ELNOR REPO READER TEXT MIRROR Original path: Memory Rebuild Docs/Flattening/Execution Ledger/Conflict Register/Conflict_Disagreement_Register.md Source repo: /Users/OpenClaw1/Elnor/Elnor Specs Git branch: main Git commit: dbaa25962edc11ab30e8d4ca1715f9ae5bf77331 Generated: 2026-06-09T01:23:58.539Z --- # DOC80 Memory Rebuild — Conflict / Disagreement Register **Repository:** github.com/wbrody/Elnor-Specs — branch `main` **Initialized:** 2026-05-25 (Stage 4) **Last updated:** 2026-05-27 (Stage 5R2b — DR-007 + DR-008 opened as matrix-row gap entries for `AlternativeExtractionRouting` and `AssertionCandidateEmission`, mirroring DR-002 pattern; immediately marked resolved with matrix-row addition deferred to next authorized matrix update. Plus: DR-003 / DR-004 / DR-005 resolution dispositions recorded per synthesis #8 + #2(b); DR-006 RE-OPENED with corrected ABC §7.8 enum values verified against source.) **Open conflicts:** 0 (8 resolved — 6 at Stage 5R2 + 2 added at Stage 5R2b) Per plan §10, a conflict row is mandatory whenever two agents, two slices, or two ledger sections classify the same concept differently, including disagreements over owner, function, status, tier, disposition, target landing, or preservation proof. **Stage 3 Codex review:** Will and Codex aligned on every finding after architect nuance on proposal status was applied. No disputes recorded. ## Open conflicts _(none currently open)_ ## Resolved ### DR-001 — SM-202 AssertionCandidateDisposition vintage misalignment (resolved 2026-05-25) **source_refs:** SM-202 (Supersession Matrix); ABC R0.2 §7.8; Concept Model §17.3; Adjudication Delta §1.6 **disagreement_kind:** supersession_disposition (specifically: which target-package vintage is senior) **position_a:** Concept Model §17.3 defines an 11-value `AssertionCandidateDisposition` enum (initial Stage 4 SM-202 row used this). **position_b:** ABC R0.2 §7.8 explicitly supersedes Concept Model §17.3 with a 7-value enum plus a separate `AssertionDedupeOutcome`; ABC §7.8 states verbatim 'do not allow multiple overlapping resolution/disposition enums'. **proposed_resolution:** ABC R0.2 is the senior structural patch on top of the Concept Model. SM-202 retargeted to the ABC §7.8 7-value enum + separate `AssertionDedupeOutcome`; Concept Model §17.3 is now lineage only. Recorded as a `semantic_change` migration with provisional preservation proof and architect_decision_ref ADQ-219 (CU/DOC73 alignment, which depends on the same target-package vintage discipline). **architect_required:** no (the rule was already in ABC §7.8; the agent had simply read Concept Model §17.3 as co-equal). Reviewer surfaced; patch applied; logged here as the lesson — when target-package files conflict by vintage, the structural patch (ABC) is senior to the conceptual baseline (Concept Model). **status:** resolved **architect_decision_ref:** ADQ-219 (related — CU/DOC73 alignment is the next instance of the same vintage discipline) **Implication for Stage 4+ work:** when reading target-package files, treat ABC R0.2 + Round D R0.2 as senior structural patches; the Concept Model and DAMS V5 Outline are conceptual context, not co-equal sources. Adjudication Delta is lineage when superseded by ABC/Round D. ### DR-002 — SM-040 (Library / Corpus intermixed) missing from Supersession Matrix (resolved 2026-05-26) **source_refs:** SM-040 (Supersession Matrix — referenced from `DOC80_Retired_Names.md`); Round D R0.2 §1.x; Stage 5R review §E-2. **disagreement_kind:** supersession_disposition (matrix gap — retired family without an explicit matrix row). **position_a:** Stage 5 round-1 Retired Names header claimed "22 entries" but the table had 21 rows; the Library/Corpus intermixed family was missing. **position_b:** Per Stage 5R review (Claude-synthesizer Part I §E-2): the SM-040 retired-name row must be added; the matrix gap must be logged here per the Stage 5 review boundary (do not edit the Matrix directly in Stage 