Finite bit construction: whole six-wire synthesis saves 38 further ADDs beyond PR936 (no new kappa)

research/r439-six-wire-boolean/CLAIM.json

@@ -0,0 +1,48 @@
+{
+  "status": "submitted",
+  "assignment": {
+    "pass_number": 13,
+    "assignment_seed": 19,
+    "drawn_at_utc": "2026-10-13T13:55:13.232283+00:00",
+    "track": "bit supplier",
+    "source": "secrets.randbelow(100), OS randomness"
+  },
+  "base": "47481c89fd30a43f8f882eb4b391e89c29273fd0",
+  "track": "bit supplier",
+  "candidates": [
+    {
+      "task": "Shared-frame target delivery factoring in phase B",
+      "decision": "rejected by census: every target-writing same-frame run has one target; positive z blocks contain no targets"
+    },
+    {
+      "task": "Whole six-wire phase-A Boolean synthesis beyond the five-wire windows",
+      "decision": "chosen: larger Boolean equivalence class than fixed positive-transfer PR925, whole-block rather than PR904/936 windows"
+    },
+    {
+      "task": "Five-wire phase-B window synthesis",
+      "decision": "deferred: closer to PR936 window method and many already treated B frames"
+    }
+  ],
+  "scope": "Classify the 60 six-wire phase-A blocks and nominate complete Boolean-equivalent replacements for the pinned PR936 word using bounded deterministic row/column Gaussian reductions. Admit only shorter programs dominated entrywise in both positive orientations; preserve frames and all original H1-H5. Compare inherited savings separately.",
+  "search": "All 720 pivot orders in each of row-only and mixed row/column elimination modes, with two fixed donor priorities, at each of 60 blocks. Assignment seed 19 rotates tie order; one complete scan and no restarts. Baseline and seed nominations saved; no optimality claim.",
+  "optimistic_cost": "A replacement of length m for a block of length L saves 2(L-m) ADDs including required literal C. Six existing wires at the unchanged frame; no helper, copy, paid move or recursive child added. Both-direction positive domination prevents a larger absolute prefix after composition.",
+  "predecessors": [
+    904,
+    925,
+    936,
+    891,
+    848,
+    933
+  ],
+  "coordination": "No pinned sprint coordination issue: non-PR issues are 650 (Discussions) and 913 (prose quality). Local claim used.",
+  "expected_deliverable": "Exact candidate or bounded negative census with independently checked local maps, full original interface replay, complete costs and native stage evidence for any construction.",
+  "expected_finish_utc": "2026-10-13T15:00:00Z",
+  "claimed_at_utc": "2026-10-13T14:09:38.019451+00:00",
+  "completed_at_utc": "2026-10-13T14:27:59.103198+00:00",
+  "pr_url": "https://github.com/CrocSwap/integer-mult-bounds/pull/952",
+  "mathematical_head": "2b1f30ade78ad5e9b8b8fb9bca80879ba5c7ba07",
+  "publication_base": "007664306fff84fbb1c9a1b4053b69f439764034",
+  "result": "Seven complete six-wire A blocks save 19 ADDs plus 19 required cleanup ADDs beyond PR936: 23,711 to 23,673, total saving 632 versus selected PR589. Five positive integer maps change; Boolean maps, both positive guards, frame paths, original H1-H5 and paid profile preserved.",
+  "validation": "Full receipt identical on ordinary CPython 3.14.4 and isolated optimized read-only CPython 3.12.3; all three complete words and native 5/9/11 stages pass. 144 bridge controls, 12 local controls, seven package tests. Selected-record and source-object checks pass.",
+  "overlap_refresh": "Through PR951; pinned PR936 head unchanged, no duplicate whole-six-wire result. New PR950 is historical central-factor tooling and PR951 is a recursive cyclic-inverse construction."
+}