Plan, apply and roll back changes to a production ontology, with a blast radius report and a proof an auditor can re-check.
Rust ai-nativedescription-logicsformal-verificationinfrastructure-as-codeisabelle-languageknowledge-graphleanlean4mcpmcp-serverontologyowlowl2reasoningrustshaclsparqltheorem-proving
Scope the DCAT-US binding claim in RQ8 and its case study
FABIOTESSc09bd52a2524
diff --git a/README.md b/README.mdindex b9fa31b1..f3105e0d 100644--- a/README.md+++ b/README.md@@ -335,7 +335,7 @@ The benchmarks below are not a feature tour. Each one exists to answer a specifi| **RQ6** | Can the SHACL-SPARQL governance rules an ontology ships actually be executed as written at national scale, or does the reference store force a workaround? | [`investment-fund-ontology`](case-studies/investment-fund-ontology/): 1.29M-triple US fund universe, 4 layer-3 rules, rdflib 7.6.0 A/B | **Store-specific.** rdflib finishes the anti-join (0.1s) and nested aggregate (0.8s) but times out (>300s) on both multi-way self-joins — exactly the cross-source reconciliation rules the ontology exists for. This repo's Oxigraph engine runs load + all four rules, unmodified, in 2.1s, reproducing the reference pipeline's published counts (73 / 0 / 214 / 2,148) exactly. The shapes also forced two validator fixes here: `sh:inversePath` support and `sh:severity` reporting. || **RQ7** | When the graph is register-scale (276k triples) rather than fund-universe-scale (1.29M), does the reference store still force the set-based workaround, and do the rule counts survive entity resolution? | [`insurance-register-ontology`](case-studies/insurance-register-ontology/): 276,683-triple EEA insurance register, 6 layer-3 rules, rdflib 7.6.0 A/B | **The workaround becomes speed, not survival.** At this scale rdflib completes all six rules (133.2s total, 97s of it in one double anti-join) against 2s to 9s for the Oxigraph engine; nothing times out. Counts match the reference pipeline exactly on R1-R4 (643 / 4 / 118 / 42). R5 falls from 227 to 20 and R6 rises from 283 to 291 because the graph runs the same rule text after entity resolution, which absorbs the legitimate branch shares and attributes 8 more zombie passports via the LEI join; both differences were verified set-based. The shapes forced two more validator fixes: `sh:pattern` and `sh:hasValue`. |-| **RQ8** | When a published standard ships a validator but no RDF binding, does a green SHACL result mean the data is valid, or that nothing was checked? | [`dcat-us-binding`](case-studies/dcat-us-binding/): DCAT-US 3.0, the US federal metadata profile, 115 published examples, pySHACL vs this engine | **It means nothing was checked, and the distinction is invisible without reporting reach.** The corpus as published expands to **76 triples, 1 predicate, 0 DCAT**, so the shapes GSA deleted select **0 focus nodes** and return `conforms=True`. Generating a JSON-LD context from the same JSON Schema that already carries the RDF terms (231 of 270 properties) lifts the corpus to **1,510 triples and 228 DCAT statements**, and the *same deleted shapes*, unchanged, then select **228 focus nodes and report 287 violations**. This is why `onto_shacl` reports `focus_nodes` next to the verdict. The run also cost us two defects here: `sh:class` and `sh:nodeKind` are not evaluated at all (287 of pySHACL's 316 findings invisible), and `sh:datatype` rejects `"1024"^^xsd:nonNegativeInteger` against `sh:datatype xsd:nonNegativeInteger`. |+| **RQ8** | When a published standard ships a validator but no RDF binding, does a green SHACL result mean the data is valid, or that nothing was checked? | [`dcat-us-binding`](case-studies/dcat-us-binding/): DCAT-US 3.0, the US federal metadata profile, 115 published examples from its live repository, pySHACL vs this engine | **It means nothing was checked, and the distinction is invisible without reporting reach.** The corpus as published expands to **76 triples, 1 predicate, 0 DCAT**, so the shapes GSA deleted select **0 focus nodes** and return `conforms=True`. Generating a JSON-LD context from the same JSON Schema that already carries the RDF terms (231 of 270 properties) lifts the corpus to **1,510 triples and 228 DCAT statements**, and the *same deleted shapes*, unchanged, then select **228 focus nodes and report 287 violations**. This is why `onto_shacl` reports `focus_nodes` next to the verdict. The run also cost us two defects here: `sh:class` and `sh:nodeKind` are not evaluated at all (287 of pySHACL's 316 findings invisible), and `sh:datatype` rejects `"1024"^^xsd:nonNegativeInteger` against `sh:datatype xsd:nonNegativeInteger`. |RQ2 and RQ4 are the ones worth reading if you are deciding whether to trust this repo. Both are negative results about work done here.diff --git a/case-studies/dcat-us-binding/README.md b/case-studies/dcat-us-binding/README.mdindex 8d8d4fce..7fa6827e 100644--- a/case-studies/dcat-us-binding/README.md+++ b/case-studies/dcat-us-binding/README.md@@ -17,9 +17,20 @@ implementation of the World Wide Web Consortium's (W3C) DCAT standard". DCAT v3The project inherited a website, a JSON Schema and a SHACL definition that were out ofsync, with nobody left who knew which was authoritative, and collapsed to the JSON-Schema as the single source of truth. The SHACL file was deleted in PR #120. What was-never published is a JSON-LD context, so the JSON has no defined RDF interpretation at-all.+Schema as the single source of truth. The SHACL file was deleted in PR #120. The live+repository publishes no JSON-LD context, so the JSON it ships has no defined RDF+interpretation.++A context for the profile does exist, at+[DOI-DO/dcat-us](https://github.com/DOI-DO/dcat-us), 48,311 bytes alongside Turtle+serialisations of every example. That repository is marked deprecated in its own README,+which points to <https://resources.data.gov/resources/dcat-us3/>, and `GSA/dcat-us`+re-checked at `main` on 4 September 2026 still publishes no context, JSON-LD, SHACL or+Turtle artefact outside its `DEPRECATED/` tree. The measurement below is against the live+repository. The deprecated context is worth reading as the control case rather than a+counterexample: it is hand-maintained and it has drifted, declaring a namespace host that+does not resolve and minting the W3C Organization Ontology under `w3c.org` instead of+`w3.org`. A context generated from the schema cannot acquire either defect.## The result