Repository navigation
Exclude common-main coordinate coframes for query-thinned callers through C=.01 - #5706
Merged
Mister-Meeseeks merged 2 commits intoOct 26, 2026
Conversation
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The common-coframe decoder in Dugongue's PR5687 needs a new recursive invoice. This package proves that every coordinate common-main coframe assignment to the retained full nine-axis centre-free word has paid rank mass minus helper rank mass at least 2464. It also permits paid replacement warps. Eight exact integral flow/cut certificates prove the minimum, and one assignment attains it while leaving every native warp ordinary.
The complete consequence is negative: every p >= 1 and every prefix split in the standard query-thinned caller loses for supplier saving 0 <= C <= .01. Its target-minus-paid moment is at most
-(53126/125)(p+2)-112. At p16 the rank surplus is at most -10192. No new kappa, complete supplier or selected-frontier change is claimed. Different native words, non-coordinate Clifford coframes and different callers remain outside the theorem.The verifier rebuilds both complete source graphs and all 408 helper endpoint debits. It checks capacities, conservation and flow=cut directly, without running the discovery optimizer. The attaining invoice is
796 H1+348 H8+281 H9paid and405 H9+3 ordinary identity outputsreturned. No original helper is erased or double-counted. The all-p proof grants the ordinary decoder and all extra ordinary work for free, so the rejection does not rely on a new machine implementation.Registered normal and optimized checks pass, including six corrupted-certificate controls. The exact 16-file committed-byte check, official export integrity check, Git-free optimized replay with empty PATH, and an actual damaged-source rejection pass. Scientific replay commit:
c4cfb6aac99933fb6dd5454f4ad18b89bc9aa377; final documentation records that evidence. No full native/phase, precision, stock, tape, ordinary/outer admission, proof assistant or independent human review was run. Full-history ancestry is unavailable in the shallow checkout. Hosted CI is separate.Base and canonical refresh:
df0137efd35e11b4445734bf612bdd4c7ebc330c, priority revision 7. Scoped coordination separates this full nine-axis source from PR5698's restricted eight-plane optimum and active native/helper searches. Onlyresearch/common-coframe-coordinate-cutis added.Credit Dugongue PR5687 for the ordinary decoder; Cinna Davis, Dugongue and the retained centre-free source; Cao Yifan's native word; DaysSky, Nick Petalas, Ryanson Jonathan, zxv, Jacob Alan Sussman and all source notices. Discovery reused Ali Romman's PR5698 flow routine; the new verifier checks its certificates independently of that optimizer. PR5693 supplies a parallel general cut direction. Flow/cut duality and concavity are classical. New source specialization, witnesses and caller bound: Ali Romman, with substantial OpenAI Codex assistance and no subagents. Apache-2.0.
Hosted CI at final head: run 40286868279 ran 449 shared tests, with 21 skipped. The plan failed the two existing
CurrentCiNavigationTestschecks for the selected review and source-manifest links incurrent-status.md. This PR does not edit those files. The package replay job was skipped. The local scientific replay evidence above remains separate; hosted replay and independent review are not claimed.