Skip to content

Structural proof: overcomplete factor-linear E8 decoders retain the puncture copy-cost floor - #1180

Merged
Mister-Meeseeks merged 1 commit into
CrocSwap:mainfrom
woklerx:research/sprint30-r40-s47
Oct 14, 2026
Merged

Mister-Meeseeks merged 1 commit into
CrocSwap:mainfrom
woklerx:research/sprint30-r40-s47

Conversation

@woklerx

@woklerx woklerx commented Oct 13, 2026

Copy link
Copy Markdown
Contributor

The punctured complex E8 copy-cost floor from PR968/PR1077 was stated for exactly
eight terms. This structural research proof extends the same lower bound,
max(72-p,56) for p<=24, to any finite number of active terms provided both factor
matrices remain in the live root code. More terms in that class cannot improve
the prior attaining curve or its best 7/13 ratio.

The new argument uses rank counting and handles the exceptional p8/m9 case by
a positive inverse-Gram diagonal contradiction. An exact nine-term factorization
with outside-code columns demonstrates why unrestricted overcomplete decoders
remain open. The inherited real-flat classification is explicitly retained.

Standard-library replay checks the source-bound predecessor geometric facts,
four complex overcomplete examples over all root pairs and 25 punctures,
1024 root-pair costs, 16 rational Gram certificates, a 12544-entry outside-code
identity, five mathematical controls and six package mutations. Reproduction,
proof, source pins, exact receipt and predecessor credit are in the package.

No physical decoder construction, unrestricted factorization lower bound, native
admission or new kappa is claimed. Only one new research directory is added;
selected PR1094 and upstream remain unchanged. Credit belongs to PR968/1077 and
their retained geometric lineage; PR765 and 1071 are distinguished explicitly.
Prepared for @woklerx with substantial OpenAI Codex assistance, Apache-2.0.

burntout76 added a commit to burntout76/integer-mult-bounds that referenced this pull request Oct 13, 2026
… in the rigidity package

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
@Mister-Meeseeks
Mister-Meeseeks merged commit 997cd27 into CrocSwap:main Oct 14, 2026
60 checks passed
@Mister-Meeseeks

Copy link
Copy Markdown
Contributor

100-PR batch review at 75b408646d558147a4eb250fc407d1fbfcb8df4e.

Merged to main, preserving the original contributor commits and package attribution.

Reviewed scope: Arbitrary-term factor-linear E8 copy floor via rank counting and exceptional inverse-Gram diagonal obstruction; unrestricted outside-code factors remain open.

The pinned scoped command passed; the batch ledger contains its command, input hashes, environment and log. This incorporates the research/tooling contribution and does not change the selected κ or discharge the inherited multiplication-theorem assumptions.

Batch review, attribution and evidence. Thank you for the contribution.

DreamingOfClouds added a commit to DreamingOfClouds/integer-mult-bounds that referenced this pull request Oct 15, 2026
…, factors in col M of a rank-8 realisation (P112 64, E8 72, P104 (CrocSwap#968's unit) 56; two-call batches 128) — no κ claim

Adds research/decoder-floors-frontier-format: rigidity of the minimal-rank label table (every fitting pairing of F2-rank <= 9 on E8/P112/P104 (CrocSwap#968's unit)/P111 and <= 7 on E7/E6 is the dense one; 16 DRAT-verified UNSAT; the P111 row is prior work of CrocSwap#620 and CrocSwap#1097), P112 >= 64 in every rank-8 realisation with factors in col M, <= 8 terms with factors anywhere, CrocSwap#1180's example (65) and the co-star dead sets, two-call and k-call batching bounds, the every-B minrank of the coordinate block, puncture clique-hitting bounds, and an explicit OPEN section (h >= 10 beyond the Gram root code, rank-9 square and outside col M with m >= 9, puncture windows p >= 46, batches of m >= 17 outside col M, other P104 units). Overnight closures from the author's working cards are RECORDED (not re-run): bit windows of 13-14 totals, the h = 10 Gram root code, p = 45 for m >= 8, and the projected-copy pricing; the frame-module theorem (PROVEN) is added as a tool. No kappa claim. verify.sh: 74 PASS on a fresh export.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
DreamingOfClouds added a commit to DreamingOfClouds/integer-mult-bounds that referenced this pull request Oct 15, 2026
…, factors in col M of a rank-8 realisation (P112 64, E8 72, P104 (CrocSwap#968's unit) 56; two-call batches 128) — no κ claim

Adds research/decoder-floors-frontier-format: rigidity of the minimal-rank label table (every fitting pairing of F2-rank <= 9 on E8/P112/P104 (CrocSwap#968's unit)/P111 and <= 7 on E7/E6 is the dense one; 16 DRAT-verified UNSAT; the P111 row is prior work of CrocSwap#620 and CrocSwap#1097), P112 >= 64 in every rank-8 realisation with factors in col M, <= 8 terms with factors anywhere, CrocSwap#1180's example (65) and the co-star dead sets, two-call and k-call batching bounds, the every-B minrank of the coordinate block, puncture clique-hitting bounds, and an explicit OPEN section (h >= 10 beyond the Gram root code, rank-9 square and outside col M with m >= 9, puncture windows p >= 46, batches of m >= 17 outside col M, other P104 units). Overnight closures from the author's working cards are RECORDED (not re-run): bit windows of 13-14 totals, the h = 10 Gram root code, p = 45 for m >= 8, and the projected-copy pricing; the frame-module theorem (PROVEN) is added as a tool. No kappa claim. verify.sh: 74 PASS on a fresh export.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants