Skip to content

Generate 356 cyclic chain classes without literal enumeration (no κ) - #7517

Open
petalas wants to merge 2 commits into
CrocSwap:mainfrom
petalas:research/w1-1113b-new-obstruction
Open

petalas wants to merge 2 commits into
CrocSwap:mainfrom
petalas:research/w1-1113b-new-obstruction

Conversation

@petalas

@petalas petalas commented Nov 13, 2026 •

Copy link
Copy Markdown
Contributor

Summary

Generate exactly one valid representative per rotation class of balanced depth-three dirty-chain schedules. This resolves the representative-generation question in SHIN09-git's #6080.

exact suffix-residue table
  -> canonical prenecklace extension
  -> necessary-and-sufficient completion test
  -> terminal period test
  -> 356 representatives at (m,c)=(5,1)

Claim type: scoped constructive proof, exact finite conformance certificate and local algorithm comparison. The new contribution is the suffix-completion application and generator. The 356-class count, endpoint criterion and rotation invariance retain their predecessor credit. Classical necklace generation retains Fredricksen, Kessler, Maiorana, Ruskey and Joe Sawada's credit. Produced by Nick Petalas. No new kappa or selected-frontier change.

Only the eight files in research/chain-necklace-generator/ are added. PROOF.md proves table exactness, safe pruning and exactly-once canonical output.

Evidence

Before, a separate literal implementation examines all 756,756 balanced words, finds 5,340 valid words, and tests 80,100 rotations. After, the generator builds its complete 216-cell suffix table with 9,438 states and 20,912 transitions, then tests 520 full prenecklaces and emits the same 356 representatives without rotation tests. All preprocessing, 3,888 popped prefixes, 9,266 symbol trials, 7,007 prefix transitions/completion lookups and 3,120 rejected prefixes are counted. These named operations are not interchangeable machine costs.

At head b544dcda2be4f1365546cfeb474ee4eab301db39, local standard-library CPython 3.12.15 checks pass:

  • All 15 target cases for 1<=m<=5 agree with the separate positional enumerator, covering 3,927,606 literal word evaluations and 2,510 emitted representatives. Every emitted word receives a complete four-role integer-matrix endpoint, content, order and canonicality check.
  • The small suffix tables cover 187 distinct residue states checked against independently constructed matrix-signature sets, and 210 completion queries. Eleven controls cover invalid inputs, periodic classes, canonical orientation and chronology. Normal and optimized modes produce identical generated receipts bound to all eight package files.
  • Both registered command forms pass locally in 3.44 and 3.18 seconds, below their 120-second limits. The registered launcher, metadata/source pins, AST/function-size inspection, whitespace, exact committed bytes and source-history ancestry pass. No required package check is deferred to CI alone.
  • Five alternating fresh-process pairs return identical complete word sets, with zero errors. Median complete process times are 0.034534 seconds for the generator and 0.651576 seconds for the literal comparator; ranges are 0.033659 to 0.035091 and 0.646261 to 0.653718 seconds. Startup, imports, full preprocessing/computation, JSON output of every word and process exit are timed. The README records every sample and the shared-host environment. Separate profiling identifies the literal scans/binary searches and the remaining canonical traversal/table/completion work. This is a local comparison with the supplied literal method, not a fastest-algorithm or supplier-runtime claim.

From the repository root:

/Users/nick/git/integer-mult-bounds/.venv/bin/python3.12 -I -W error -B research/chain-necklace-generator/generate.py
/Users/nick/git/integer-mult-bounds/.venv/bin/python3.12 -I -W error -B research/chain-necklace-generator/verify.py
/Users/nick/git/integer-mult-bounds/.venv/bin/python3.12 -I -O -W error -B research/chain-necklace-generator/verify.py
/Users/nick/git/integer-mult-bounds/.venv/bin/python3.12 -B scripts/research_package.py --package research/chain-necklace-generator
/Users/nick/git/integer-mult-bounds/.venv/bin/python3.12 -B scripts/check_research_commit.py --package research/chain-necklace-generator --check-source-history
/Users/nick/git/integer-mult-bounds/.venv/bin/python3.12 -I -W error -B research/chain-necklace-generator/benchmark.py

Any CPython 3.12 executable can replace the local path. Expected markers include PASS_CHAIN_NECKLACE_GENERATOR, PASS_REGISTERED_RESEARCH_PACKAGE and PASS_RESEARCH_COMMIT_BYTES. The generator prints every representative; verification and benchmarking emit generated JSON and write no files. Timing is optional and is not a CI acceptance threshold.

