TT Claude Memory

project_zeta_q1_locked_option_c_2026_04_30

Properties1
descriptionFirst architectural amendment (A-126) under 2026-04-30 ζ.2 re-derivation methodology pivot. Locks 14-criterion frame derived from Bass-Clements-Kazman 4e + ISO/IEC 25010:2023 + AWS WAF + 2024-2026 SOTA (BEA 2e + MCDA SOTA + AI architecture evaluation). Includes Bass acquisition + T82 synthesis + utility tree v1.0 (2/14 criteria fully built out). ζ-Q2 next-session under MCDA SOTA sensitivity discipline.

5 min read

⚠️ PRE-REFINED-PROMPT-archived 2026-05-01T20:45 BST per Full Phase-2-3 Reversion — this memory file records a lock that has been DISCARDED. The Q is queued for re-ask under refined-prompt v3.0 discipline. Original lock content preserved below for forensic continuity. Do not act on this memory’s recommendations as if locked. Audit-record: ~/testatetech/docs-strategy/docs/superpowers/specs/2026-04-29-multi-phase-audit/audit-records/2026-05-01-full-phase-2-3-re-ask-under-refined-prompt.md v1.0. Refined-prompt v3.0 reference: ~/testatetech/docs-strategy/docs/superpowers/specs/2026-04-29-multi-phase-audit/refined-end-of-turn-directive.md.

ζ-Q1 LOCKED — Option C 14-criterion frame Thursday 30 April 2026 evening

§1 Outcome

Rich-locked Option C re-derived 14-criterion frame from primary sources, 2026-04-30T18:50 BST. Convergent decision (Claude recommendation was also C; both arrived at C from independent reasoning paths). 1.55 weighted-sum margin to F (decisively rejected per Bass §14.2 critique-by-name of ISO 25010 wholesale); 0.74 margin to G (TT-product-derived; honest fallback strengthened by Bass §14.2 “Iowability” precedent). 10/10 sensitivity perturbations preserve C.

§2 The 14 locked criteria

  1. Functional Suitability [ISO 25010:2023 + Jackson] — completeness + correctness + appropriateness
  2. Reliability/Faultlessness [ISO 25010:2023 sub-char rename]
  3. Safety [NEW ISO 25010:2023 9th characteristic 2023] — operational constraint + fail-safe + hazard warning + safe integration
  4. Formal-verification provability [T80 split] — SymCC + Lean spec + cvc5 SMT + Catala typing
  5. Temporal precision — datetime-precise + bi-temporal + 80-year replay
  6. Maintainability [ISO 25010:2023]
  7. Operational Excellence [AWS WAF 2025] — two-person-operability + automated CI gates
  8. Cost Optimization [AWS WAF 2025] — runway-discipline binding
  9. Compatibility [ISO 25010:2023]
  10. Conformance-test-suite presence [T63] — Bowtie + AAIF + per-engine harness
  11. Standards-track positioning [T41 split + OASIS] — PAS-route + IPR-mode + per-body-cost + standards-influence
  12. AI-vendor procurement attestability [T80 + A-122 + 2025 SOTA] — TRiSM + AgentArcEval + CSS + TUE
  13. Tooling ecosystem maturity [T81 lesson — empirical-verified]
  14. Acquirer-narrative strength [v1.9 #16 reframed]

§3 What was excluded under ζ.2

5 v1.9 criteria dropped:

  • v1.9 #6 Developer-joy [Jackson engineering-internal — folded into Maintainability]
  • v1.9 #13 100yr horizon [temporally mismatched]
  • v1.9 #14 Talent-pool [engineering-internal — folded into Operational Excellence]
  • v1.9 #15 LLM-moat [v1.2 no-fine-tune constraint inverts]
  • #4 Continuity-with-v1.9 [NEW exclusion under ζ.2 — explicitly does not anchor on G+]

§4 Cascade artefacts (committed)

  • arch-state Tier-1 v3.10 → v3.11 with A-126 amendment (commit 03990e9)
  • ζ-Q1 cascade file moved current-questions/ → answered-questions/ (commit 03990e9)
  • 2026-04-18 architecture-options scorecard v1.9 → v2.0 status: superseded (commit 03990e9)
  • Utility tree v1.0 authored: Q-001-zeta-criteria-frame-locked-utility-tree.md (commit 03990e9)
  • INDEX v2.1 → v2.2 (commit 03990e9)
  • Stage-1 §K answer commit 901a6f5

§5 Library work (NOT in git per filing-system discipline)

