TT Claude Memory

project_shape_x_full_lock_2026_04_23

Properties1
descriptionAll 8 Shape X structural angles locked after deep-session research-grounded restart. 24 amendments (A-1..A-23). 4-layer standards stack (OMG Commons + SKOS + XKOS + CIDOC E55 Type bridge). Alignment-axiom-only modularity. Per-target axiom-strength matrix. Shape X is now ready for Phase-1 implementation.

7 min read

Shape X — INHERIT v2’s Asset taxonomy structure — is fully locked. The session of Thu 23 April 2026 resolved all 8 structural angles via research-grounded restart after the 3-way agent triangulation revealed the initial methodology sequencing was inverted. Most substantive content survived; what needed revision was revised.

The 8 angles locked

AngleLockAmendmentKey artefact
(i) Top-level class counti-ζ 9 classes (FinancialAsset, RealPropertyAsset, AgriculturalPropertyAsset, VehicleAsset, BusinessAsset, ChattelAsset, DigitalAsset, IntellectualPropertyAsset, Receivable)A-109-class under prov:Entity → Asset per angle-(iii) upper lock
(ii) Class boundariesRole+RoleInstance ODP coupled pair (14th canonical primitive)A-20 + A-20.1 + A-22 axiom downgrade5 boundary questions resolved (OWL 2 punning + FRBR Work/Item + Obligation shared URI + DigitalAsset 3-way composition + Role pattern for Q1/Q3/Q5)
(iii) Upper-ontologyC = PROV-O upper + BFO peerShape X baseEmpirical EXP-1/EXP-2 spike; C=87% vs A=66% / E=78%
(iv) Alignment axiom strengthiv-3 per-target matrix + verification tiersA-22Default rdfs:subClassOf; owl:equivalentClass sparingly for genuine 1:1; rdfs:seeAlso for loose; DOLCE skipped. Verification rigor scales with axiom strength.
(v) T-box layeringα 3-layer (PROV-O upper + OWL 2 EL middle per module + SROIQ application for specific cases)Shape X baseELK reasoner for EL; HermiT/Pellet for SROIQ
(vi) Faceted classificationvi-6 OMG Commons Classifiers 4-layer stackA-23cmns-cls (ISO 1087-1 + ISO 11179-3) T-Box + SKOS A-Box + XKOS (ISO 25964) hierarchy refinements + CIDOC CRM E55 Type interop bridge
(vii) Extension/modularityvii-3 file-per-module + alignment-axiom-only + ROBOT-managed variants + 7-gate CIA-217 module files + separate ontology/schemes/*.ttl A-Box + separate ontology/alignments/*.ttl for external; NO runtime import of external reference ontologies
(viii) OWL 2 fragment per layerviii-3 profile-discipline authoring + multi-profile publication via ROBOT-derived variantsShape X base + A-21 reframing-edit/-base/-full/-simple per module from CI pipeline, not hand-authored

File layout (per vii-3 + A-23)

ontology/
├── core/inherit-core.ttl              # 14 canonical primitives T-Box
├── wills/inherit-wills.ttl            # owl:imports core
├── trusts/inherit-trusts.ttl          # owl:imports core, wills
├── probate/inherit-probate.ttl        # owl:imports core, wills, trusts
├── assets/inherit-assets.ttl          # owl:imports core
├── catalogue/inherit-catalogue.ttl    # owl:imports core, assets
├── delegation/inherit-delegation.ttl  # owl:imports core
├── schemes/                           # A-Box controlled vocabularies (SKOS + XKOS concepts)
│   ├── asset-category-scheme.ttl
│   ├── organisation-role-scheme.ttl
│   ├── jurisdiction-scheme.ttl
│   ├── faith-tradition-scheme.ttl
│   ├── life-state-scheme.ttl
│   ├── liability-type-scheme.ttl
│   └── partner-badge-scheme.ttl
└── alignments/                        # External-ontology alignments (opt-in; no runtime import)
    ├── inherit-to-bfo.ttl
    ├── inherit-to-provo.ttl
    ├── inherit-to-commons.ttl
    ├── inherit-to-fibo.ttl
    ├── inherit-to-cidoc.ttl           # CC-BY-SA 3.0 — alignment-ONLY, never bundle
    └── inherit-to-schema-org.ttl      # CC-BY-SA 3.0 — alignment-ONLY, never bundle

Per-target alignment-axiom-strength matrix (A-22 iv-3)

TargetAxiomLicenseVerification tier
OMG Commons (cmns-*)rdfs:subClassOfOMG Non-AssertFull spike (Role foundational)
BFO 2020rdfs:subClassOfCC-BY 4.0Full spike (upper)
PROV-Ordfs:subClassOf default + owl:equivalentClass only for 1:1 (per Beverley 2025)W3C Doc LicenseFull spike per equivClass target
FIBO specificrdfs:subClassOf for specialisation; rdfs:seeAlso otherwiseMITMedium (existence + version)
CIDOC CRMrdfs:subClassOf sparingly; mostly rdfs:seeAlsoCC-BY-SA 3.0 — alignment-ONLY never bundleMedium
schema.orgrdfs:seeAlso onlyCC-BY-SA 3.0 — alignment-ONLY never bundleLight (existence only)
DOLCESkip — choose BFO over DOLCE (peer annotational at most)unspecified (write LOA-CNR)—

