{"schema_version":1,"research_id":"eoa_inverse_innovation_exp06_external_evaluation_20260803","source_assessment_id":"representation_independent_interface_contract__chemistry_materials:P2:v0","cell_id":"representation_independent_interface_contract__chemistry_materials","search_queries":["thermophysical property software interface standard CAPE-OPEN thermodynamic package specification property package","CoolProp AbstractState backends REFPROP interface derivatives phase reference state official documentation","NIST REFPROP DLL interface thermophysical properties official documentation equations errors","thermophysical property data standards ThermoML IUPAC official","site:coolprop.org/coolprop HighLevelAPI reference states derivatives phase impose units documentation","site:nist.gov/srd/refprop REFPROP uncertainty range of validity DLL official","IDAES thermophysical property package interface property metadata units official documentation","thermophysical property software uncertainty reference state interoperability comparison REFPROP CoolProp paper","NIST REFPROP 10 DLL documentation PDF error codes limits reference state official","site:pages.nist.gov REFPROP DLL documentation reference state units errors","CoolProp paper abstract open source thermophysical property library DOI 2014","CAPE OPEN adoption companies official members thermodynamic property packages expressed need interoperability"],"sources":[{"source_id":"S1","title":"Thermodynamics and Physical Properties Interface Specification v1.1","publisher":"CAPE-OPEN Laboratories Network","url":"https://www.colan.org/specifications/thermodynamics-and-physical-properties-interface-specification-v1-1/","source_class":"STANDARD","publication_date":"2011-05","accessed_at":"2026-08-03","claims_supported":["CAPE-OPEN already standardizes interfaces for thermodynamic software components used by process-modeling environments.","The specification defines expected behavior for property-package managers, material objects, equilibrium calls, and related edge cases.","The specification was prototyped as a plug-and-socket interface and subsequently clarified in response to implementation experience."]},{"source_id":"S2","title":"About CAPE-OPEN","publisher":"CAPE-OPEN Laboratories Network","url":"https://www.colan.org/general-information-on-co-lan/","source_class":"OFFICIAL_ORGANIZATION_DATA","publication_date":"undated","accessed_at":"2026-08-03","claims_supported":["Industrial operators and simulator vendors explicitly sought plug-and-play interoperability without bespoke interfaces.","Identifiable participants included Bayer, BASF, BP, DuPont, Aspen Technology, SimSci, and academic partners.","Users often need to substitute third-party thermodynamic models for native models, and CAPE-OPEN supplies maintained standard interfaces and use cases for that purpose."]},{"source_id":"S3","title":"Backends in CoolProp","publisher":"CoolProp Project","url":"https://coolprop.org/develop/backends.html","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"2026","accessed_at":"2026-08-03","claims_supported":["CoolProp's AbstractState already separates a public thermophysical-property interface from backend implementations.","Multiple implementations, including CoolProp Helmholtz, incompressible, REFPROP, and IF97 backends, implement the same abstract protocol.","A representation-independent backend interface is therefore existing product practice rather than a new general concept."]},{"source_id":"S4","title":"High-Level Interface — CoolProp 8.0.0 Documentation","publisher":"CoolProp Project","url":"https://coolprop.org/coolprop/HighLevelAPI.html","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"2026","accessed_at":"2026-08-03","claims_supported":["Thermophysical interfaces expose phase identification and first, second, and saturation derivatives.","Derivative validity depends on phase and query type; CoolProp warns that ordinary derivative formulations are invalid or undefined for many two-phase combinations.","The documented restrictions demonstrate that domain, phase, derivative, and error semantics are material parts of a usable property interface."]},{"source_id":"S5","title":"NIST Standard Reference Database 23: REFPROP Version 10","publisher":"National Institute of Standards and Technology","url":"https://www.nist.gov/srd/refprop","source_class":"GOVERNMENT_OR_REGULATOR","publication_date":"2018","accessed_at":"2026-08-03","claims_supported":["REFPROP is an authoritative thermophysical-property implementation that exposes calculations through source subroutines, a DLL, spreadsheets, and application integrations.","NIST instructs users to identify and cite the specific equation-of-state or transport model used, supporting the importance of model-level provenance.","REFPROP licensing distinguishes individual, site, intranet, and distributor uses, creating a licensing consideration for real integrations."]