects script integrity, editorial coherence, time, and cost, but the prevalence and consequence of the specific timeout-to-OPEN failure are unmeasured.","source_ids":["S1","S2","S4"]},"stakeholder_pull":{"score":2,"rationale":"Credible users and authorizers express a general continuity need, but no stakeholder requests, funds, or commits to this formal protocol.","source_ids":["S1","S2"]},"incremental_advantage":{"score":3,"rationale":"Proof-bearing labels and explicit UNKNOWN could improve auditability over rule repositories and probabilistic checkers, but advantage over a well-configured ontology reasoner plus human review has not been tested.","source_ids":["S4","S5","S7","S8"]},"distinctiveness_plausibility":{"score":2,"rationale":"Narrative model checking, syntactic reasoning profiles, acyclicity restrictions, and evidence-grounded consistency checking substantially cover the mechanism set; the remaining distinction is a domain workflow and label contract.","source_ids":["S4","S6","S7","S8"]},"technical_implementability":{"score":4,"rationale":"A synthetic finite evaluator, fragment recognizer, bounded proof scheduler, certificate checker, and router are technically credible, although story-semantic fidelity and the proposed reduction are unresolved.","source_ids":["S6","S7","S8"]},"adoption_authority_feasibility":{"score":3,"rationale":"Production managers, line producers, directors, and script supervisors form an identifiable authority chain, but showrunner or studio governance for serialized canon and procurement authority were not directly evidenced.","source_ids":["S1","S2"]},"evidence_readiness":{"score":3,"rationale":"Synthetic stories and recent benchmark structures make a bounded laboratory test ready, but decisive demand and workflow evidence requires access to practitioners or proprietary processes.","source_ids":["S3","S4","S6"]},"safety_net_benefit":{"score":4,"rationale":"Keeping UNKNOWN and bounded countermodels distinct from permission directly limits false confidence, provided interfaces and human authorities preserve those labels.","source_ids":["S1","S2","S7","S8"]},"scalability":{"score":2,"rationale":"Reasoning over enforced profiles can scale, but formalizing evolving canon, reviewing semantic changes, handling confidential unreleased material, and maintaining proofs could impose production-specific labor that has not been measured.","source_ids":["S1","S3","S7","S8"]}},"score_confidence":"MODERATE","costs":{"first_evidence":{"band_2026_usd":"50K_TO_250K","scope":"A six-to-twelve-week synthetic evaluation: freeze one grammar and semantics, independently review the reduction, implement minimal fragment checking, proof scheduling, certificate validation, and routing, and test 20 micro-canons against three comparators.","confidence":"MODERATE","assumptions":["Approximately 0.8-1.5 loaded engineer/researcher years in aggregate plus part-time formal-methods and continuity-domain review.","No active production data, production-system integration, procurement, or security certification.","Open-source reasoning components may be reused, but semantic adapters and test oracles are custom."],"source_ids":["S3","S6","S7","S8"]},"initial_deployment_startup":{"band_2026_usd":"250K_TO_1M","scope":"Build a production-oriented alpha for one consenting serialized project, including grammar tooling, provenance, access control, versioning, result-label UI, audit logs, and workflow integration.","confidence":"LOW","assumptions":["Four-to-six people for roughly six to nine months, combining formal reasoning, product engineering, security, and continuity expertise.","One production workflow and one canon format only.","Excludes licensing disputes, enterprise procurement delays, and migration of a large historical franchise bible."],"source_ids":["S1","S2","S5","S7"]},"operational_launch":{"band_2026_usd":"1M_TO_5M","scope":"Harden and launch across several productions or one major franchise, with independent assurance, incident handling, confidential-data controls, integrations, training, accessibility, and operational support.","confidence":"LOW","assumptions":["A multidisciplinary team of roughly six-to-twelve people for 12-18 months.","Formalization and migration require human review rather than automatic extraction alone.","Launch does not authorize automated script approval and retains human creative authority."],"source_ids":["S1","S2","S3","S4","S7"]},"annual_recurring":{"band_2026_usd":"250K_TO_1M","scope":"Operate a limited service for several productions: engineering maintenance, semantic/version review, security operations, user support, independent proof audits, and incident response.","confidence":"LOW","assumptions":["Two-to-five loaded full-time equivalents plus compute and storage.","UNKNOWN and formalization exceptions remain human-reviewed.","Major new language expressiveness or franchise-wide migration would be separately funded."],"source_ids":["S1","S2","S7","S8"]}},"verified_pipeline_gates":{"externally_supported_problem":{"status":"UNCERTAIN","reason":"Continuity burden and narrative-consistency errors are externally supported, but the defining failure—an unrestricted recursive production canon service coercing resource exhaustion into OPEN—was not found.","source_ids":["S1","S2","S3","S4"]},"externally_credible_adopter_or_authorizer":{"status":"YES","reason":"Script supervisors are credible users; production managers or line producers hire them with director approval, and they report to producers and collaborate with directors.","source_ids":["S1","S2"]},"distinct_testable_incremental_claim":{"status":"YES","reason":"The protocol can be tested for fragment-enforcement accuracy, certificate validity, proof-search fairness, exact finite classifications, and preservation of UNKNOWN and bounded-countermodel labels against explicit comparators.","source_ids":["S6","S7","S8"]},"bounded_next_evidence_step":{"status":"YES","reason":"A fixed grammar, independently reviewed reduction, 20 synthetic cases, named comparators, outcome