  • Bass-Clements-Kazman 4e ingested at ~/off-github/library/indexed/bass-software-architecture-4e/ (PDF + extracted text + TOC + metadata + notes with 15 ULID-cited verbatim quotes + exclusion-log)
  • T82 architecture-evaluation methodology synthesis at ~/off-github/library/projects/inherit/T82-architecture-evaluation-methodology-2026-04-30.md v1.0 (12 sections; 14 primary sources)
  • Project CLAUDE.md T-file quick-reference updated to include T82 (commit 9cd1931)

§6 Research findings for ζ-Q2 option-set

5 carry-forward + 1 NEW honest 6th-option candidate:

  • Option B — JSON Schema family (current direction; baseline)
  • Option C — Semantic Web purist (RDF + OWL 2 DL + SHACL + SPARQL)
  • Option D — Pragmatic hybrid (KG-canonical, LLM-assists)
  • Option F — Catala-heavy maximalist
  • Option G+ — Commercial-Catala (current canonical lock candidate)
  • Option I [NEW] — Graph-RAG-Native ontology-driven hierarchical-temporal (arxiv:2505.00039 May 2025; 70% pass-rate vs RAG baseline 37.5% on Indian legal-norms case study; matches T13/T37/T44 LinkML direction + Bass §19.4 utility-tree pattern)

Option E remains INVALIDATED (TT-fine-tune cost prohibitive per feedback_no_tt_training_data_for_inherit).

Rejected as 6th-option candidates (honest discipline):

  • Lean 4 native architecture — too pioneer-territory; no legal-domain precedent (Harmonic AI $100M + Mathlib 210K theorems but no legal applications)
  • Smart legal contracts — regulatory gap (no state recognises blockchain wills); not viable as canonical architecture
  • WASM 3.0 (Sept 2025) — deployment substrate not paradigm; relevant as Catala compile target but doesn’t constitute distinct option

§7 Next session work

Stage-3 utility tree v1.0 → v1.1: full ASR scenarios (2-3 per criterion) + fitness functions for criteria 2 + 4-14 (12 criteria). ~5-7 hours inline OR 10-12 hours subagent. Required before ζ-Q2 per-option SCORING begins.

Stage-4 ζ-Q2 cascade: 6 options (B/C/D/F/G+/I) scored against locked utility tree’s leaf scenarios under MCDA SOTA sensitivity discipline (COMSAM multi-criterion simultaneous perturbation + weight stability intervals + 8-mode PySensMCDA-style analysis per ζ-Q1 §I).

§8 Methodological observations

  • Convergent decision (Rich + Claude both arrived at C from independent reasoning paths). Suggests methodology pivot’s discipline is working — when fresh evidence is engaged + presented honestly + sensitivity-tested, Rich + Claude tend to converge.
  • Bass acquisition was decisively right. Single-book purchase (~£50-70 + 3 hours indexing+notes) closed the most consequential library-coverage gap. Recommendation strength rose from Claude-synthesised to SEI-canonical-anchored. Compare to alternative (5-book set ~£525-660 + multi-week wait).
  • First A-numbered amendment (A-126) under ζ.2 methodology registered cleanly through pin-drift hook + Stage-1/Stage-2 separation discipline. Methodology pivot validated through actual cascade execution.
  • ζ-Q1 took 4 hours wall-clock from MQ-000 lock to ζ-Q1 lock (16:30 → 18:50 BST + research/cascade through 19:30). Includes Bass acquisition + indexing + T82 synthesis + 4 web searches. Pace is realistic for ζ-Q2..ζ-Q10 if Rich-availability is consistent.
  • Utility tree v1.0 with 2/14 fully built is sufficient for ζ-Q2 SCOPING work but not for full per-option SCORING. v1.1 expansion is the gating task for ζ-Q2 to begin.

§9 Cross-references

  • Locked answer: docs/superpowers/specs/2026-04-29-multi-phase-audit/answered-questions/Q-001-zeta-criteria-frame-for-broader-options.md v1.1
  • Utility tree: docs/superpowers/specs/2026-04-29-multi-phase-audit/Q-001-zeta-criteria-frame-locked-utility-tree.md v1.0
  • arch-state amendment: docs/superpowers/specs/inherit-v2-architecture-state.md v3.11 A-126
  • T82 methodology: ~/off-github/library/projects/inherit/T82-architecture-evaluation-methodology-2026-04-30.md v1.0
  • Bass per-book: ~/off-github/library/indexed/bass-software-architecture-4e/
  • Design spec: docs/superpowers/specs/2026-04-30-architecture-rederivation-design.md v1.1
  • MQ-000: docs/superpowers/specs/2026-04-29-multi-phase-audit/meta-questions/MQ-000-prior-work-revisit-scope.md v1.0

Graph View

Backlinks