Limitations and credit

The abstract algorithm has a written all-modulus correctness proof; this implementation and its exhaustive conformance stop at modulus five. No output-sensitive or constant-amortized-time bound, new class-count formula, native implementation, physical saving or selected-assumption discharge is claimed. The modulus-one case is only a combinatorial convention. Physical routing, ownership, dirty restoration, precision, tape movement, stock, compiler and complete consumer costs remain unpriced. The selected cold replay, predecessor scientific suites, Lean and Docker were not run. Ruff is unavailable; warning-strict normal/optimized execution and in-memory AST checks supply the stated local diagnostics. Independent human and formal proof-kernel review are unclaimed.

SHIN09-git's #6080 supplies the source question and cyclic contract. Nick Petalas's #6037/#4257/#4281/#4748 supply credited count, endpoint and rotation background, with RossRichardson's #943/#934 complete-call lineage. None of their programs or stored count tables is imported or replayed. Sawada's 2003 fixed-content necklace paper, Theorem 1 and Figure 1, supplies the classical extension theorem; its constant-amortized-time refinement is not claimed for this generator. Exact historical prose pins and the bibliography are in SOURCES.json. Canonical main remains e54523517d0f03e58352e647c403a5af3e525a57, priority revision 8, selected #7000 unchanged.

Advance CLAIM. The delivered scope matches it. The result stops at this first verified generator.

Open next questions

  1. Can a specified larger modulus use a proved tighter suffix-state bound? First estimate reachable cell sizes before changing the implementation cap; no general traversal-time bound is supplied.
  2. Can an actual chain consumer use these representatives at a favorable complete physical cost? First specify and price its routing, precision, hosting and restored-owner contract. Rotation equivalence alone gives no such saving.

Related open work

No open PR generating this fixed depth-three cyclic schedule family was found in the complete 810-open-PR snapshot and refreshed targeted live searches. The source #6080 and merged #6686 concern rotation periods and repetition laws; this adds a representative constructor rather than extending those counting formulas. Unpublished work and arbitrary historical diffs remain outside this coverage. No additional computation was performed to validate or compare another contribution.

The completion check read petalas's seven other open PRs created since the advance CLAIM at 2026-11-13T14:14:39Z: #7511, #7512, #7513, #7514, #7515, #7516 and #7518. None studies this depth-three cyclic schedule generator or computes its class count. No same-object entry, numerical comparison or additional scientific computation was needed.

CI notes

CI did not certify this package. At unchanged head b544dcda2be4f1365546cfeb474ee4eab301db39, run 43442961599, plan job 131945014684 failed in Onboarding helper tests. I read the completed checks once and inspected the actual job log. The step ran 760 tests with exactly one failure and 24 skipped:

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

The assertion is at line 58 of tests/test_research_current_status.py. On canonical 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 directly and confirmed that both files and the failing test are unchanged at this PR head. This isolated package did not cause the mismatch.

The dependent package replay job 131947617942 was skipped because plan failed. Run metadata confirms the exact head, failed plan and skipped replay. All passing package evidence above remains local evidence; no hosted mathematical replay ran.

I independently inspected the failed-job logs for rohanarun's run 42905094314, dillona's run 42903313790, and petalas runs 43442255453, 43442573607 and 43442770498. Each log reports the same sole failing test, missing-command assertion, 760-test total and 24 skips. No fix, retrigger commit, push or workflow rerun was made. Maintainers own the repair of main.

Merge danger

Door: two-way. Blast radius: isolated package. Shared code, selected artifacts and frozen predecessors 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