Skip to content

Research audit: tighten the packed-bit repair envelope (no new kappa) - #735

Merged
Mister-Meeseeks merged 1 commit into
CrocSwap:mainfrom
AryanSudhirDev:research/packed-guard-envelope-audit
Oct 14, 2026
Merged

Mister-Meeseeks merged 1 commit into
CrocSwap:mainfrom
AryanSudhirDev:research/packed-guard-envelope-audit

Conversation

@AryanSudhirDev

Copy link
Copy Markdown
Contributor

The original packed selected-bit proof bounds the correction fraction by min(1, 80(f-1)2^-K). This research audit derives an invariant carry/borrow guard envelope with exact density

1 - (1 - 2^(1-K) + 2^(1-2K))^(f-1) <= min(1, 2(f-1)2^-K).

The note proves agreement off that envelope, its invariance under both the ideal and actual permutations, and maximality within a precisely defined local no-escape guard class. It retains the eight rotations and existing correction argument. The factor of 40 applies to the conservative union-bound coefficient; this is not a runtime benchmark or a new kappa. The envelope is not the exact failure set, and the note distinguishes complete-rectangle density from fixed-source-fiber density.

The contribution is confined to research/packed-guard-envelope-audit/. It includes a written proof, source pins, exact receipt, and an independent standard-library verifier. Upstream sources and the selected construction are unchanged. This audits the original primitive, not the different compact-control primitive or the current supplier surcharge.

Validation:

  • 581,888 exhaustive packed states and 37,760 boundary/seeded states passed.
  • Four removed guard conditions, two broken rotation programs, and reversed correction order have concrete failure witnesses, all at K=6.
  • All 14 relevant existing tests and make selected-record-check passed; 2,240 comparisons with the pre-existing simulator agreed.
  • Full make verify and the separate selected-supplier cold replay were not run.

Reproduce from the repository root:

python3 -B research/packed-guard-envelope-audit/verify.py

Expected: PASS_PACKED_GUARD_ENVELOPE_AUDIT, with all evidence matching RESULT.json.

Base commit: 66849e1475c930fd7732b1a0067ffee7e1e95f60. Track: adversarial mathematical audit; OS-random assignment seed 62. No pinned sprint issue was present; the claim and overlap review are recorded locally in the package.

Attribution: Aryan Sudhir (AryanSudhirDev), with substantial OpenAI Codex assistance. The note credits the retained OpenAI #109 manuscript and the repository's preceding finite, packed-movement, and short-guard audits. New material is Apache-2.0. This is a written address-map proof with finite corroboration, not a Lean formalization or a full multiplication theorem.

@Mister-Meeseeks

Copy link
Copy Markdown
Contributor

100-PR batch review at 8d36139e49add1e55795e283b3bfa45e1cfe1bf0.

Merged to main, preserving the original contributor commits and package attribution.

Reviewed scope: Sharper invariant no-escape envelope for the eight-rotation map; envelope density is not the actual failure probability, and original streaming assumptions remain.

The pinned scoped command passed; the batch ledger contains its command, input hashes, environment and log. This incorporates the research/tooling contribution and does not change the selected κ or discharge the inherited multiplication-theorem assumptions.

Batch review, attribution and evidence. Thank you for the contribution.

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.

2 participants