5R). **proposed_resolution:** Add the SM-040 retired-name row in `DOC80_Retired_Names.md` (done at Stage 5R per F1; brings the table to 22 entries). The Supersession Matrix itself is not edited in Stage 5R — the SM-040 row will be added at the next matrix update authorization. **architect_required:** no (the resolution is documented and applied to Stage 5R artifacts; the matrix-row addition is an authorized future matrix update). **status:** resolved (Stage 5R artifacts patched; matrix-row addition deferred to next authorized matrix update). **architect_decision_ref:** (none — direct accept per Stage 5R F1 patch.) ### DR-003 — `VersionedClaim` matrix gap (resolved 2026-05-26) **source_refs:** `VersionedClaim` (DOC73 legacy); `DOC80_Owner_Map.md` row currently marked PENDING; Stage 5R review §B-5 + §F2. **disagreement_kind:** owner / supersession_disposition (matrix has no SM row for `VersionedClaim`; it is either a live object needing an owner OR a retired name needing an entry). **position_a:** Round 1 Stage 5 Owner Map referenced `VersionedClaim` but did not assign it a `schema_owner` and did not retire it; the Supersession Matrix has no row. **position_b:** Per Stage 5R review §B-5 + §F2: the disposition must be decided explicitly. Two paths: (a) retire in favor of `Assertion` + `AssertionVariant` via lineage table; (b) give it a `schema_owner` (likely DOC73 or DOC82). **proposed_resolution:** Stage 5R Owner Map row for `VersionedClaim` is set to PENDING with a pointer to this register entry. Stage 6 charter authoring (specifically the E3/E4 DOC82 charter and the DOC73 reconciliation per ADQ-219) resolves the disposition. Matrix gap is logged here; matrix-row addition deferred to next authorized matrix update. **architect_required:** deferred to Stage 6 charter authoring. **status:** resolved (Stage 5R artifact pointer in place; disposition resolved at Stage 6). **architect_decision_ref:** ADQ-219 (the DOC73 CU reconciliation will surface the `VersionedClaim` disposition). ### DR-004 — `DynamicHeaderLedger` vs ADQ-305 ambiguity (resolved 2026-05-26) **source_refs:** `DynamicHeaderLedger` (DOC80_Owner_Map.md); ADQ-305 (resolved Stage 4 — stable headers only if hash-pinned AND policy-invariant); Stage 5R review §B-6. **disagreement_kind:** owner (round 1 Stage 5 Owner Map placement disagreed with ADQ-305's resolution). **position_a:** Round 1 Stage 5 Owner Map placed `DynamicHeaderLedger` ambiguously (DOC84 owner with unclear contract). **position_b:** Per Stage 5R review §B-6: reconcile to ADQ-305 resolution. ADQ-305 specifies the stable-vs-volatile rule but does NOT explicitly name a `schema_owner` for the ledger itself. **proposed_resolution:** `DynamicHeaderLedger` is owned by DOC84 (delivery — the ledger is consumed at render time). The ADQ-305 rule (stable only if hash-pinned AND policy-invariant) is the operational contract; DOC84 owns the contract; DOC24 + KDA consume. If Stage 6 E7/E8 charter authoring surfaces further ambiguity about the ownership, escalate to architect. **architect_required:** no (ADQ-305 resolution is sufficient; the ownership assignment to DOC84 is the natural consequence). **status:** resolved (DOC84 owner with ADQ-305 rule recorded in `DOC80_Owner_Map.md` row). **architect_decision_ref:** ADQ-305. ### DR-005 — `WorkSession` unaccounted (resolved 2026-05-26) **source_refs:** `WorkSession` (legacy DOC72/DOC73 working-state); `DOC80_Owner_Map.md`; Stage 5R review §B-4 + §F2. **disagreement_kind:** owner / supersession_disposition (matrix has no row; round 1 Stage 5 did not assign). **position_a:** Round 1 Stage 5 did not list `WorkSession` in the Owner Map or Retired Names. **position_b:** Per Stage 5R review §B-4 + §F2: decide explicitly — assign to DOC83 (working-state) OR retire in favor of `WorkEpisode`. **proposed_resolution:** Stage 5R assigns `WorkSession` to **DOC83** (working-state) pending Stage 6 E6 charter confirmation. If E6 finds `WorkSession` redundant