tick-history: 2026-04-29T06:49Z — Aaron correction (human-lineage anchors stay) + 8-thread triage

docs/hygiene-history/ticks/2026/04/29/0649Z.md

@@ -0,0 +1 @@
+| 2026-04-29T06:49:10Z | claude-opus-4-7 | 26f978a2 | Aaron-correction + multi-PR-thread-triage tick. (1) **Aaron's mid-tick correction** reversed one direction of Amara's round-4 push to drop Conway-Kochen: *"we still need to link to human lineage so external observerse have a frame of references without fully understading our engineering"* (typos preserved). Synthesis: cite the lineage, do not dress engineering claims with it. Memory file landed at `memory/feedback_human_lineage_anchors_always_stay_attribution_for_observer_legibility_aaron_2026_04_29.md` + paired-edit MEMORY.md update. Round-4 absorb research note updated to honor both framings (anchors stay; metaphors don't prove). (2) **Multi-PR thread-triage**: 5 unresolved threads on #815 (2 P0 + 3 P1), 2 unresolved on #811, 1 unresolved on #818. All 8 triaged + resolved per the round-4 SNAPSHOT_MISMATCH split. P0 fixes: B-0103 dropped a literal incorrect example + handled multiple shard-name shapes (HHMMZ / HHMMZ-NN / HHMMSSZ-suffix) + NUL-delimited iteration. P1 fixes: B-0098 strict-POSIX example uses only POSIX features (case statement, [], printf — no bash [[ ]] / glob match). | [#811 POSIX fix pushed](https://github.com/Lucent-Financial-Group/Zeta/pull/811), [#815 P0+P1 fixes + 5 threads resolved](https://github.com/Lucent-Financial-Group/Zeta/pull/815), [#818 FORWARD_CROSS_PR_REFERENCE classification + thread resolved](https://github.com/Lucent-Financial-Group/Zeta/pull/818), [#819 round-4 absorb + Aaron correction](https://github.com/Lucent-Financial-Group/Zeta/pull/819) | Two durable observations: (1) **Aaron's correction reasserts the Beacon-promotion pattern at the rendering layer** — once a load-bearing rule earns external lineage, the lineage stays in prose for observer legibility; the rule that gets rejected is using metaphors AS engineering proof, not citing them as attribution. The two failure modes (proof-dressing + lineage-stripping) are both real; the rule that prevents both is "cite the lineage, do not dress engineering claims with it." (2) **The B-0101 SNAPSHOT_MISMATCH split is operationally validated this tick** — backward-stale (verify-and-resolve) and forward-dependent (encode-dependency) are now the two remedies in actual use; #815's threads triaged via both sub-classes correctly. The taxonomy split was filed as a research-note class refinement and validated as operational discipline within the same round. Cross-model consensus (round-3) prioritized; substrate verification (this tick) decided. |