Skip to content

Finite bit construction: fixed-prefix helper470 retirement and five cut obstructions (no new kappa) - #1256

Merged
Mister-Meeseeks merged 2 commits into
CrocSwap:mainfrom
RossRichardson:research/bit-seed08-2044
Oct 14, 2026
Merged

Mister-Meeseeks merged 2 commits into
CrocSwap:mainfrom
RossRichardson:research/bit-seed08-2044

Conversation

@RossRichardson

@RossRichardson RossRichardson commented Oct 13, 2026 •

Copy link
Copy Markdown

PR1128 leaves six rank-seven helper candidates at the fixed PR589 D/A/scatter cut. This contribution tests an explicit early-restoration mechanism: five candidates lose a required target dependency even at their least first restoration frame; helper470 admits a complete finite exception using rank two before rank seven.

Status: finite construction and scoped negative result. No new kappa or selected supplier adoption.

  • Keep D, A, retained copies, scatter and the original B unchanged. Insert three gates after scatter, replace one cleanup donor and omit one cleanup gate. Every one of the 661 symbolic input columns is correct in both scalar directions.
  • Helper470's frame dimensions change from 0 -> 1 -> 8 to 0 -> 1 -> 2 -> 7. The other 660 frame itineraries are identical. Complete ADDs are 24,305 -> 24,307, MOVEs 2,911 -> 2,912, and paid local rank mass 5,129 -> 5,128; all nine COPY/ERASE pairs remain.
  • Give exact target-access witnesses for all six single-rank-seven retirement failures, the five least-frame failures, and the working lower-frame exception. This does not exclude lower-frame preparation, later target-read cuts or other interfaces.
  • Price the explicit two-slot relocation relaxation with all returned/consumed idle rank. It is favorable at the pinned bit exponent, but no physical window assignment or new bit endpoint is claimed.

Ali Romman's PR1043 already deletes helper470 through an earlier source-boundary change and includes a hosted invoice. This does not improve that result or combine its savings. The distinct evidence is the unchanged-prefix endpoint and the six-candidate restoration classification. PR1128's boundary model, Hanson Sun's PR1079 census, Sagar Thakrar's PR1023 hosting incentive, and DreamingOfClouds/eumemic's PR589/555 lineage are credited. PR1179's kernel entrance is a separate mechanism.

Catalogue card bit-helper-contract: scope changed from the card's H2/K repair to an H4/terminal-frame investigation, after overlap with PR1194. The donor's bit-supplier assignment is retained. Task card and exact dependency pins record the reason and bounded domain.

Exact research base: 49b3157d1d89adcef16df8be98445451672be5ba; submission refresh: dadfea6e8a78ff6ba901705f87532a71c6f5bbb8, with identical input and the same selected PR1094. The sole live input has Git/SHA256 pins; external predecessor heads and historical context are in SOURCES.json. Seed8, bit-supplier track. Only research/bit-early-retirement-cut/ is added.

Proof, literal patch, certificate, scope and next steps.

Reproduce from the repository root:

python3.12 scripts/research_package.py --package research/bit-early-retirement-cut

Expected: PASS_REGISTERED_RESEARCH_PACKAGE research/bit-early-retirement-cut outcome=completed.

Local validation passed: registered Python3.12 replay; both checkers and eleven test groups on Python3.9.6; byte-identical receipt regeneration on Python3.14.6; selected-record integrity and all five selected-result tests; whitespace and source-pin checks. The two implementations use packed rows/Fraction geometry and set rows/fraction-free integer geometry. Tests include 22 star omissions, three retirement omissions, both cleanup errors, frame faults, source/receipt corruptions and read-only offline relocation. Hosted CI is separate; no hosted pass is claimed.

Unresolved: changed H4 cleanup admission, native 5/9/11-stage replay, actual owner/window assignment, packed precision/tape and outer/all-size obligations. The current complex supplier remains limiting. No full make verify, selected cold replay, or Lean theorem proof was run. Upstream and the selected frontier are untouched.

Smallest useful follow-up: test helper470's 0,79,186,513 path against one actual rank-seven idle window and the H4 cleanup contract before pricing a supplier. This is blocked by native/physical admission, not by the finite scalar identity. The README and task card also identify the exact later-target-read cuts for the remaining candidates.

Prepared by Ross Richardson with substantial OpenAI Codex assistance in reasoning, both implementations, tests and writing; Apache-2.0. Separate code paths do not imply independent human review.

@Mister-Meeseeks
Mister-Meeseeks merged commit 4848dbf into CrocSwap:main Oct 14, 2026
61 checks passed
@Mister-Meeseeks

Copy link
Copy Markdown
Contributor

100-PR batch review at d6d5a994fcdcd8d04ec439ff988daa8f37b17ece.

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

Reviewed scope: Complete661-column early retirement of helper470 and five fixed-first-block exclusions; local mass5129→5128 with extra ADDs, native/H4/owner admission still open.

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.

thakrarsagar added a commit to thakrarsagar/integer-mult-bounds that referenced this pull request Oct 14, 2026
…al review

- windows.py: match the nine-stage schedule event by event with the one the
  pinned global_k.py produces (recipe() and its idle-insertion rule extracted
  with ast and replayed); a dropped bridge is now rejected (new control).
- verify.py: max-flow brute force over every distribution of the 6k
  residuals (462 / 6188); record binding no longer writes the matched path
  and accepts a checkpoint record once the frontier moves; certificate bound
  at module level; CrocSwap#1281 priced rows split into local / pooled32 flags.
- test_package.py: real k = 0 tightness check; schedule-anchor control.
- VERIFY/SOURCES: declare global_k.py and the selected record as inputs.
- README/TASK: per-replica, frozen-recipe scope of CrocSwap#1281's open question 1;
  CrocSwap#1281's optimum transcribed (not re-derived), 148 rows priced here;
  overstatement about 1.5x (optimistic) / 1.7x (pessimistic); placement built
  at run time; CrocSwap#1256 relation; credits (chafreaky PR439/443/444, PR386/436);
  inherited assumptions; environment; feasible open-question first steps.
thakrarsagar added a commit to thakrarsagar/integer-mult-bounds that referenced this pull request Oct 14, 2026
- The synthetic witness is described exactly: its post-scatter reads undo
  W_syn's moved compensation, so after B it equals CrocSwap#1179's prefix entrance
  on PR589; it now also fails without its late reads (new control).
- The open triple's obstacle is scoped to A6R (B++C fixed); LC-fav also
  allows a rewritten B/C. 'Settles/decides' becomes 'narrows to one open
  triple'.
- PROOF.md: cone lemma is CrocSwap#1128 section 1 (Lemma 2 is the dual of section
  2); Lemma 1 states only x-free, pivot-free D' contributions; Step 4 folds
  scatter-extension snapshot terms into u; frames at sigma equal raw frames;
  reads during C reduce to the end of B (asserted: C at FULL, no target).
- Donors: 5,212 enumerated, 5,059 placeable. W_t are coordinate subspaces
  (recorded); the structural reason is restated accordingly.
- Relaxation checked to contain every admission; census sanity checks are
  labelled as such; flipped-byte control recorded; engine-2 coverage stated
  precisely; credits for eumemic, DreamingOfClouds, CrocSwap#967, CrocSwap#915/CrocSwap#926, CrocSwap#1043,
  CrocSwap#1079, CrocSwap#1256.
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