with `WorkEpisode`, retire it then via a future matrix update. Matrix gap logged here. **architect_required:** no (deferred to Stage 6 E6 charter — the charter resolves whether `WorkSession` is needed alongside `WorkEpisode`). **status:** resolved (Stage 5R Owner Map row assigned; final disposition at Stage 6). **architect_decision_ref:** (none — direct accept per Stage 5R F2 patch.) ### DR-006 — SM-202 carry-forward verification (RE-OPENED at Stage 5R2, 2026-05-27; corrected resolution below) **Re-open trigger (Stage 5R2, per synthesis must-fix #7):** Round-2 reviewers (Claude + Codex, independently) confirmed that the Stage 5R "verified consistent" check verified the *structure* of SM-202 (7 values + separate `AssertionDedupeOutcome`) but NOT the actual *enum values*. The parenthetical in SM-202 / Retired Names / Owner Map referenced `discard_not_memory / retain_ephemeral_observation / retain_session_note / record_friction_event / update_issue_frame / retain_as_evidence_only / create_or_merge_assertion` — these are paraphrases of the ABC §7.7 decision-tree *outcomes*, NOT the ABC §7.8 disposition enum. This is a transcription error, not an architectural reversal. **Stage 5R2 corrected resolution (verified against ABC R0.2 §7.8 source 2026-05-27):** The actual ABC §7.8 `AssertionCandidateDisposition` enum has 7 values: - `durable_eligible` - `durable_eligible_confirmation_required` - `session_scoped` - `ephemeral_reroute` - `candidate_only_pending_issueframe` - `review_queue` - `reject_not_memory` `AssertionDedupeOutcome` is a separate orthogonal axis: - `outcome` ∈ {`merge_to_existing_variant`, `add_variant_to_existing_assertion`, `create_new_assertion`, `reroute_misclassified`} - `dedupe_basis` ∈ {`source_span_duplicate`, `canonical_question_match`, `semantic_equivalence`, `authority_citation_match`, `manual_user_merge`, `no_match`} **Patched at Stage 5R2 (this round):** - `Supersession_Matrix.md` SM-202 row — corrected with actual ABC §7.8 values - `DOC80_Retired_Names.md` `AssertionCandidateDisposition` row — corrected - `DOC80_Owner_Map.md` references — consistent with corrected SM-202 The 4 truly-retired Concept Model §17.3 values (`retain_as_cu_component`, `create_directive_candidate`, `create_procedure_candidate`, `evaluation_only`) remain retired; their capabilities re-home via `AlternativeExtractionRouting` at DOC83 (B9 — renamed from `NonAssertionExtractionOutcome` at Stage 5R2b for semantic clarity; lineage-only retired entry recorded in `DOC80_Retired_Names.md`). **Lesson recorded as Stage 5R2 process discipline:** "verified consistent" verifications must enumerate the actual values, not just confirm the structural shape. Future round-N verifications cite the source values verbatim. **status:** resolved (Stage 5R2, 2026-05-27 — corrected against ABC §7.8 source). **architect_decision_ref:** ADQ-219 (related — CU/DOC73 alignment depends on the same vintage discipline); SM-202 (the row itself). --- ### DR-006 (superseded entry from Stage 5R, 2026-05-26 — kept for lineage) **source_refs:** SM-202 (Supersession Matrix); Stage 5R review carry-forward verification. **disagreement_kind:** verification (does the Stage 5 Owner Map / Retired Names match the Stage 4 Supersession Matrix SM-202 patch?). **position_a:** Stage 5R review §Carry-forward asked: does the Supersession Matrix SM-202 row reflect the ABC §7.8 7-value enum + separate AssertionDedupeOutcome per Stage 4 Codex review patches? **position_b:** Verification done at Stage 5R: yes, the SM-202 row was correctly patched at Stage 4 — *structurally*. The verification did NOT check that the actual enum values matched ABC §7.8 source; it verified shape only. This was insufficient — see Stage 5R2 re-opening above. **proposed_resolution (superseded):** No further matrix edit needed. **architect_required:** no (verification only). **status:** **superseded by Stage 5R2 re-opening above (2026-05-27).