Skip to content

Prove all-width kind A removability and the separate-call conflict (no κ) - #7677

Open
aliromman wants to merge 1 commit into
CrocSwap:mainfrom
aliromman:codex/consecutive-result-022
Open

aliromman wants to merge 1 commit into
CrocSwap:mainfrom
aliromman:codex/consecutive-result-022

Conversation

@aliromman

@aliromman aliromman commented Nov 18, 2026 •

Copy link
Copy Markdown
Contributor

Result and scope

This package proves the kind A removability rule at every address width in the formal frame game. A part y=WD+cW gamma D can be removed from an admissible other-array last call T exactly when its two middle terms have the same frame and (-c)^ord(gamma)=1. Here c is nonzero rational, gamma is free, level(WD)=2 and level(TW)=2. The criterion does not depend on which admissible T is used.

PR #7629 gave this rule as finite evidence, including a spare-bit sample. PR #7649 proved the complementary finishing rule. Their opposite binomial tests now prove that a nonraw kind A part cannot satisfy both separate single-output jobs, even with arbitrarily many spare bits. This does not exclude joint decoding or general two-in-five programs.

Read the proof, exact receipt, verifier and scope record. The proof projects onto WP while retaining possible earlier-input singletons, then shows that they cannot repair the missing target coefficient or cyclic column. At singularity it gives the explicit relation using only nonidentity preparations. Four counterexamples with level(TW)=1 show that the call restriction is essential.

No new kappa, native supplier or selected-assumption discharge is claimed. All five assumptions and selected PR #7000 remain unchanged. This completes the advance claim on the narrowed formal-frame question of complex-rank-shape-decoder; no complete decoder is claimed.

Evidence and reproduction

Normal and optimized CPython 3.12.16 checks passed on macOS arm64. The verifier constructs all 10752 free elements at three bits, checks their symplectic pairings and uniqueness, and tests 43008 coefficient contracts by exact rational elimination. It uses complete literal coset sets for one fixed admissible removing call, with all earlier and paid columns retained. The 9983 nonplain cases reproduce the source counts 639, 767, 0 and 0 for c in {1,-1,2,1/2}. Those finite counts belong to the predecessor; the new contribution is their all-width removal proof.

It also checks 1407 explicit cyclic deletion relations, 150 contracts at four through six bits through a separate membership route, and five boundary controls. Package metadata, committed bytes, executable modes, source ancestry, the official seven-file export and its integrity check passed. An optimized read-only export replay passed without Git or a usable PATH, with all eight payload/manifest hashes and modes unchanged. No external run-time input is used. The proof and README were rendered with GitHub Markdown and checked.

From the repository root, with CPython 3.12:

python3.12 -I -W error -B research/kind-a-removability-coset/verify.py
python3.12 -I -W error -B -O research/kind-a-removability-coset/verify.py
python3.12 -B scripts/research_package.py --package research/kind-a-removability-coset
python3.12 -B scripts/check_research_commit.py --package research/kind-a-removability-coset --check-source-history --check-modes

Expected status markers:

PASS_KIND_A_REMOVABILITY_COSET
PASS_REGISTERED_RESEARCH_PACKAGE
PASS_RESEARCH_COMMIT_BYTES

Hosted CI did not certify this package. The plan job failed in the shared test_current_ci_section_matches_selected_record: the base guide names frontier6030 while the selected record requires make frontier7000-verify. The log reports 760 tests, one failure and 24 skipped; package replay was skipped. This package changes no shared CI file. Local replay remains the scientific evidence. Predecessor scientific replays, the selected cold replay, Lean and Docker were not run. Independent human and formal-kernel review are unclaimed.

Credit and remaining obligations

Prepared by @aliromman with substantial OpenAI Codex assistance. Jacob Sussman (@jacobalansussman) defined the removal problem, stated the finite rule and posed its all-width question in PR #7629, exact source. Nick Petalas (@petalas) supplied the coset method and complementary finishing proof in PR #7649, exact proof. The original frame context includes Jacob Sussman's PR #4628 and Dugongue's PR #4743. The needed geometric facts are derived in this proof. No predecessor code was copied. New work is Apache-2.0.

Kind B, joint decoding, post-call free processing, Pauli phases, native realization, precision and a full physical invoice remain open. The next useful question is whether the two output families can cancel jointly while preserving their shared preparation constraints. Base: canonical main e54523517d0f03e58352e647c403a5af3e525a57, priority revision 8. Live open/closed PR and Discussion overlap was checked; sparse local notes and unpublished work remain coverage limits.

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.

1 participant