Skip to content

Certify the dimension-1 clean support minimum is 41 - #289

Draft
unpaidmoon wants to merge 1 commit into
cursor/upstream-e545235-base-cdfbfrom
cursor/dimension-one-clean-minimum-cdfb
Draft

unpaidmoon wants to merge 1 commit into
cursor/upstream-e545235-base-cdfbfrom
cursor/dimension-one-clean-minimum-cdfb

Conversation

@unpaidmoon

Copy link
Copy Markdown
Owner

Change and mathematical scope

On the fixed PR CrocSwap#5745 source, every nonzero clean relation among the 119 dimension-1 frames has cancelled frame support at least 41, and two explicit rays meet 41. This is an exhaustive finite census of the 28 rational rays whose projected weight is at most 40, together with an exact 6-span enumeration that proves the list is complete. It is not a heuristic search, a channel schedule, or a change to the selected frontier. The selected record remains κ = 363005817/100000000000.

The ten-channel certificate already gives 4 ≤ γ ≤ 41. This package shows that 41 is optimal inside the dimension-1 subfamily, so any shorter clean relation has to use a frame of response dimension 2 or more. Supports 4 through 10 stay open.

Task and overlap

Base commit: e54523517d0f03e58352e647c403a5af3e525a57.

Task card: joint-query-controller, scope refined. The card's free-injection search and the three-frame boundary are already certified. The actual question is the dimension-1 minimum left open by research/ten-channel-triple-span.

Nearest work: the ten-channel package (upstream CrocSwap#6295) proves no three low-grade frames meet the clean mode and exhibits one support-41 witness. It asks whether 41 is the minimum inside that kernel. This package answers that question and does not repeat the triple screen. A local overlap search for the dimension-1 kernel hit only that package. A direct issue search for ten-channel plus dimension-1 returned only closed CrocSwap#6295. The fork branch cursor/ten-channel-triple-span-9b63 is the predecessor, not this census.

Validation

python3 -B research/dimension-one-clean-minimum/verify.py
python3 -O -B research/dimension-one-clean-minimum/verify.py
python3 scripts/research_package.py --package research/dimension-one-clean-minimum --check-only

Both verifier commands printed PASS_DIMENSION_ONE_CLEAN_MINIMUM on CPython 3.12 with gcc, in about 96 seconds each. Metadata pins and check_research_commit.py --check-source-history passed. A C compiler (cc, gcc, or clang) is required for span_bound.c.

The census checks the orthogonal replay (15,752 contacts, 879 frames, 119 dimension-1 frames), the 14-dimensional kernel, denominator scale 52, six disjoint mode-zero supports, and the published witness of support 41 and mode (0,0,0,0,14,-14,-14,14). Projected rank over F_3 is 8. Exactly 28 projective rays have projected weight at most 40. Their cancelled supports are 41 (two rays), 50 (the weight-32 ray), 52 (one ray) and 53 (24 rays). The 6-span maxima on the three heavier zero-sets are 31, 31 and 33, all below 49, which is the completeness threshold.

Not run: make frontier7000-verify, the ten-channel triple screen, and any mixed-dimension or four-frame search.

Credit

The orthogonal census and the nine-channel pair theorem are Ali Romman's. The 41-frame upper bound and the three-frame obstruction are the ten-channel certificate's. Prepared by unpaidmoon with substantial Cursor Grok assistance. Apache-2.0.

Open next questions

  • Can a sum that uses at least one frame of response dimension 2 or more meet the clean mode on fewer than 41 frames? First step: one dimension-2 frame whose nuisance image meets a dimension-1 frame. The census here does not see mixed supports.
  • Is γ at least 11 on this source? First step: a four-frame screen on the 3,782 intersecting pairs. Supports 4 through 10 remain open.
  • Can either 41-frame mode be scheduled below 8 H_8 after transport and restoration? First step: price one witness under the PR Prove sharp 8H8 active-mode delivery cost with temporal reuse CrocSwap/integer-mult-bounds#5745 contract. No schedule is constructed.

Workflow feedback (optional)

The dimension-1 kernel splits into eight clean generators and six disjoint mode-zero supports, so the minimum-support question is a finite ray census rather than another triple screen. Naming that split in the ten-channel open question made the follow-up checkable in one package.

Open in Web Open in Cursor 

On the fixed PR CrocSwap#5745 source, every nonzero clean relation among the 119
dimension-1 frames has cancelled frame support at least 41, and two rays
meet 41. The selected frontier is unchanged.

Prepared by unpaidmoon with substantial Cursor Grok assistance.

Co-authored-by: unpaidmoon <unpaidmoon@users.noreply.github.com>
@cursor
cursor Bot changed the base branch from main to cursor/upstream-e545235-base-cdfb November 11, 2026 14:08
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