** **architect_decision_ref:** ADQ-219 (related); SM-202 (the row itself). ### DR-007 — `AlternativeExtractionRouting` matrix-row gap (opened + resolved at Stage 5R2b, 2026-05-27) **source_refs:** `AlternativeExtractionRouting` (DOC83-owned; renamed from `NonAssertionExtractionOutcome` at Stage 5R2b); `DOC80_Owner_Map.md` §4; `DOC80_Import_Graph.md` §2.1 / §5.1 Stage 5R2b footnote; Stage 5R2 audit-gap items 11 + 17. **disagreement_kind:** supersession_disposition (matrix gap — schema cited in Owner Map and Import Graph without a Supersession Matrix row). **position_a:** `AlternativeExtractionRouting` (under its earlier name `NonAssertionExtractionOutcome`) was created at Stage 5R per B9 / F-3 to re-home the capabilities behind the 4 retired Concept Model §17.3 enum values. It was cited in the Owner Map and Import Graph but never received a Supersession Matrix row. **position_b:** Per Stage 5R2b audit (items 11 + 17): every schema cited in the Import Graph or referenced by a supersession should be either a Matrix row or have its matrix gap explicitly logged (parallel to DR-002 for SM-040). Without that, the lint `supersession.import_graph_object_without_matrix_row` cannot be added at Stage 9. **proposed_resolution:** Mirror the DR-002 pattern. The schema row in `DOC80_Owner_Map.md` (line 103, after rename) is the artifact pointer; the Stage 5R2b footnote in `DOC80_Import_Graph.md` §5.1 names the gap and points to this register entry. Matrix-row addition deferred to next authorized matrix update. **architect_required:** no (the resolution is documented and applied; matrix-row addition is an authorized future matrix update). **status:** resolved (Stage 5R2b artifacts patched; matrix-row addition deferred to next authorized matrix update). **architect_decision_ref:** (none — direct accept per Stage 5R2b audit-gap fold-in.) ### DR-008 — `AssertionCandidateEmission` matrix-row gap (opened + resolved at Stage 5R2b, 2026-05-27) **source_refs:** `AssertionCandidateEmission` (DOC82-owned; DOC83 generates); `DOC80_Owner_Map.md` §1; `DOC80_Import_Graph.md` §2.1 + §2.2 + §5.1 Stage 5R2b footnote; Stage 5R2 audit-gap item 11. **disagreement_kind:** supersession_disposition (matrix gap — named E3↔E5 handoff object cited across artifacts without a Supersession Matrix row). **position_a:** `AssertionCandidateEmission` was created at Stage 5R per B9 as the named E3↔E5 handoff object (DOC82 owns the contract; DOC83 generates instances). It is referenced in the Owner Map and Import Graph but never received a Supersession Matrix row. **position_b:** Per Stage 5R2b audit (item 11): the same matrix-gap logging discipline that applied to DR-002 / DR-003 / DR-007 applies here. **proposed_resolution:** Mirror the DR-002 / DR-007 pattern. The schema row in `DOC80_Owner_Map.md` (DOC82 family) is the artifact pointer; the Stage 5R2b footnote in `DOC80_Import_Graph.md` §5.1 names the gap and points to this register entry. Matrix-row addition deferred to next authorized matrix update. **architect_required:** no (the resolution is documented and applied; matrix-row addition is an authorized future matrix update). **status:** resolved (Stage 5R2b artifacts patched; matrix-row addition deferred to next authorized matrix update). **architect_decision_ref:** (none — direct accept per Stage 5R2b audit-gap fold-in.) ## Schema (plan §10 AuditDisagreementRecord) | field | description | |---|---| | disagreement_id | stable ID | | source_refs | rows / sections being disputed | | disagreement_kind | owner / function / schema_shape / source_status / normative_weight / capability_tier / supersession_disposition / target_landing / preservation_proof / op_a_disposition / other | | position_a | first position with rationale | | position_b | second position with rationale | | proposed_resolution | optional | | architect_required | bool | | status | open / resolved / deferred | | architect_decision_ref | pointer to Architect Decision Queue item if escalated |