{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp09_archetype_breadth150_20260804","research_id":"eoa_inverse_innovation_exp09_light_prior_art_20260804","cell_id":"backcasting_pathway_design__computer_science","search_lanes":{"direct_problem_and_intervention":{"queries":["computational reproducibility clean room replay immutable dependencies provenance workflow official standard","backcasting reproducible research computational workflow provenance"],"source_ids":["SRC1","SRC2","SRC3"],"no_result_note":null},"synonyms_and_historical_terms":{"queries":["reproducible computational experiment capture environment inputs random seeds external services ReproZip","CDE CARE reproducible execution package capture dependencies original paper"],"source_ids":["SRC1","SRC2"],"no_result_note":null},"products_practices_and_standards":{"queries":["Workflow Run RO-Crate specification official profile provenance inputs outputs code","site:acm.org artifact review badging reproducibility official computational artifacts"],"source_ids":["SRC2","SRC3"],"no_result_note":null},"component_combination":{"queries":["backcasting software systems roadmap future state prerequisites dependency mapping","backcasting reproducible research computational workflow provenance","\"backcasting\" \"software\" roadmap dependency"],"source_ids":["SRC3","SRC4"],"no_result_note":"No retained source directly combined workflow-run provenance and independent replay with a backcasting workshop, dependency board, and present-commitment register."}},"sources":[{"source_id":"SRC1","title":"Reproducibility and Replicability in Science — Appendix C: Recommendations Grouped by Stakeholder","publisher":"National Academies of Sciences, Engineering, and Medicine / National Academies Press","url":"https://www.nationalacademies.org/read/25303/chapter/14","source_type":"OFFICIAL_GUIDANCE","claims_supported":["Computational reproducibility requires complete information about inputs, methods, parameters, operating environment, architecture, and library dependencies.","Nondeterministic steps may require retained intermediate or output data, and bitwise reproducibility is not always reasonable.","Nonpublic-data policies can legitimately restrict what is retained or disclosed."]},{"source_id":"SRC2","title":"ReproZip! The Reproducibility Packer","publisher":"NYU VIDA","url":"https://www.reprozip.org/","source_type":"FIRST_PARTY_PRODUCT","claims_supported":["ReproZip captures data files, libraries, environment variables, and options in a self-contained package.","A captured package can reconstruct an execution environment on another machine or operating system through sandboxing or virtualization backends.","Automatic boundary-time capture and independent re-execution are established product capabilities rather than novel components."]},{"source_id":"SRC3","title":"Recording provenance of workflow runs with RO-Crate","publisher":"PLOS ONE","url":"https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0309210","source_type":"PRIMARY_RESEARCH","claims_supported":["Workflow Run RO-Crate packages prospective and retrospective provenance, inputs, outputs, code, parameters, and configuration with support for re-execution.","The model was implemented in six workflow-management systems and evaluated through use cases.","Full re-execution is not always possible; it depends on workflow formalism and captured execution-environment details, and containerization can help.","The work discusses authorization-controlled workflow execution for sensitive data through Trusted Research Environment extensions."]},{"source_id":"SRC4","title":"The Futures Toolkit — Backcasting","publisher":"UK Government Office for Science","url":"https://www.gov.uk/government/publications/futures-toolkit-for-policy-makers-and-analysts/the-futures-toolkit-html","source_type":"OFFICIAL_GUIDANCE","claims_supported":["Backcasting starts from a preferred future and works backward to identify enabling events and present actions.","The method explicitly asks what had to happen beforehand, identifies critical events, and distinguishes controllable from externally controlled changes.","Backcasting workshops, stakeholder participation, feasibility testing, and conversion of results into an action plan are established planning practices."]}],"problem_evidence":{"status":"SUPPORTED","finding":"The problem is visible: authoritative guidance identifies inputs, parameters, software dependencies, execution environments, nondeterminism, and access restrictions as prerequisites for computational reproducibility; established packaging and provenance systems exist specifically because incomplete or drifting execution state obstructs independent re-execution. The sources do not establish prevalence on the proposed platform, but they directly support the stated failure mechanism.","source_ids":["SRC1","SRC2","SRC3"]},"closest_prior_art":[{"name":"Workflow Run RO-Crate and its workflow-management-system implementations","source_ids":["SRC3"],"overlap":"Captures workflow-run provenance, code, inputs, outputs, parameters, configuration, and some environment information in portable packages designed to support independent re-execution and cross-run comparison.","remaining_difference":"It specifies provenance representation and implementations, not a future-to-present workshop that derives capture, retention, authorization, and ownership prerequisites before selecting near-term platform commitments."