Repository navigation
Prove exact canonical-checkpoint costs for cyclic output frames (no new κ) - #4860
Merged
Mister-Meeseeks merged 2 commits intoOct 23, 2026
Merged
Conversation
This was referenced Oct 22, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Scoped proof with finite corroboration: solve the canonical-checkpoint question left in Ross Richardson's merged #1287. The fixed cyclic-equivariant child contract and its aligned recurrence are cited premises. No predecessor result is audited or replayed.
An explicit schedule attains every prescribed-cut optimum. With A=3, B=1 and five children, a canonical cut after child 2 costs 8 boundary units, retaining the entire 15-to-8 saving. A cut after child 1 costs 9. Canonical batch frames do not imply completed source-only partial echoes.
Only
research/cyclic-checkpoint-cost/is added. This contribution was produced by Nick Petalas. Credit Ross Richardson #1287/#1145, Sagar Thakrar #1227, mightyOil/Philipp #1203, Hanson Sun #1488, Cinna Davis #1391 and the retained restoration/echo lineage.Evidence
Before: #1287 leaves prescribed canonical checkpoints open and supplies a three-state recurrence. After: a quantified closed form, arbitrary-cut optimum, sharp surcharge and emitted attaining schedules, with a generated source-hash-bound receipt.
Local exact checks pass 16,320 cut-mask/price optima, 9,216 one-cut pairs, 72 rational-price cases, 756 literal fixtures and 10,962 arbitrary-dirty basis/rational probes. Fourteen input/word rejections and three formula controls pass. Warning-strict normal and optimized checks each take about 1.1 seconds, below the registered 120-second limit. Both registered commands pass on CPython 3.12.15 / macOS 27.0.1 arm64. Committed-byte/source-history preflight, official export integrity and an optimized Git-free replay of the final commit with empty PATH pass. All exported bytes are preserved. A relocated optimized replay passes; changed-source, changed-receipt and changed-formula copies reject and preserve every byte.
From the worktree root, exact reproduction commands:
Expected markers:
PASS_CYCLIC_CHECKPOINT_COST,PASS_REGISTERED_RESEARCH_PACKAGE, and committed-byte/source-history success. Elsewhere substitute the local Python 3.12 path.Limitations
The theorem concerns fixed indivisible order-three children, uniform equivariant readers, symmetric nonzero-power prices and read-only canonical batch cuts. It does not cover adaptive mutation, opened child bodies or completed external-source partial echoes at those cuts. The inherited child/alignment contract and six-ADD rotation upper bound are cited premises. Physical caller, routing, coefficient access, precision, native hosting, proof-kernel validation and complete multiplication obligations were not run or discharged. No admitted supplier, new kappa or selected-frontier change is claimed.
Open next questions
Related open work
Refreshed 2026-10-22T22:28:57.032236+00:00. Read this account's ten open PRs created since the CLAIM at 2026-10-22T22:14:16Z. None studies this same order-three canonical-checkpoint optimum. The merged #1287 is the direct predecessor; merged #1488 and #1391 treat different transport groups. Search coverage does not include unpublished work or arbitrary diffs.
CI notes
Both hosted checks passed at
e36ff5b6431d3014b9ca0a1a49a5274d82788252:planand the registered package replay. Successful CI run. Local registered normal/optimized checks also pass. CI verifies these registered checks; it does not establish proof-kernel or complete multiplication-theorem acceptance.The card's historical baseline command fails with
KeyError: bit_loweragainst the selected record; the current selected-baseline source-binding positive control passes. This is an existing orientation mismatch, not a package check failure.Merge danger
Door: two-way. Blast radius: isolated. Only the new six-file research package is added.