dot-iu-cutter v0.5 — WS-3 Gap & Readiness Review

Date: 2026-05-18 Phase: v0.5 WS-3 — gap & readiness assessment Nature: design_only — assessment of whether WS-1 + WS-2 authority is sufficient for the WS-3 logical proof. No schema, no code, no mutation. Conflicts are flagged, not self-resolved.

QG7 — plain terms. This file answers one question: "Did WS-1 and WS-2 give us enough to do the WS-3 paper proof, and what is still missing for later?" It sorts every gap into three buckets: things to fix before any database schema is designed, things to fix before production, and things that can wait until after the pilot.

1. Verdict on sufficiency

ws1_ws2_sufficient_for_ws3_logical_proof: YES
ws3_logical_proof_result: PASS_WITH_BLOCKERS
  # logical proof itself stands; blockers are for the NEXT phases, not for the paper proof
reason: >
  WS-1 supplied the assembly_profile logical contract (incl. cross_source_topic_thread
  sketch G4), the topic registry logical contract, the binding boundary, and the
  edge-minimization result. WS-2 supplied source_family registry, authority_semantics
  default + D4 override, grammar profiles, entity_reference_registry, and the canonical
  address scheme. Together these were sufficient to express every node, every relation,
  every authority label, and the full P11D pipeline for sample topic collection_structure
  WITHOUT redesigning anything (QG1) and WITHOUT any forbidden action.

2. Blocker / non-blocker classification

Buckets per the user's required taxonomy: must_resolve_before_schema, must_resolve_before_production, can_defer_post_pilot.

2.1 must_resolve_before_schema

- id: BR-A1-canonical-address-separator-final-confirm
  what: OD-A1 separator (slash docprefix vs hyphen levels) is a GPT recommendation
        PASS_WITH_RULING, not a final User/GPT lock.
  why_blocks_schema: any physical address column / iu_id derivation hardcodes the
        separator; must be locked before a schema cycle (v0.5 Q5).
  ws3_impact: NONE on the paper proof (scheme is expressible as-is).
  resolution_owner: GPT/User — NOT self-resolved.

- id: BR-FA1-assembly_profile-physical-shape
  what: OD-FA1 logical contract resolved (resolve_in_WS1) but physical storage
        (new table vs ledger vs JSONB) is defer_WS3/Q5 — undecided.
  why_blocks_schema: schema cycle cannot start without choosing storage option.
  ws3_impact: NONE (logical instance suffices for the proof).
  resolution_owner: schema-design cycle Q5 (sovereign gate).

- id: BR-FA3-topic-registry-physical
  what: OD-FA3 Option C ratified logically; physical topic_vocab deferred to
        Pilot Tier 1 + duplicate-rate + APR Đ32 (P44-6).
  why_blocks_schema: topic_code referential integrity needs the physical decision.
  ws3_impact: NONE — topic_code used as free-text + mandatory provenance (P11D).
  note: topic registry logical-only IS sufficient for the pilot proof (see §4).

2.2 must_resolve_before_production

- id: BP-FA4-evidenced_by-APR
  what: evidenced_by is the single minimized new-edge candidate (OD-FA4 → {evidenced_by}).
        GPT ruling NOT_APPROVED_AS_6_EDGE_BATCH. APR NOT taken.
  why_blocks_production: a queryable/durable "X is proven by Y" edge cannot exist in
        production until APR cấp medium (Đ32) + edge_type vocab amend (NT4).
  ws3_impact: NONE on the paper proof — WS-3 uses alternative B (assembly-local,
        output-only). evidenced_by appears only as APR-gated logical example (QG5/QG10).
  is_it_blocking_ws3: NO. is_it_blocking_production_compliance: YES.
  resolution_owner: GPT/User APR — NOT self-resolved.

- id: BP-WS2-2-candidate-edge-activation  (risk R-WS2-2)
  what: edge-minimization reuses governed_by (constrains) and derived_from (lesson_from);
        both are P44-4A §3.2 CANDIDATE edges pending User v1.0 ratification.
  why_blocks_production: if candidates are NEVER activated, those resolutions MUST be
        revisited before WS-3 execution (GPT caution, verbatim).
  ws3_impact: NONE on paper proof — if inactive, R1/R7 fall back to assembly-local
        (reduced queryability, proof still holds).
  resolution_owner: P44-4A v1.0 ratification (User) — NOT self-resolved.

- id: BP-FA8-generic-executor-prototype
  what: OD-FA8 generic executor over P11D §6.1 = logical prototype only, code defer_WS3.
  why_blocks_production: no engine code exists; production needs the executor built
        (separate, gated; no code now).
  ws3_impact: NONE — pipeline proved logically.