},{"name":"ReproZip","source_ids":["SRC2"],"overlap":"Automatically captures execution dependencies and data into a portable bundle that can be reconstructed and run on another machine in container or virtualized environments.","remaining_difference":"It is a capture-and-replay product, not a governance-aware method for backcasting from a replay acceptance test, classifying missing prerequisites in historical records, or ordering organizational commitments."},{"name":"Government Futures Toolkit backcasting method","source_ids":["SRC4"],"overlap":"Uses a preferred future, backward prerequisite questions, milestone ordering, stakeholder participation, control boundaries, feasibility checks, and an action plan—the central planning logic of the proposal.","remaining_difference":"It is domain-general and does not supply the clean-room compute-replay endpoint, provenance schema, immutable artifact requirements, or blinded replay test."},{"name":"National Academies computational-reproducibility recommendations","source_ids":["SRC1"],"overlap":"Provides a requirements-oriented account of the data, methods, parameters, environments, dependencies, uncertainty, and policy limitations needed for reproducibility.","remaining_difference":"It does not derive those requirements through a platform-specific reverse dependency map or compare resulting commitments with an existing forward telemetry backlog."}],"prior_art_disposition":"ADJACENT_PRIOR_ART","contrastive_claim_remaining":"For a bounded scientific-computing workflow class, deriving capture and governance commitments backward from a predeclared independent clean-room replay test will identify at least one necessary, correctly ordered capture, retention, authorization, or ownership prerequisite that an equally resourced forward provenance backlog misses, and this will improve prerequisite coverage in subsequent blinded replay attempts. This is an application-and-effect claim; the replay technologies and backcasting method themselves are established.","contrastive_claim_falsifier":"Pre-register replay criteria and compare a backcasting workshop with an equally resourced forward-backlog exercise using the same workflow class and equivalent expertise. The claim is falsified if the backcast yields no necessary prerequisite absent from the forward backlog, yields no more accurate prerequisite ordering, or fails to improve blinded clean-room replay coverage without increasing unauthorized retention.","gates":{"adequate_source_search":{"status":"PASS","rationale":"The bounded search covered direct formulations, older capture-and-package terminology, products and provenance specifications, backcasting practice, and component combinations. Exactly four opened sources span four publishers and include official guidance, primary research, and a first-party product.","source_ids":["SRC1","SRC2","SRC3","SRC4"]},"supported_problem":{"status":"PASS","rationale":"Independent sources directly identify mutable or missing data, parameters, dependencies, environments, nondeterminism, and access restrictions as barriers to computational re-execution.","source_ids":["SRC1","SRC2","SRC3"]},"distinct_testable_claim":{"status":"PASS","rationale":"Although both technical replay and backcasting are established, the remaining comparative claim is narrow and measurable: whether reverse derivation discovers necessary prerequisites or ordering links missed by a matched forward backlog and improves blinded replay coverage.","source_ids":["SRC1","SRC2","SRC3","SRC4"]},"bounded_next_test":{"status":"PASS","rationale":"A two-week test using one synthetic job and up to five consenting low-sensitivity jobs from one workflow class is bounded, reversible, and capable of recording predicted prerequisites, stopping points, authorization status, and replay outcomes. It cannot establish general effectiveness but can falsify the proposed local mechanism.","source_ids":["SRC1","SRC2","SRC3","SRC4"]},"no_obvious_safety_or_authority_stop":{"status":"PASS","rationale":"The proposed first step is limited to consenting low-sensitivity workflows, shadow capture, and isolated replay; it reserves retention and access decisions to the relevant owners and includes halt and quarantine procedures. Existing guidance confirms that policy restrictions, external control, and sensitive-data authorization must remain explicit. No obvious stop applies if those limits are enforced.","source_ids":["SRC1","SRC3","SRC4"]}},"screen_survival":true,"world_novelty_boundary":"This bounded public-web screen found adjacent prior art but no opened source that directly integrated clean-room workflow replay with the proposal's complete backcasting-and-commitment process. That phrase-level absence does not establish world novelty, patentability, market size, expert acceptance, realized value, or absence of unindexed, proprietary, or differently described prior art."}