Repository navigation
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Every legal row order of the fixed fifteen-pivot native graph needs at least 1,507 snapshot banks under the retained immutable-snapshot contract. The existing order from PR #7724 uses 2,106. The global minimum remains in [1,507, 2,106], with a 599-bank gap. This is a lower-bound proof, not a new schedule. No new kappa follows.
The package also proves a precise method limit: 1,507 is the strongest bound obtainable from the maximal universally live set, its rank, the full endpoint information rank and a separate result bank at one fixed row. All 7,643 anchors are covered. A stronger proof must use additional constraints, such as several boundaries or the actual complementary snapshots available. This is not a ceiling on the global minimum or on other proof methods.
Read the proof, 24-owner certificate, receipt, verifier, source pins and credits.
At row 88772, 24 snapshots must already exist and must remain for a read at or after that row, in every legal order. Their normalized forms have exact rank 19. Five explicit rational relations, including four equal-form pairs, prove the redundancy. The full endpoint has rank 1,501, so at least 1,482 other banks must retain the missing information. The new result bank must be separate before any donor read:
The finite-field ranks give rational-rank lower bounds. Only the named witness needs exact rational equality to prove the ceiling is attained; the receipt does not claim every other modular rank equals rational rank. Corrupting a relation or dependency path causes the named check to fail.
Python 3.12, standard library only:
Both normal and optimized commands print:
Registered normal/optimized replay, committed bytes/source history/modes, export verification and read-only sealed serial replay pass. The sealed export uses no Git and leaves all hashes and modes unchanged. Hosted CI did not certify this package. The plan job failed in the shared
test_current_ci_section_matches_selected_record: the base guide names frontier6030 while the selected record requiresmake frontier7000-verify. The log reports 760 tests, one failure and 24 skipped; package replay was skipped. This package changes no shared CI file. Local replay remains the scientific evidence.The contract retains all private entrance copies, distinct immutable token identities, exact donor lists, whole-row donor retention, result reservation before reads and common output installation. It excludes recomputation, in-place writes, merged equal forms, early output installation and public fields used as workspace. Only the snapshot component is bounded; source accumulator, scratch, guides, guards, child/adapter backing and parent storage keep their separate costs. Physical/caller replays, historical suites, selected cold replay, Lean and Docker were not run. All five framework interfaces remain inherited. No complete budget gain, global optimum or selected-frontier change is claimed.
Nick Petalas supplied the order and open question in the PR7724 source proof. Ali Romman's PR #7674 supplied the graph; PR #7681 supplied the storage contract. SYLforge supplied the word and caller; Clear1nsight supplied the comparator encoding and digest in PR #7682. All retained notices preserve predecessor credit.
The graph hash is
d81ec9f47740852a4bf7bd83100feff5cb66b70e1973466e1609bbac8af75098. The consumed PR7724 head isef52676eb20c252f5786f42a70ef0dd814400a03; the contract pin is1753a3ff56be34b09a217b12b590f51eba15defb. This PR starts from canonical maine54523517d0f03e58352e647c403a5af3e525a57. All 15 retained files are byte-pinned and unchanged. The graph reader and upper-bound allocator/replay are shared dependencies; no unmerged source ancestry is asserted.Prepared by Ali Romman with substantial OpenAI Codex assistance in research, implementation, verification and writing. This is source-aware author checking, not independent human or formal-kernel acceptance. Apache-2.0. Canonical claim and handoff.