},{"source_id":"S6","title":"State Definition — IDAES Generic Property Package Framework","publisher":"IDAES Project, U.S. Department of Energy sponsored","url":"https://idaes-pse.readthedocs.io/en/stable/explanations/components/property_package/general/state_definition.html","source_class":"OFFICIAL_GUIDANCE","publication_date":"2026","accessed_at":"2026-08-03","claims_supported":["IDAES requires property packages to define state variables and supports standard state-definition modules.","IDAES says empirical thermophysical correlations are valid only over specific ranges and directs users to declare bounds and units.","IDAES requires explicit thermodynamic reference temperature and pressure because many thermophysical quantities are relative."]},{"source_id":"S7","title":"ThermoML","publisher":"National Institute of Standards and Technology","url":"https://www.nist.gov/mml/acmd/trc/thermoml","source_class":"STANDARD","publication_date":"2026-05-29","accessed_at":"2026-08-03","claims_supported":["ThermoML is an IUPAC-standard XML format for exchanging thermophysical and thermochemical data.","The format supports uncertainty, methods, constraints, bibliographic information, and machine-to-machine communication.","ThermoML standardizes data representation but NIST explicitly distinguishes representational checking from critical scientific evaluation."]},{"source_id":"S8","title":"Pure and Pseudo-pure Fluid Thermophysical Property Evaluation and the Open-Source Thermophysical Property Library CoolProp","publisher":"American Chemical Society; record hosted by Technical University of Denmark","url":"https://orbit.dtu.dk/en/publications/pure-and-pseudo-pure-fluid-thermophysical-property-evaluation-and/","source_class":"PRIMARY_RESEARCH","publication_date":"2014","accessed_at":"2026-08-03","claims_supported":["Peer-reviewed primary research documents implementation of many thermophysical-property correlations in one open-source library.","The library used C++ with wrappers for many programming languages and platforms, demonstrating technical portability of a common query surface.","The work confirms substantial diversity in thermodynamic and transport-property formulations behind software interfaces."]}],"problem_evidence":{"support":"MODERATE","rationale":"The problem's semantic ingredients are externally visible: IDAES requires declared validity bounds, units, and reference states; CoolProp documents phase-dependent derivative restrictions; and NIST requires model-specific provenance. CAPE-OPEN reports implementation errors and the need for behavioral clarifications. These sources establish that thermophysical interfaces can fail at domains, phases, references, derivatives, and integration boundaries. They do not establish the proposal's asserted prevalence of clients directly reading coefficients or quantify simulation failures caused specifically by representation-only substitutions.","source_ids":["S1","S4","S5","S6"]},"stakeholder_evidence":{"support":"STRONG","rationale":"CAPE-OPEN identifies process companies, simulator vendors, and academics that funded and maintained standard thermodynamic-component interfaces because users wanted to substitute third-party models without bespoke integration. IDAES is an identifiable DOE-sponsored modeling framework with explicit property-package requirements. This demonstrates strong pull for interoperability, although no source specifically requests the proposal's release-scoped behavioral oracle, uncertainty-preserving substitution rule, or immutable ThermophysicalSurface object.","source_ids":["S1","S2","S6"]},"prior_art":{"proximity":"SUBSTANTIAL_COLLISION","closest_analogues":[{"name":"CAPE-OPEN Thermodynamics and Physical Properties Interface","similarity":"A maintained industry standard already defines implementation-independent interfaces and expected behavior for interchangeable thermodynamic property packages in process-modeling environments.","remaining_difference":"The opened material does not show a release-scoped numerical conformance oracle that requires independently encoded evaluators of one fixed scientific surface to preserve preregistered tolerances, uncertainty, provenance, derivative behavior, and typed errors.","source_ids":["S1","S2"]},{"name":"CoolProp AbstractState and multiple backends","similarity":"An abstract public state/property protocol is already implemented by several distinct thermophysical