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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
A scoped written proof and literal reversible table reduce the negative stochastic half's exact worst-case cost from
6W+2to4W+1transitions per direction for all signed widthsW >= 2. The bound is attained atx=0with either coin. The work tape uses the same six glyphs and one retained parity cell. Both heads return and the resident coin is read only.For
x=2q+p, the shift writes~qdirectly and the ripple adds1-pu. Their sum is-q-pu, so the table needs no separate negation scan. The written proof covers both directions, the exact inverse image and the timing bound. Proof and tableProduced by Nick Petalas. Credit ToikY's #6809 for the negative-half baseline and transition-constant question, Nick Petalas's #6801 for the marked shift, and zxv's #6092 and #6066 for the signed-half and restoration contract. Advance claim
Evidence
On CPython 3.12.15, both direct checks and both registered commands passed. Each direct invocation replayed 70 expanded rows, 4,088 exhaustive and 240 translated round trips, all 2,032 inverse boundary points through width eight, and five rejecting controls. The direct invocations took 0.22 and 0.21 seconds locally. Committed bytes, modes and source ancestry passed for all six package files.
Run from the repository root. On another machine use its CPython 3.12 executable.
The supplied marks and resident coin are part of the local contract. Optimality over other machines, physical encoding, coin delivery, native MOVE and a complete supplier invoice remain open. The selected cold replay, predecessor scientific suites and formal proof-kernel checks are outside this package's replay scope. The result is local and carries no κ claim.
Open next questions
Bind one reached negative half to this layout. Charge signed width, marker setup, coin delivery, retention of parity and return-state sequencing before comparing complete caller costs.
Related open work
CI notes
This PR carries the shared CI-doc and replay-gap sync used by other open PRs.
The sync updates the current selected-family CI section and runs the historical replay against its pinned checkout. All five supplied hashes matched. The six current-status tests ended OK, and the historical replay returned
PASS_REGISTERED_RESEARCH_PACKAGE. The new research package's two registered checks passed locally. All three hosted checks passed atbee924a7de5f3130f6f5d005d816986c2bf16482. The run completed the plan and both registered package replays forresearch/historical-selected-replay-gapandresearch/fused-negative-half-tape. CI run 44199622181Merge danger
The research commit adds one opt-in package. The separate sync updates the CI guide and the historical replay-gap package.