Faceted classification 4-layer stack (A-23 vi-6)

  • Layer 1 — cmns-cls: (OMG Commons Classifiers; ISO 1087-1 + ISO 11179-3): T-Box frame with ClassificationScheme, Classifier, isClassifiedBy
  • Layer 2 — skos: (W3C REC 2009): A-Box skos:Concept individuals; multilingual prefLabel/altLabel/notation; broader/narrower
  • Layer 3 — xkos: (DDI Alliance; ISO 25964 SKOS extension): hierarchy refinements — generalizes/specializes, partitionOf, instanceOf, Correspondence (cross-scheme mapping)
  • Layer 4 — crm:E55_Type (CIDOC CRM; ISO 21127): interop bridge — E55 Type ⊒ skos:Concept per CIDOC RDF v1.1. Alignment-axiom-only per vii-3 (no wholesale Erlangen import).

Every facet relation is rdfs:subPropertyOf cmns-cls:isClassifiedBy. No local invention of predicates.

FIBO uses this exact pattern: fibo-fnd-arr-cls:IndustrySectorClassificationScheme rdfs:subClassOf cmns-cls:ClassificationScheme with SKOS for annotations only.

14 canonical Apache-core primitives

Asset, Liability (A-4), AssetCollection, Space, Valuation, Provenance, Completeness, Visibility, OwnerIntent, Identifier+ExternalLink, Media, AuditLog, Organisation (A-11), Role+RoleInstance coupled pair (A-20+A-20.1+A-22 — one canonical counted as pair).

Plus 5 cross-module primitives (Capacity A-3, Person A-9, 3 SRI-3 health primitives A-15) = 19 effective Core primitives.

Phase-1 module-authoring sequence LOCKED (A-24, 2026-04-24)

Seq G1 — gap-discovery-optimal (scorecard 94.0% vs alternatives 40-84%):

  1. Core (~2-3 weeks) — 14 canonical + 5 cross-module primitives
  2. Assets (~2-3 weeks) — 9 top-level classes + 3 domain roles + AssetCategory facet
  3. Wills — E&W minimal (~1-2 weeks) — E&W Wills Act 1837 + AJA 1982 + IHTA Part III only; Scotland + Swiss + karta + faith-tradition Catala as InheritKit crates AFTER
  4. Delegation (~1-2 weeks) — DelegationCredential + W3C VC 2.0
  5. Catalogue (~1-2 weeks) — AssetCollection + wishlist + SoftSale + dealer roles
  6. Trusts (~3 weeks) — E&W full + karta HUF
  7. Probate (~2-3 weeks) — E&W full

Total Phase-1: ~15-20 weeks. Rich+Claude solo-bootstrap.

Rationale: Seq G1 tests 9 of 11 gap-type axes by position 5 (structural + axiom-strength + facet-stack + role-pattern + legal-jurisdictional + external-spec + commercial + Apache-deferred-movement). Revisions are cheap before Trusts+Probate are committed. Today’s session (FIBO spike + A-22 + A-23) demonstrated architecture benefits from early stress-testing.

“Wills E&W initial scope at position 3” is the key structural choice — tests the per-jurisdiction InheritKit crate pattern with smallest tractable scope. E&W is initial scope for commercial tractability, NOT the reference case per feedback_inherit_v2_international_from_day_one.

Phase 1.5 GATE (A-25, 2026-04-24) — hard gate before v2 Phase-1 complete

INHERIT v2 is an international standard covering every jurisdiction / succession regime / tax treatment / faith tradition. Phase-1 sequence (Seq G1) authors an E&W baseline for commercial tractability; Phase 1.5 is the mandatory validation that prevents this sequencing from contaminating the international-standard positioning.

Trigger: Phase-1 position 7 (Probate E&W) completes + all 7 CI gates pass.

Scope: v6.6 has 21 extensions already authored (~8,185 LOC at ~/openinherit/code-standard/v3/extensions/). Phase 1.5 systematically imports each extension into v2’s 7-module architecture. 21 extensions × 7 modules = 147 module-extension cells to stress-test.

v6.6 extensions to import: UK-EW, US, AU-NZ, Brazil, EU-Succession, Israel, Japan, Scotland, Canada, Hong Kong, India, Ireland, Latin America, PRC China, Switzerland, UAE, Singapore/Malaysia + Africa Customary, Hindu, Islamic, Jewish.

Outcome criteria:

  1. Every v6.6 extension roundtrips through v2 without data loss
  2. Architectural adjustments committed as A-26+ amendments
  3. Module-boundary moves committed (per feedback_inherit_v2_international_from_day_one: entities move freely)
  4. Alignment axioms updated per any new external-ontology references
  5. Gap-discovery report + residual-risk assessment
  6. IRI-verification spikes for any new external alignment

Permissions during Phase 1.5: move entities between modules; expand Core primitives; add facet values / domain roles; revise per-target matrix.