backends, including REFPROP and IF97, closely matching the proposal's hidden-representation and substitutable-implementation structure.","remaining_difference":"The documentation establishes a shared protocol, not equivalence of different encodings of one immutable assessed model release or mutant-tested acceptance of numerical and metadata behavior.","source_ids":["S3","S4","S8"]},{"name":"IDAES Generic Property Package Framework","similarity":"IDAES formalizes state variables, bounds, units, reference states, and reusable property-package integration inside a major process-modeling framework.","remaining_difference":"It is a modeling-framework contract rather than an evaluator-neutral certification suite for multiple representations of the same released property surface.","source_ids":["S6"]},{"name":"ThermoML","similarity":"ThermoML standardizes machine-readable thermophysical data with uncertainty, constraints, methods, and bibliographic provenance.","remaining_difference":"It governs storage and exchange and explicitly does not critically evaluate the underlying science; it does not specify query behavior, derivatives, errors, numerical substitutability, or evaluator conformance.","source_ids":["S7"]}],"distinctive_claim_remaining":"For one fixed, preregistered thermophysical-model release, a behavioral contract covering domains, phases, values, units, reference states, supported derivatives, typed errors, uncertainty, and provenance—combined with a representation-blind, mutant-tested oracle—will accept independently coded analytic and table/spline evaluators when they are behaviorally equivalent and reject known coefficient, domain, derivative, and error-policy defects that API/schema-only compatibility checks admit.","confidence":"HIGH"},"implementation_evidence":{"support":"STRONG","rationale":"Technical feasibility is supported by CoolProp's existing abstract backends and derivative APIs, REFPROP's callable DLL/subroutines, IDAES's explicit state/bounds/unit/reference constructs, and ThermoML's uncertainty/provenance schema. An isolated synthetic trial requires no production-data authority and has low physical safety risk. Workflow feasibility is credible for an evaluation-library owner, but production adoption would require the property-release steward and simulator owners to agree on tolerances and version boundaries. REFPROP-backed implementations require license review; open or synthetic implementations avoid that issue. No source validates the proposed tolerance relation, oracle completeness, performance envelope, or ability to represent covariance-rich uncertainty without loss.","source_ids":["S1","S3","S4","S5","S6","S7","S8"]},"scores":{"meaningful_impact":{"score":3,"rationale":"Incorrect domains, phases, references, and derivatives can affect simulation results, but prevalence and realized harm from representation-only changes were not quantified.","source_ids":["S4","S6"]},"stakeholder_pull":{"score":4,"rationale":"An identifiable industrial coalition and DOE-sponsored framework have expressed and operationalized the need for replaceable thermodynamic components and explicit property-package semantics.","source_ids":["S1","S2","S6"]},"incremental_advantage":{"score":2,"rationale":"CAPE-OPEN, CoolProp, IDAES, and ThermoML already cover much of the interface, backend, metadata, and interoperability surface; the plausible increment is the fixed-release, mutant-tested behavioral oracle.","source_ids":["S1","S3","S6","S7"]},"distinctiveness_plausibility":{"score":2,"rationale":"The remaining claim is testable but narrow, and the search did not establish that existing conformance programs omit equivalent numerical or metadata tests.","source_ids":["S1","S3"]},"technical_implementability":{"score":4,"rationale":"Existing abstract backends, callable property libraries, derivative APIs, state bounds, units, and metadata standards provide most required building blocks.","source_ids":["S3","S4","S5","S6","S7","S8"]},"adoption_authority_feasibility":{"score":3,"rationale":"Release stewards, library owners, and simulator owners are identifiable authority classes, but no specific organization has agreed to steward this proposed contract or its tolerances.","source_ids":["S1","S2","S6"]},"evidence_readiness":{"score":4,"rationale":"A synthetic analytic surface, independent spline implementation, seeded queries, and corrupted controls form a bounded and reproducible experiment, although no results yet exist.","source_ids":["S3","S4","S6"]},"safety_net_benefit":{"score":4,"rationale":"Offline conformance, explicit out-of-domain failures, provenance retention, and unchanged incumbent software provide a strong