Skip to content

Prove Hadamard column completion for two-prime odd alphabets - #7744

Open
petalas wants to merge 2 commits into
CrocSwap:mainfrom
petalas:research/w4-1118d-scoped-bound
Open

petalas wants to merge 2 commits into
CrocSwap:mainfrom
petalas:research/w4-1118d-scoped-bound

Conversation

@petalas

@petalas petalas commented Nov 18, 2026 •

Copy link
Copy Markdown
Contributor

Summary

Scoped proof. Let A be an n-by-(n-1) partial Butson matrix over the mth roots, with m = p^a q^b for distinct odd primes. If 3 <= n < pq and n is divisible by either prime, the normalized missing column uses only mth roots and preserves every supplied column. At any n >= 3, the same conclusion holds if projection to either prime-power factor preserves column orthogonality.

Below pq, divisibility forces every vanishing column inner product to split into cycles of the dividing prime. Projecting to that prime-power component preserves these cancellations. Reduction modulo the discarded prime then separates positive and negative roots in the recovered column. This contribution was produced by Nick Petalas and answers a subcase of the odd-composite question. Advance claim

Credit PR #5865 for the normalized column and prime-power sign results, Dugongue for atlas B097 and the pilot application, and Lam and Leung for the two-prime cycle decomposition. PR #5865 · Atlas B097 · Lam and Leung

Evidence

Both registered modes passed locally on CPython 3.12.15 using the standard library. Each checked six mixed-phase fixtures, 86 erasures, 1,574 recovered entries and their opposite signs. The controls show that mixed cycles can fail to project and that neither projection is guaranteed at n = pq. Each direct run took 0.03 seconds locally. Committed bytes and source history passed for all five package files.

Run from the repository root.

python3.12 -I -W error -B research/two-prime-column-completion/verify.py
python3.12 -I -W error -B -O research/two-prime-column-completion/verify.py
python3.12 -B scripts/research_package.py --package research/two-prime-column-completion
python3.12 -B scripts/check_research_commit.py --package research/two-prime-column-completion --check-source-history

The direct checks print PASS_TWO_PRIME_COLUMN_COMPLETION. The package and committed-byte checks print PASS_REGISTERED_RESEARCH_PACKAGE and PASS_RESEARCH_COMMIT_BYTES.

Limitations

The theorem concerns exact recovery of one column. Mixed-cycle row counts may require a different argument. Noisy recovery, physical execution, supplier pricing and independent human or formal proof-kernel review remain open. The selected cold replay, Lean and Docker were unrun. The proof establishes no new κ.

Open next questions

  • Resolve the eight-row case over the fifteenth roots. Start by dephasing the first column and using the one-three-cycle and one-five-cycle condition on every pair of column differences. A mixed vanishing sum by itself supplies no valid partial matrix.

Related open work

The live PR and Discussion searches, including the open PRs created by this account since the claim, found no open PR on the same missing-column alphabet question. The merged proof in #5865 leaves the odd-composite case open. PR #5865

CI notes

This PR carries the shared CI-doc and replay-gap sync used by other open PRs.

All three hosted checks passed on 7dc85513e8ee9b771a5856f027cf199936f85af0. The plan job and the historical-selected-replay-gap and two-prime-column-completion replays passed. Hosted run 44400163009

On the base commit, the local onboarding test test_current_ci_section_matches_selected_record failed because docs/ci-verification.md names the #6030 replay and omits make frontier7000-verify, which the selected record requires. The same test failed in the plan job on PR #7740. After the sync, the local current-status tests and registered historical replay-gap check passed. PR #7740 plan job

Merge danger

Two-way. The research result is isolated in its package. A separate commit carries the shared CI sync.

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