NOT permitted: revoke Shape X locks without scorecard; add jurisdictional content to Core Apache surface (InheritKit crates only); skip IRI-verification; bundle CC-BY-SA content.

Budget: ~4-6 weeks. NOT optional. Phase-1 is NOT complete until Phase 1.5 passes.

Output artefact: docs/superpowers/specs/2026-XX-phase-1-5-stress-test-report.md — companion to architecture-proposal v1.12 for ISO/OASIS PAS submission defence.

Downstream decisions UNBLOCKED by Shape X FULL LOCK

  • Apache-commitment-timing per-module table (Z24 W4-C1)
  • Cross-module $ref coherence table (Z24 W4-C2)
  • Asset.sourcedVia enum (Z20 W4-C18)
  • @inherit/core/ontology/*.ttl + *.jsonld authoring (Phase-1 implementation can begin)
  • ontology/schemes/*.ttl A-Box controlled-vocabulary authoring
  • Z23 TCK test-vector freeze for Asset taxonomy
  • Partner-model v2.5 → v2.6 revision
  • Pre-public-release deliverables (NOTICE, THIRD-PARTY-ONTOLOGIES.md, license-audit CI gate, LOA-CNR DOLCE grant request)

Key commits (Thu 23 April 2026 session)

  • 1b11ab0 A-20 RoleInstance as 14th canonical (angle ii)
  • 508c5cf FIBO Role version-pin spike findings
  • ecd2f2d A-20.1 FIBO→OMG Commons IRI correction
  • 5e84f2d IRI-verification-before-lock memory + research
  • 3afba2c IRI terminology discipline adopted (Q2)
  • 8ca6902 Shape X restart plan (research-grounded)
  • 43e8a7f A-21 angle (vii) vii-3 alignment-axiom-only modularity
  • 7cecaa9 A-22 angle (iv) iv-3 per-target matrix
  • b44abb3 Shape X FULL LOCK (A-23 angle (vi) vi-6)

Research notes (library-local; not git-tracked)

  • ~/off-github/library/projects/inherit/class-boundaries-research-2026-04-23.md (A-20 basis)
  • ~/off-github/library/projects/inherit/angle-sequencing-research-2026-04-23.md (methodology agent)
  • ~/off-github/library/projects/inherit/alignment-axiom-current-practice-2026-04-23.md (Agent 1)
  • ~/off-github/library/projects/inherit/modularity-patterns-current-practice-2026-04-23.md (Agent 2)
  • ~/off-github/library/projects/inherit/license-compatibility-current-practice-2026-04-23.md (Agent 3)
  • ~/off-github/library/projects/inherit/faceted-classification-standard-2026-04-23.md (Agent 4)
  • ~/off-github/library/projects/inherit/research-synthesis-in-progress-2026-04-23.md (my corroboration)

Spike docs (git-tracked)

  • docs/superpowers/specs/spike-upper-ontology-2026-04-23/ (EXP-1/2/3 empirical + EXP-3 materials)
  • docs/superpowers/specs/spike-fibo-role-2026-04-23/findings.md (A-20.1 trigger)

How to apply

  • When authoring any INHERIT v2 ontology file: follow the file-layout (vii-3) + per-target matrix (iv-3) + cmns-cls facet pattern (vi-6) + Role+RoleInstance pattern (ii) unchanged.
  • When aligning to any external ontology: use the per-target matrix in A-22; run IRI-verification spike before committing equivClass; medium verification for subClassOf; light for seeAlso.
  • When adding a new facet: it’s a cmns-cls:Classifier subclass with a cmns-cls:ClassificationScheme; A-Box individuals are skos:Concept; relation rdfs:subPropertyOf cmns-cls:isClassifiedBy; hierarchy refinements via XKOS if needed.
  • When adding a new canonical primitive: update this memory + feedback_v2_asset_catalogue_scope + architecture-proposal v1.12 §Core primitives.

Cross-references

  • Memory feedback_v2_asset_catalogue_scope — 14 canonical primitives list (updated with A-23 cmns-cls pattern)
  • Memory feedback_iri_verification_before_lock — verification discipline that compounds with A-22 verification tiers
  • Memory feedback_iri_terminology_discipline — IRI over URI in new writing
  • Memory feedback_v2_apache_deferred_until_public_release — why license-compatibility matters
  • Memory project_v2_architecture_locked_2026_04_22 — parent architectural lock (this is a child in the sense that Shape X is a component of the Option-G+/Ⅲ/D-ii architecture)
  • docs/superpowers/specs/2026-04-22-inherit-v2-architecture-proposal.md v1.12 (Shape X FULL LOCK 8/8)
  • docs/superpowers/scoping/2026-04-23-v1-12-amendments.md (24 amendments)
  • docs/superpowers/scoping/2026-04-23-shape-x-restart-plan.md v1.0 (research-grounded restart)
  • docs/superpowers/scoping/2026-04-23-session-close-summary.md (full arc of the day)

Graph View

Backlinks