Skip to content

Prove orthogonal Schur residual rigidity under sector padding (no κ) - #7534

Open
petalas wants to merge 1 commit into
CrocSwap:mainfrom
petalas:research/w6-1113c-frontier-step
Open

petalas wants to merge 1 commit into
CrocSwap:mainfrom
petalas:research/w6-1113c-frontier-step

Conversation

@petalas

@petalas petalas commented Nov 13, 2026 •

Copy link
Copy Markdown
Contributor

Summary

Scoped written proof with exact finite corroboration, produced by Nick Petalas. For commuting orthogonal projections with fixed invertible compressed pivots A,B and compressed product H, sector padding cannot change the noncommuting Schur residual core.

T00=I-A-B+H, T10=A-H, T01=B-H, T11=H
existence: every Tj is positive semidefinite, n >= sum rank(Tj)
residual pair: one fixed core plus scalar pairs (0,0), (1,0), (0,1), (1,1)
consequence: one commutator rank and fixed nonzero singular values at every n

The positive-operator dilation and Gram-isometry method are classical. The new contribution is their explicit application to the prescribed-pivot Schur interface and the resulting obstruction to orthogonal rank tuning. At scalar pivots a=b=h=1, every orthogonal Schur pair commutes, even though the cited unrestricted maximal-rank theorem allows rank two at dimension five. For scalar effects generally, the exact commutator norm is sqrt(c(1-c)), c=(a-h)(b-h)/(ab).

Credit Krzysztof Wojt / krzysztofwojt #5128 for the interface, sector products, maximal-rank result and question; Nick Petalas #6031 and apoursh #6053 for the already-settled unrestricted rank frontier; and the Schur/sector lineage credited there. Their scientific executables were not imported or replayed. No literature-wide priority or new kappa is claimed.

Evidence

  • Before: the unrestricted oblique rank/dimension frontier permits extra sector dimensions to alter ranks. After: the written orthogonal decomposition proves that extra dimensions contribute only commuting scalar summands at fixed A,B,H.
  • Exact rational full-matrix checks admit 137 of 267 specified candidates; each of the other 130 has a singular pivot. Literal ordered Schur decompositions pass for every admission, with commutator-rank histogram {0:110, 2:25, 4:2}.
  • All 93 scalar identities, 32 padded/rotated realizations, explicit Gram-isometry intertwiners, the separate norm-1/2 fixture with core dimension three, and eight hypothesis/positivity controls pass. The fixture families are not exhaustive over all projections or compressed matrices.
  • CPython 3.12.15, standard library only. Both registered warning-strict commands pass locally; final direct check durations were 0.818 s and 0.882 s. These are check durations, not a speed comparison. Metadata/source pins, exact committed bytes and declared source history pass at d0777ae.
  • Official export integrity and normal/optimized replay from the read-only, Git-free export with PATH=/nonexistent pass, with byte-identical generated receipts. Receipts bind all seven package files; none is handwritten. AST compilation, function-size inspection and whitespace checks pass.

Exact reproduction from the repository root:

/Users/nick/git/integer-mult-bounds/.venv/bin/python3.12 -I -W error -B research/orthogonal-schur-rigidity/verify.py
/Users/nick/git/integer-mult-bounds/.venv/bin/python3.12 -I -W error -O -B research/orthogonal-schur-rigidity/verify.py
/Users/nick/git/integer-mult-bounds/.venv/bin/python3.12 -B scripts/research_package.py --package research/orthogonal-schur-rigidity
/Users/nick/git/integer-mult-bounds/.venv/bin/python3.12 -B scripts/check_research_commit.py --package research/orthogonal-schur-rigidity --check-source-history

Any CPython 3.12 path can replace the displayed interpreter. Expected markers are PASS_ORTHOGONAL_SCHUR_RIGIDITY, PASS_REGISTERED_RESEARCH_PACKAGE and PASS_RESEARCH_COMMIT_BYTES.

Advance claim. Base canonical main e545235 was refreshed unchanged before committing. The formal-interface-lemma card explicitly changes to this algebraic side theorem. Only research/orthogonal-schur-rigidity is added.

Limitations and unrun obligations

The proof assumes finite real or complex Hilbert spaces, exact compressed effects, commuting orthogonal projections and invertible pivots. Finite checks use real rational matrices. Minimum-dimensional orthonormal realizations over Q, singular pivots, indefinite metrics and approximate data are not covered.

No actual selected caller has been shown to impose this orthogonal contract. Physical preparation, native realization, dirty restoration, precision, stock, compiler, header frequency and outer transfer remain separate. Independent human and formal proof-kernel review are unclaimed. No predecessor scientific suite, selected cold replay, native replay, Lean or Docker was run. The selected frontier and its five framework assumptions remain unchanged.

Open next questions

  • Bind the self-adjoint hypothesis to one actual header/native caller, including its inner product and entrance/exit bases, before using this obstruction in a paid construction.
  • On a fixed effect-rank stratum with pivots bounded away from singularity, derive a quantitative perturbation bound for the residual core. No rounded-data stability theorem is supplied here.

Related open work

The closeout check inspected this account's seven other open PRs created since the CLAIM: #7527, #7528, #7530, #7531, #7532, #7533 and #7535. They study parity suffix norms, Euclidean resolvents, permutation layouts, block query dimension, source counts, shifted COPY coding and tensor splitting. None studies the fixed-compression orthogonal Schur residual pair or computes a competing value here. No additional scientific computations were run to compare their results.

CI notes

CI did not certify this package. At unchanged head d0777ae8e8b5ded706d0fa6d4ec6dc0fa7d8b13d, run 43468190113 failed in plan job 132032847132, step "Onboarding helper tests". I read the checks once and inspected the failed job log. It reports exactly one failing test among 760 tests, with 24 skipped:

test_research_current_status.CurrentCiNavigationTests.test_current_ci_section_matches_selected_record
AssertionError: '`make frontier7000-verify`' not found

The assertion is at tests/test_research_current_status.py:58. At base main e54523517d0f03e58352e647c403a5af3e525a57, certificates/selected-result.json names make frontier7000-verify, while docs/ci-verification.md still describes #6030 and make frontier6030-verify. I read those base files and confirmed that this PR leaves the selected record, CI document and failing test byte-identical to the base. This is an existing repository-level documentation mismatch.

The dependent replay job 132035445481 was skipped because plan failed. Package execution evidence remains the local checks recorded above. No fix, retrigger commit, push or job rerun was made.

I also inspected the failed logs of all five comparison runs below. Each reports the same sole failing test, missing command, and totals of 760 tests, one failure and 24 skipped:

Merge danger

Door: two-way. Blast radius: isolated.

One new research package and its opt-in replay. Existing scientific sources, credit and the selected frontier are unchanged.

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