metrics, and hard failure conditions can be completed without touching active canon.","source_ids":["S3","S4","S6","S7","S8"]},"no_unresolved_safety_or_authority_stop":{"status":"YES","reason":"The synthetic next step can exclude confidential scripts and operational decisions, retain director or producer authority, and halt on label collapse or invalid certificates. Production deployment would require a separate confidentiality and rights review.","source_ids":["S1","S2"]},"credible_cost_scope_and_range":{"status":"YES","reason":"The ranges are broad resource-equivalent estimates tied to explicit team size, duration, integration, assurance, and support assumptions, although no vendor quotes or wage benchmark was included and confidence is therefore low to moderate.","source_ids":["S1","S2","S7","S8"]}},"next_evidence_step":"Run a maximum 12-week, non-production study. Freeze one temporal canon grammar, open/closed-world and negation semantics, contradiction policy, fragment recognizer, and operational search bound. Have an independent formal-methods reviewer assess the program-input-to-canon translation for totality, well-formedness, step preservation, and the HALTED biconditional. Test 20 preregistered synthetic micro-canons using four comparators: manual continuity review, flat/closed-world fact lookup, an evidence-grounded LLM contradiction checker, and ordinary finite temporal model checking without the proposed router. Primary measures are exact-fragment classification accuracy, invalid-certificate acceptance, known-proof starvation, UNKNOWN preservation through the UI, formalization time, reviewer agreement, and output latency. Falsify the intervention if any seeded invalid proof is accepted, any complete finite case is misclassified, any known finite proof is starved under the declared fair schedule, resource exhaustion becomes OPEN or permission, fragment-invalid syntax reaches exact mode, or independent review rejects the reduction. In parallel, conduct consented retrospective workflow interviews with 6-10 script supervisors, story editors, producers, or continuity staff from at least two organizations, without collecting unreleased story content. Falsify the stated problem premise if none uses or is procuring rule-based canon automation, none has seen unresolved search represented definitively, and existing workflows already distinguish absence, timeout, ambiguity, and creative override.","blocking_evidence":["No direct evidence was found that a film or serialized-production canon service currently accepts unrestricted recursive rules and promises total exact three-way entailment.","No direct evidence was found that a deployed tool converts timeout, failed proof search, or a bounded countermodel into OPEN or script permission.","No production company, showrunner, producer, or continuity department has expressed willingness to adopt or fund the proposed protocol.","The computation-to-canon reduction and its semantic preservation obligations have not been independently checked.","The finite grammar, contradiction semantics, proof format, fairness schedule, and fragment recognizer have not been implemented or validated.","Formalization labor, UNKNOWN frequency, user comprehension, override behavior, and comparative decision quality are unmeasured.","Confidentiality, copyright, data-retention, and access-control requirements for unreleased story bibles and plot revelations have not been assessed with a production partner.","Cost ranges lack vendor quotes, measured prototype effort, and production-specific integration estimates."],"research_disposition":"PROBLEM_PREVALENCE_STUDY","world_novelty_boundary":"The search establishes substantial adjacent and colliding art, not world novelty. Narrative model checking, structured story bibles, screenplay knowledge graphs, evidence-grounded consistency checking, syntactically restricted ontology profiles, and acyclicity-based termination controls all predate or independently parallel the proposal. Whether the precise serialized-canon routing and result-label combination exists elsewhere, is patentable, has freedom to operate, has a measurable market, or produces realized impact remains unmeasured.","arm":"COMPLETE_PROPOSAL_PORTFOLIO","candidate_version":0,"controller_recommendation":{"action":"STOP_EMPIRICAL_RESEARCH_NEEDED","repairable":false,"material_progress_observed":true,"progress_targets":["Establish through consented fieldwork whether any real serialized-production workflow uses or is procuring formal or rule-based canon reasoning and whether incomplete search is ever rendered as a definitive continuity result.","Obtain documented pilot interest and authority boundaries from at least one production organization, including who may configure semantics and who may approve scripts.","Complete independent review of the computation-to-canon reduction and publish every assumption, failed obligation, and scope limitation.","Implement the bounded synthetic comparison and meet zero-tolerance falsifiers for invalid proofs, finite-case misclassification, proof starvation, fragment bypass, and UNKNOWN-to-permission collapse.","Measure formalization time, inter-reviewer agreement, UNKNOWN rate, user interpretation, false-clearance and false-block rates, and comparative performance against manual review, fact lookup, LLM checking, and ungoverned finite model checking.","Obtain production-specific security, confidentiality, rights, integration, and cost estimates before any live-canon pilot."],"reason":"Web evidence verifies that continuity matters, identifies credible users and authorizers, and shows that the technical mechanisms are largely established. It does not verify the proposal's defining deployed failure or stakeholder demand. Resolving those gaps requires practitioner fieldwork, proprietary workflow evidence, and live or prototype testing rather than further bounded web search; under the controller rule this requires an empirical-research stop with repairable set to false."},"proposal_index":5}