Skip to content

SUB-241C-SPAN: native V3 sealing round taking up Lemma 24.1C (control plane, pre-F) - #988

Closed
amaybaum wants to merge 1 commit into
mainfrom
v3/sub-241c-span
Closed

amaybaum wants to merge 1 commit into
mainfrom
v3/sub-241c-span

Conversation

@amaybaum

Copy link
Copy Markdown
Owner

This is the guiding draft PR for the native V3 sealing round SUB-241C-SPAN (AGENTS.md §A.39). The planning issue is #986; the sibling round is #985 / #987.

State: pre-F draft, not yet a candidate F. The head is 542d8cac40526c1b409e976041536437ce45f522, a single-parent child of D = 0d917665 (main; verify run 40535620231, green).

  • The owner approved drafting on 2026-10-28.
  • Designating F, any exact-head dispatch and any merge each need separate owner authorization. This PR is not merged into main.

Content. The branch holds one file, verification/programmes/substratum/round-sub-241c-spanning-v3/preregistration.md.

  • v3 blocks: round SUB-241C-SPAN, kind sealing.
  • Record files:
    • the record directory and the receipt (AM);
    • the seal record verification/v3-seals/SUB-241C-SPAN.json (A, G12; this is the directory's first use).
  • Execution paths:
    • OIBridge/SpanningClassReach.lean (A);
    • OIBridge.lean, the census, ROADMAP.md and README.md (M).
  • Provenance:
  • R7-SCD → controls.py: each content check is mapped to a named failure code (table in the preregistration).
    • The ancestry and archive modes are not carried; the verifier's T1–T7 cover that ground.
    • There are no _SCD_* constants, no guard clause and no pin commit.
    • The legacy seal roles map to the receipt fields f, e and landing. A deterministic seal record is added at Q.
  • Frozen controls text. controls.py (blob bcdf6670…) is pinned in the preregistration and published for review on the linked design branch design/sub-241c-span-controls (855857f7). The verifier admits only the preregistration and amendments before F, so stage 1 adds the file to the record directory.
  • Discrepancies: X1–X12 are recorded. X8 is a new halt rule for SC5 cells that have no truthful frozen sentence, and it is flagged for owner review.

Owed before candidate F:

  1. Elaboration evidence for the eight frozen label statements and the header. This means one owner-authorized workflow_dispatch on a disposable branch, with an ill-typed countercontrol. Until then the X9 Lean transcription is unchecked.
  2. Owner confirmation of the X8 halt rule and of readings X3, X9, X10, X11 and X12.
  3. Independent review.

Local checks at the head:

  • controls freeze and controls start-state D: OK.
  • controls.py --self-test: 252 synthetic executions over the 54 outcome rows hold, and all 83 mutation controls fail with their named codes.
  • Lifecycle rehearsal on scratch commits: --verify-round Q → HOLDS. The rehearsal covered the halting and rebuild cases and 29 git-mode mutations.
  • v3_verifier self-test and corpus: OK. legacy_records_check: 303 intact.
  • Release gate: every step passes except lean-axioms (no local toolchain).

These are design evidence, not check-run attestations.


Generated by Claude Code

…ecision as a native sealing round

Takes up the legacy freeze lemma-24-1c-spanning-class-decision (blob b6f3837, merged by #643 as
e21a42c, never executed) as native V3 round SUB-241C-SPAN, kind sealing (issue #986). Adopts
SC0-SC6 and their rules by verbatim quotation; re-takes the start state at D; freezes the kernel
statement pins, surfaces, result-note requirements and the blob of the round-local controls.py,
which the first execution commit adds. Records the legacy-to-native vocabulary, the R7-SCD to
controls.py mapping and the seal-role to receipt mapping. Candidate F is not designated.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Y87wExth9RoUb3dtbumhTb

amaybaum commented Nov 1, 2026

Copy link
Copy Markdown
Owner Author

Owner decision recorded: pre-F release for retirement

The owner (amaybaum) gave this decision in the coordinating session on 2026-11-01, just before 17:35Z. It is quoted verbatim:

I release the pre-F SUB-241C-SPAN draft at 542d8cac for retirement to issue #986. No F designation, scientific execution or merge is authorized. Preserve the branch, commits, design controls, frozen sources and dependent evidence. Closure without merge is permitted only after the preservation checks pass.

The owner also stated: "These decisions release the administrative reservations. They do not authorize automatic closure or change the other research holds."


Generated by Claude Code

amaybaum commented Nov 1, 2026

Copy link
Copy Markdown
Owner Author

Preservation checks for retirement to #986 (read 2026-11-01, read-only). All checks pass.

The preservation results were independently re-checked: verdict confirmed, ALL-PASS.


Generated by Claude Code

amaybaum commented Nov 1, 2026

Copy link
Copy Markdown
Owner Author

Closing without merge, on the owner's decision (amaybaum, 2026-11-01): "#988: Approve closure without merge. Preserve both branches, especially the design branch containing the only copy of controls.py."

  • Preservation: all checks passed (comment 6561688613) and were confirmed by an independent re-check.
  • Branches kept, not deleted:
    • v3/sub-241c-span at 542d8cac40526c1b409e976041536437ce45f522
    • design/sub-241c-span-controls at 855857f773e2a37c7bd823df4c4c1f240771eee5, which holds the only copy of controls.py (blob bcdf66701b3383602477cd855674451db7f5102a). The preregistration pins its blob id but does not embed its text.
  • Open work: carried by SUB-241C-SPAN: native V3 round taking up Lemma 24.1C (spanning-class decision) #986, per the retirement plan there. A resumption requires a new guiding PR from a current D, citing this PR's control plane and commits as provenance (§A.43.6).
  • No change to: any F designation, scientific execution, or the standing research holds.

Generated by Claude Code

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