rollback and attribution safety net.","source_ids":["S4","S5","S6","S7"]},"scalability":{"score":3,"rationale":"Common interfaces and cross-language wrappers scale technically, but every scientific release needs domain-specific tolerances, relations, derivative policies, and stewardship.","source_ids":["S1","S3","S8"]}},"score_confidence":"MODERATE","costs":{"first_evidence":{"band_2026_usd":"10K_TO_50K","scope":"Preregister one synthetic single-phase release; implement direct analytic, table/spline, and deliberately corrupted evaluators; build the contract harness; and execute seeded value, derivative, unit, metadata, boundary, and error tests.","confidence":"MODERATE","assumptions":["Six to ten person-weeks split between a thermophysical scientist and software/test engineer.","Uses open-source or newly written code and synthetic data, avoiding REFPROP redistribution fees.","Excludes production integration, physical experiments, scientific assessment, and regulatory validation."],"source_ids":["S3","S4","S6","S7"]},"initial_deployment_startup":{"band_2026_usd":"50K_TO_250K","scope":"Turn the prototype into a versioned contract specification, generated bindings, reusable conformance suite, two real-package adapters, provenance mapping, documentation, and CI integration.","confidence":"LOW","assumptions":["One bounded property family and two evaluator implementations.","Existing release metadata and tests can be reused.","Does not include broad simulator migration or safety-critical qualification."],"source_ids":["S1","S3","S5","S7"]},"operational_launch":{"band_2026_usd":"250K_TO_1M","scope":"Qualify one real assessed release across multiple simulator consumers, complete performance and numerical reviews, establish stewardship and deprecation procedures, remediate client reach-throughs, and conduct staged non-production then production acceptance.","confidence":"LOW","assumptions":["Two to four consuming applications and a small number of property backends.","No new physical-property assessment campaign is required.","Licensing and redistribution permissions are obtainable; REFPROP distributor use may require a separate agreement."],"source_ids":["S1","S2","S5","S6"]},"annual_recurring":{"band_2026_usd":"50K_TO_250K","scope":"Maintain the schema, bindings, test fixtures, reference evaluator, compatibility matrix, release reviews, security updates, and divergence triage.","confidence":"LOW","assumptions":["A part-time scientific steward plus engineering maintenance.","A limited release cadence and no continuous regulatory certification.","Costs exclude licensing fees and major scientific-model revisions."],"source_ids":["S1","S3","S5"]}},"verified_pipeline_gates":{"externally_supported_problem":{"status":"YES","reason":"Official documentation independently confirms that validity ranges, reference states, phases, derivatives, units, implementation errors, and model provenance materially affect thermophysical-property use; prevalence of coefficient-level client coupling remains unmeasured.","source_ids":["S1","S4","S5","S6"]},"externally_credible_adopter_or_authorizer":{"status":"YES","reason":"CAPE-OPEN identifies industrial operators and simulator vendors that sought substitutable thermodynamic components, while IDAES is a credible DOE-sponsored property-package adopter; neither has endorsed this exact proposal.","source_ids":["S1","S2","S6"]},"distinct_testable_incremental_claim":{"status":"YES","reason":"The remaining claim is explicitly contrastive against API/schema compatibility and is falsifiable using two independent representations plus corrupted controls under preregistered acceptance rules.","source_ids":["S1","S3","S7"]},"bounded_next_evidence_step":{"status":"YES","reason":"The synthetic single-phase trial has fixed comparators, query counts, boundaries, relations, tolerances, corrupted controls, halt conditions, and no production mutation.","source_ids":["S3","S4","S6"]},"no_unresolved_safety_or_authority_stop":{"status":"YES","reason":"The first experiment is offline, synthetic, reversible, and excludes operational or materials-selection use. Production deployment would require separate steward, simulator-owner, licensing, and any applicable safety-review authority.","source_ids":["S2","S5","S6"]},"credible_cost_scope_and_range":{"status":"UNCERTAIN","reason":"Four resource-equivalent bands are scoped with labor and integration assumptions, but no direct source provides comparable implementation costs and the number of releases, consumers, adapters, and licensing arrangements is not yet fixed.","source_ids":["S1","S3","S5"]}},"next_evidence_step":"Preregister a synthetic single-phase release over