Skip to content

Prove idle helpers do not lower the span-(m+1) line bound - #122

Draft
unpaidmoon wants to merge 1 commit into
cursor/upstream-626e936-pin-fc56from
cursor/span-lemma-idle-helpers-fc56
Draft

unpaidmoon wants to merge 1 commit into
cursor/upstream-626e936-pin-fc56from
cursor/span-lemma-idle-helpers-fc56

Conversation

@unpaidmoon

Copy link
Copy Markdown
Owner

Change and mathematical scope

Scoped proof, no new κ. In the abelian scalar-idempotent model, t invertible m×m rational targets plus any number of clean or dirty monomial helper registers still need at least m t paid calls when the frame span has dimension at most m+1 and the normalized extra frame satisfies |K| ≠ 2. The direct coordinate word attains m t and restores every untouched helper to the group identity.

This answers open question 3 of research/span-lemma-every-dimension: Lemma A's line projection survives identity helper targets. Extra registers do not escape that part of the three-register obstruction. The |K| = 2 flag case is not extended. The selected record stays 781451/250000000.

Task and overlap

Base commit: 626e936ce4d61f2584316914fc71baf96940c512 on canonical CrocSwap/integer-mult-bounds main. Priority revision 8.

Session r74, item 7 of the ordered eligible list:

  1. The 19 October handoffs (discussion comments 19030045, 19030047, 19030049, 19030053) are excluded as already delivered.
  2. Later structured handoffs are excluded when the same post is DONE, a later merged result answers them, or the canonical artifact is an open PR. The 31 October–1 November H25 requests (comments 19307764, 19308680, 19309328, 19311401, 19330344) are open PRs Archive H25 cache sources and remaining N22, P07 and P02 work CrocSwap/integer-mult-bounds#6321, Extract bounded H25 serializer with 140 finite controls and retained failure evidence CrocSwap/integer-mult-bounds#6326, Tooling: reuse same-call inventory SHA with retained checks CrocSwap/integer-mult-bounds#6328, Add authentic H25 serializer sources and retain first runtime authentication failure CrocSwap/integer-mult-bounds#6332 and Archive serializer source repairs and retained timing-review history CrocSwap/integer-mult-bounds#6335.
  3. Revision-8 "Next bounded decisions" then supplies six rows. Those rows name p104-side-correction, coverage-law-change and bit-helper-contract.
  4. Catalogue cards not named by those rows follow in catalogue order. Item 7 is bit-new-register-model.

Scope status: refined. The card allows a fixed larger register count. The alphabet stays the rational matrix group; the new hypothesis is r helper registers with arbitrary monomial outputs. The result is a lower bound, not a native word.

Nearest work, and why this is distinct:

Validation

CPython 3.12, standard library, Linux. From the repository root:

python3 -B scripts/research_selected_baseline.py
python3 -I -W error -B research/span-lemma-idle-helpers/verify.py
python3 -I -O -W error -B research/span-lemma-idle-helpers/verify.py
python3 -B scripts/research_package.py --package research/span-lemma-idle-helpers

Orientation printed PASS_RESEARCH_BASELINE_BINDINGS_ONLY with selected κ 781451/250000000. Both verifier modes printed PASS_SPAN_LEMMA_IDLE_HELPERS with idempotents_w3=28, idempotents_w4=120, pair_products=14400, eight_call_line_covers=0. The package command printed PASS_REGISTERED_RESEARCH_PACKAGE with outcome completed.

The pair check evaluates every 3×3 minor of (I+aN)(I+bM)-I at a,b ∈ {1,2,3,5}. Degree at most 3 in each scalar makes that grid an identity certificate for Lemma R's k=2 case. The induction in PROOF.md is the general argument. No Lean replay and no selected-frontier checker were run.

Credit

eumemic's PR #288 supplies the model. Ali Romman's PR CrocSwap#468 supplies the low-span lemma and the |K|=2 flag bound that this note does not extend. Nick Petalas's span lemma states the open helper question. The one-relation arbitrary-register theorem and the coordinate auxiliary check are retained as neighbouring scope, not as inputs that were replayed.

Prepared by unpaidmoon with substantial Cursor Grok assistance. Apache-2.0. No independent human or formal-kernel acceptance is claimed.

Open next questions

  1. Does the |K|=2 flag bound still give k >= m t for clean identity helpers? First step: recompute the dimension-two flag count at m=3 with t invertible targets and r zero blocks. Blocker: Lemma C fails at |K|=2, and the cited 2W bound assumes invertible blocks.
  2. Can one idle register and two independent frame relations beat m t? First step: fix m=3, t=3, r=1 and one explicit span-5 frame list, and decide whether an eight-call word exists. Blocker: the two-relation Laurent kernel is not principal.

Workflow feedback (optional)

The handoff thread's newest requests are H25 source archives whose canonical artifacts are already open PRs, so the revision-8 table plus unnamed catalogue cards is what still orders a non-overlapping mathematical task. Feedback discussion: CrocSwap#1200

Open in Web Open in Cursor 

Extra clean or dirty monomial registers leave the direct m*t call count
optimal whenever the rational frame span has dimension at most m+1 and the
normalized extra frame is not in the |K|=2 block. No new kappa.

Co-authored-by: unpaidmoon <unpaidmoon@users.noreply.github.com>
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