2.3 can_defer_post_pilot

- OD-FA2 physical IU↔entity binding table (Option D hybrid logical-resolved)
- OD-FA3 physical topic_vocab (P44-6)
- OD-FA5 entity_reference_registry permission_policy_ref / snapshot_policy_ref (deferred placeholders)
- OD-G1 internal Architecture Constitution authority + grammar leaf detail (defer_post_pilot)
- OD-S1 KB-7294 rev44 authority / 📋 controlled-draft policy (defer_post_pilot)
- OD-L1..L5 metadata physical (ledger vs table, JSONB vs EAV, hot-key threshold number)
- P11D-α/γ/δ/ε topic vocab table / confidence threshold number / alias / lifecycle grace
- P11D-β / P11D-ζ topic source-method detail / Q-T3 edge whitelist final + uses + max hops
- P44-4A reverse-index P1/P2/P3 materialization
- span-level authority_override detailed mechanism (WS-2 D4 → mixed-authority pilot)

3. Required specific checks (user-mandated)

evidenced_by_APR_blocking:
  blocking_ws3_logical_proof: NO   # alternative B used; appears only as APR-gated example
  blocking_production_compliance_matrix: YES
  position: APR-gated, NOT approved, NOT created (QG5/QG10). Parallel/conditional route.

governed_by / derived_from candidate dependency:
  blocking_ws3_logical_proof: NO   # falls back to assembly-local if inactive
  blocking_production: YES if candidates never activated (R-WS2-2 revisit before WS-3 exec)
  position: FLAGGED, dependency tracked, NOT self-resolved.

source_family_registry sufficiency:
  sufficient_for_ws3: YES — 9-family seed covers all 7 chain families + defaults + version/status policy.
  gap: physical registry table deferred (defer_WS2/schema cycle) — non-blocking for proof.

grammar_profile sufficiency:
  sufficient_for_ws3: YES — Profile A (incomex-architecture-constitution-v4) and Profile B
    (vn-national-law) defined with level_definitions + status_marker_rules + address_template_ref.
  gap: OD-G1 — actual internal Architecture Constitution fixture grammar (Nguyên tắc/Kiến trúc/Điều,
    no Chương/Khoản/Điểm) needs source/grammar ratification before real ingestion (defer_post_pilot;
    recommendation route C in report). Non-blocking for the logical proof.

authority_override sufficiency:
  sufficient_for_ws3: YES — WS-2 D4 unit|span override + provenance proved on node N2 (QG8).
  gap: span-level detailed mechanism deferred to real mixed-authority pilot (WS-2 D4) — expected, non-blocking.

entity_binding sufficiency:
  sufficient_for_ws3: YES — WS-1 G2 iu_entity_binding + WS-2 D5 entity_reference_registry
    (core 5 fields) + entity_kind_registry seed covered N4 (sql_entity) and N5 (code_artifact).
  gap: permission_policy_ref / snapshot_policy_ref deferred (OD-FA5) — non-blocking.

topic_registry_logical_only sufficiency_for_pilot:
  sufficient_for_pilot: YES — OD-FA3 Option C logical contract + mandatory provenance + free-text
    topic_code is explicitly the pilot path; physical topic_vocab deferred to P44-6 post-pilot.
  conclusion: topic registry logical-only IS enough for the WS-3 pilot proof.

4. compliance_matrix proof — EXPLICITLY DEFERRED (QG9)

ws3_scope_actual: cross_source_topic_thread / topic_thread for sample topic collection_structure ONLY
ws3_does_NOT_prove: compliance_matrix (full logical proof)

compliance_matrix_status_in_ws3:
  - WS-1 G1/G4 SKETCHED compliance_matrix as one of the 3 mandated profile sketches.
  - WS-3 may RE-REFERENCE that sketch (it does, in the report file) but does NOT open a
    detailed logical proof for it.
  - A full compliance_matrix assembly-profile proof is DEFERRED to a separate phase AFTER
    WS-3, or AFTER evidenced_by APR is clearer.

reason_for_defer (as required, verbatim intent):
  - compliance_matrix depends more strongly on evidence/compliance relations.
  - evidenced_by is still APR-gated and NOT approved.
  - proving compliance_matrix now would force assuming an evidence relation that has no
    authority yet — scope creep and a breach of the evidenced_by APR gate.

scope_creep_guard:
  - If any WS-3 artifact contained a FULL compliance_matrix proof it would be scope creep
    and MUST be cut (QG9). Audited: none of the 4 WS-3 files contains a full
    compliance_matrix logical proof — only an authority-respecting re-reference.

post_phase_note:
  - A future "compliance_matrix logical proof" phase is offered ONLY as a recommendation
    (report file, next-step option). NOT self-advanced.

5. Conflicts (flagged, NOT self-resolved)

conflicts_requiring_GPT_or_User:
  - C1 OD-FA4 (new edge vs APR): evidenced_by needed for true compliance/evidence
       durability vs "no edge created/approved" rule. → GPT/User APR decision.
  - C2 OD-G1 / OD-S1 (source authority): internal Architecture Constitution as the
       real fixture, its grammar, and KB-7294 rev44 / 📋 controlled-draft authority are
       unresolved (defer_post_pilot). Affects real ingestion, NOT the logical proof.
  - C3 R-WS2-2: if P44-4A candidate edges (governed_by/derived_from) are never
       activated, the constrains→governed_by and lesson_from→derived_from minimization
       resolutions must be revisited before WS-3 execution.
position: all three are FLAGGED only. WS-3 does not pick a side on any of them.

6. Readiness summary

ws3_logical_proof: PASS_WITH_BLOCKERS
blockers_are_for: schema cycle / production / pilot — NOT for the paper proof
must_resolve_before_schema: [BR-A1, BR-FA1, BR-FA3]
must_resolve_before_production: [BP-FA4 evidenced_by APR, BP-WS2-2 candidate edges, BP-FA8 executor code]
can_defer_post_pilot: [OD-FA2 phys, OD-FA3 phys, OD-FA5 deferred fields, OD-G1, OD-S1, OD-L1..L5, P11D-α/β/γ/δ/ε/ζ, reverse-index, span-override detail]
self_advance: PROHIBITED — recommendations only, no decision taken