a fixed temperature interval with analytic enthalpy h(T), heat capacity cp(T)=dh/dT, SI and one reversible alternate unit system, one reference state, explicit domain endpoints, typed errors, provenance, uncertainty fields, and separate value and derivative tolerances. Independently implement (A) direct analytic evaluation and (B) a table/spline evaluator; add corrupted comparators with a shifted coefficient, shortened domain, wrong reference offset, missing uncertainty, and silent extrapolation. Run both valid implementations and every corrupt comparator through 30 fixed points including both endpoints, 500 seeded interior points, singleton-versus-batch and permutation checks, round-trip unit conversion, derivative consistency, provenance/uncertainty preservation, and invalid-query tests. Falsify the intervention if the oracle accepts any corrupt comparator, rejects the independent valid evaluator for representation-specific reasons, changes classification when comparator order is reversed, or cannot distinguish a scientific-release change from an encoding change. Stop immediately on plausible-valued extrapolation, lost provenance/uncertainty, or tolerance-order dependence; retain the specification and divergence log and leave production paths unchanged.","blocking_evidence":["No externally measured prevalence of simulators directly consuming coefficients, knots, segment indexes, or evaluator-private objects was found.","No quantified impact estimate links representation-only substitutions to solver failures, safety events, or decision changes.","The opened CAPE-OPEN materials were not exhaustively compared clause-by-clause with the proposed values, derivatives, uncertainty, provenance, error, and conformance obligations.","No empirical result shows that the proposed tolerance relation is symmetric, transitive enough for release governance, and resistant to directional bias.","No mutant-testing result demonstrates that the proposed oracle rejects under-domain, coefficient, derivative, metadata, or silent-extrapolation defects.","No named release steward or simulator owner has committed to adopt the proposed contract.","Production licensing, redistribution, validation, and safety-review requirements depend on the chosen evaluator and application and remain unverified.","World novelty, patentability, freedom to operate, market size, and realized impact were not measured."],"research_disposition":"PARTNERED_RESEARCH_PROGRAM","world_novelty_boundary":"The search establishes substantial prior art in abstract thermophysical backends, standardized property-package interfaces, explicit state/range/reference semantics, and standardized uncertainty/provenance data. It does not establish whether the exact combination of a fixed scientific-release identity, immutable behavioral surface, representation-blind numerical and metadata oracle, mutant controls, and substitution-versus-scientific-revision gate has previously been implemented. World novelty, patentability, freedom to operate, market size, and realized impact remain unmeasured.","arm":"COMPLETE_PROPOSAL_PORTFOLIO","candidate_version":0,"controller_recommendation":{"action":"STOP_EMPIRICAL_RESEARCH_NEEDED","repairable":false,"material_progress_observed":true,"progress_targets":["Complete the preregistered analytic-versus-spline-versus-corrupted-evaluator experiment and publish the fixtures, seeds, acceptance relation, and full divergence log.","Demonstrate that the oracle rejects every preregistered coefficient, domain, reference-state, derivative, metadata, and silent-extrapolation mutant while accepting both valid independent implementations.","Test symmetry, comparator-order independence, boundary behavior, numerical conditioning, and sensitivity of acceptance to value and derivative tolerances.","Obtain a clause-level differentiation matrix against CAPE-OPEN, CoolProp AbstractState, IDAES property packages, and ThermoML, identifying which obligations are absent, optional, or already standard.","Secure written interest from one property-release steward and one simulator or evaluation-library owner before any real-package integration."],"reason":"Bounded web research verified the problem ingredients, credible stakeholder demand, feasibility, and substantial prior-art collision, but it cannot determine whether the proposed oracle is discriminating, representation-independent, numerically stable, or operationally valuable. Those decisive facts require implementation and live sandbox testing. Under the evaluation rule, that requires STOP_EMPIRICAL_RESEARCH_NEEDED, and every STOP recommendation is non-repairable within this web-only evaluation."},"proposal_index":2}