T-236: Daniel reports Lean proofs of Fang's Theorems 1 and 2, `s(n² − n + 1) < n` for `n ≥ 15` — review tracking

Open
#784 1 comment 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
4/5
Estimated time
3-5 days
Newbie friendliness
28/100
Issue type
Documentation
Clarity
Clearly specified
Activity status
Active
Tech stack
python
Domain
testing, tooling

Research direction

Begin with the retained Lean source under lean/Sqpack, its README, certificate generator, and the theorem1/theorem2 modules referenced by the issue; also inspect the linked results.yaml entry for the recorded claim. The first executable step is to regenerate the 18 piece modules and build FangFinal one chunk at a time. Completion requires the stage 4 mathematical review to land, a successful reproduction build, and an axiom receipt for both theorems.

Written by the indexing model from the issue text.

Description

intake

This issue tracks the review of a registered result whose next rung is pending: T-236, credited to Daniel after Fang, Arslanov, Mustafin, Shangitbayev, Levy. The source is the Lean proof in lean/Sqpack in evand/square-packing, published on 18 October 2026 and reported on no issue. Kevin Fang's result that it formalizes, T-163, is tracked on #603. We opened this issue so that the state of this formalization's review has one public place.

The claim, from its register entry on main:

  • Evan Daniel's Lean 4 proofs of Kevin Fang's Theorems 1 and 2 (T-163), with no hypothesis: theorem1, s(n² − n + 1) < n for every integer n ≥ 15, and theorem2, that n² − n + 1 + p + 2q unit squares pack in a square of side below n whenever n ≥ 15 and 16p + 21q ≤ n − 15.
  • The proof formalizes Fang's: his Lemma 3, the gluing of Arslanov, Mustafin and Shangitbayev's decomposition; the induction on n; and the 18 pieces as exact rational certificates, each checked by decide +kernel in a generated module.
  • The source reports the axioms propext, Classical.choice and Quot.sound only, and a build of the pieces in about 19 minutes at 8 chunks in parallel. It publishes no build log.
  • At the counts this record holds, n = 15 to 18, the theorems give s(211) < 15, s(241) < 16, s(273) < 17 and s(307) < 18. T-163's construction and the case bounds improve on these, since the theorems state s < n and not the construction's side n − 1/20000.

Where it stands: V0/C0, significance S3, recorded. The proof stands as reported: nothing here has built it, and no review has read it.

  • The packet retains the theorem modules, the pieces' certificates and README, the piece-module generator and the library closure.
  • Fang's own repository narrowed its Lean statement on 19 October to Theorem 1 and the unmixed bounds of Theorem 2. This formalization states Theorem 2 in full; the bead think-494l compares that scope with T-163.

What is open:

  • C1: the stage 4 mathematical review of theorem1 and theorem2, their statements, the gluing lemma and the induction, and the 18 pieces, read from the retained packet. Owned by the bead think-57ei. Not priced.
  • V3 and C3: a build here of FangFinal from the retained bytes, with the 18 piece modules generated by the retained generator and built one chunk at a time, and the axiom receipt of theorem1 and theorem2. Owned by the bead think-dgrc. About 1.5 to 3 CPU-hours (priced 23 October), at most about 8 GB per chunk, and about 10 GB of disk. That would be reproduction with the producer's code.

Records. The case records for n = 211, 241, 273 and 307, which this entry leaves unchanged. We will post each change of this state here, and close the issue when the review lands.


Generated by Claude Code

Dominant language
Python
Stars
54
Forks
6
Avg merge
9h 11m
Merged PRs (30d)
239

Getting set up

  • No Dockerfile or Docker Compose file
  • Has a pull request template
  • No contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from jlevy/squares

All issues in jlevy/squares

Similar issues

More Python issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.