Repository navigation
Certify all-subset exposure bounds for 273 fixed BIT planes - #4064
Merged
Mister-Meeseeks merged 2 commits intoOct 20, 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
For each of120 pinned E8 root lines and153 actual rank-two R438 frames, certify
dim(E)*nullity_F2(D_S) <= 8*|exposure_E(S)|for every subset S of438 helpers. Complete colored partitions turn finite certificates into this all-subset statement without enumerating2^438 subsets or trusting augmentation optimality.The new independent checker reconstructs the semantic111-by438 dirty map with reverse adjoints and causal masks with forward helper-seed propagation. It imports no producer mathematics. A separate proof review checks the counting argument and its exact model.
Scope and limits
One package,
research/bit-fixed-plane-exposure. It covers exactly273 specified planes, full semantic D, unchanged causal common-E promotion,111 target resources and nine copied-total resources, and private copied fields keptZERO. Both capacity-eight batch optima are zero, attained by S=empty. Eight is the least integer common coefficient: frame116 with all438 helpers has nullity327 and90 exposed resources, so654>7*90. This does not determine the best real coefficient.Eight is an auxiliary partition capacity, not the actual one-rank charge of a reached resource. The inequality is not net deficit or an all-rate rejection. Source paths, pivot promotion, owner credits and complete moments need separate bounds. Mixed E, cold-target quotients, new private cuts, changed words and arbitrary rational planes are outside it. No native selective supplier, rounding/tape implementation, selected-frontier change or new κ is claimed.
Historical individual-plane examples and other weights are retained as provenance, not silently included in this replay. The earlier paired-helper package uses a different private-at-E convention and has its own caller assumptions.
Test plan
304c4e0778d0bf864b7d09c163a837d7eb4fa455, including all five mathematical stages, integrity controls and unchanged inputs; no manual rerun requested.Sources and credit
Scientific pin
20ecf2891d80b5acf18d980b2164fb44ea7faa90. Retain Sagar Thakrar's R438 transport and the eumemic/DreamingOfClouds/Ali Romman E8/BIT ancestry. The causal promotion experiment and dirty-map construction are predecessor lane work; generic colored-partition counting is not claimed as a new theorem. Related compensation-kernel/paired-helper work retains the PR2857/2908 and PR386 lineage; PR4058 contains the separate fixed-word paired-boundary package.Prepared by Ross Richardson with substantial OpenAI Codex assistance and separate AI computational, proof and integrity reviews. Apache-2.0; all original bytes, source pins and notices retained. No independent human or formal-kernel certification.