TT Claude Memory

project_zeta_q3_eps_iota_S6_am_cdm_precedent_2026_05_02

Properties1
descriptionBuffalo Manufacturing Works AM-CDM precedent — 6-module split confirmed; 173/174 classes subclass single CCO DocumentField (NOT transferable to INHERIT v2 i-ζ); zero SSSOM; zero CI; ω.η lock validation comes from Mondo (S7) NOT AM-CDM; AM-CDM is SUPPORTING precedent for ε.ε ONLY

3 min read

ζ-Q3 ε.ι S6 AM-CDM-Ontology-Map precedent inspection — outcome-VALIDATED-WITH-NOTE 2026-05-02

Outcome

outcome-VALIDATED-WITH-NOTE — sufficient transferable + non-transferable patterns surfaced for ε.ι lock-time framing input. AM-CDM is a limited precedent: it shares the CCO-mid-level-anchor pattern with INHERIT v2 ε.ε but takes a fundamentally different approach to data modelling (everything-is-a-DocumentField class) + ships zero SSSOM machinery + zero CI infrastructure.

Key surprise: 173 of 174 AM-CDM classes subclass a single CCO class (DocumentField). This is a viable but design-specific choice — not the canonical CCO pattern. INHERIT v2 should NOT adopt this single-anchor pattern.

Findings

AspectFinding
Modules6 (BASE / BUILD / MATERIAL / PROCESS / SYSTEM / TESTINSPECTION) — matches plan §5 Task 8 description
Class anchor173/174 → CCO DocumentField (single-anchor); 1 → NominalMeasurementInformationContentEntity
SSSOM filesZero (find . -name "*.sssom*" returns empty)
CI infrastructureZero (no .github/workflows/, no Makefile, no validation scripts)
ImportsRemote raw.githubusercontent.com URLs (fragile per S2 evidence; INHERIT v2 A-33 commit-SHA pinning is BETTER)
Aboutness layerToOntology/ files use OWL Restrictions like someValuesFrom is_about / designates / is_a_measurement_of
LicenseApache-2.0

Transferable patterns (6)

  1. Per-module file split (precedented for INHERIT v2 9-module split per A-124)
  2. Layered import structure (BASE foundational; BUILD/MATERIAL import BASE; etc.)
  3. Two-layer architecture (CCO maps + aboutness ↔ INHERIT v2 LinkML schema + SSSOM TSV)
  4. CCO mid-level anchor pattern
  5. Aboutness-via-restriction pattern (Year-2+ uplift candidate for INHERIT v2)
  6. Single-merged-view file (AM-CDM-ALL-MERGED.owl ↔ NEW Phase-1 Sprint S1 candidate inherit-v2-merged.owl CI artefact ~½ day)

NOT-transferable patterns (4)

A. Single-class-anchor (everything-subclasses-DocumentField) — different framing from INHERIT v2 MaterialEntity-side i-ζ B. Zero-SSSOM machinery — cannot express maturity-graded alignments per A-55 verificationTier C. Zero-CI infrastructure — INHERIT v2’s 22 A-21 CI gates are AHEAD D. Remote raw.githubusercontent.com OWL imports — fragile; use A-33 commit-SHA pinning + locally-vendored ontologies

Implications for Phase E Task 13 ε.ι lock-decision

  1. AM-CDM is a SUPPORTING precedent for ε.ε, not a model for ε.ι. ε.ε’s CCO/BFO-rooted-9-classes structure is precedented by AM-CDM’s CCO mid-level anchor pattern. AM-CDM does NOT model ε.ι’s distinctiveness layers (Layer 1 owlready2 emitter, Layer 2 SSSOM-FIBO alignment, Layer 4 universal-3-layer-pipeline) — those have other precedents (Mondo, Catala, FIBO, OG-RAG).
  2. inherit-v2-merged.owl CI artefact — NEW Phase-1 Sprint S1 task candidate per Pattern 6.
  3. Aboutness-layer Year-2+ uplift per Pattern 5 — INHERIT v2 can add inherit-v2-aboutness.owl for OWL-restriction-based descriptions.
  4. DROP candidate: any “AM-CDM as standards-track precedent” claim.
  5. NO architectural change to A-131 ε.ι amendment row.

Cross-references

  • T-file: ~/off-github/library/projects/inherit/T-spike-eps-iota-S6-am-cdm-precedent-2026-05-02.md v1.0 (22 KB, 208 lines)
  • arch-state v3.26 → v3.27 §11 S6 row LANDED + Changelog row
  • Q-003 v1.8 → v1.9 §10 S6 row LANDED + CHANGELOG entry
  • MEMORY.md +1 entry
  • Active-work-log entry
  • Companion artefacts at /tmp/spike-s6-am-cdm/: inspect_am_cdm_v2.py (rdflib-based survey) + inspect_restrictions.py + inspect-output.log

Methodological observations

  1. First PRECEDENT INSPECTION spike with NO kill condition per plan §5; outcome-VALIDATED-WITH-NOTE vocabulary correctly captures “≥6 transferable + ≥4 not-transferable patterns documented”.
  2. Alternatives-first discipline generalised beyond kill-condition-met scenarios — applied at toolchain layer (owlready2 → rdflib) when import-follow failed on remote BFO TTL. Pattern: alternatives-first becomes a generally-applicable problem-solving pattern beyond kill-condition-only triggers. Forward-looking refined-prompt v3.7 candidate.
  3. 11th spike consecutively (S2+S2.5+S3+S2.10+S2.6+S4+S2.9+S2.9b+S5+S8+S6) with logging-contract closed within same session as T-file authoring.

Graph View