Foundation-0's home for machine-checkable mathematics. As of 14 Nov 2026 (owner directive, PR #122), this repository is the consolidation point for three strands that are landing here:
| Strand | In this repo |
|---|---|
| Integer multiplication | Incoming. Audits and certificates will live here; background and links stay in f0 and fdn0.org/research/integer-multiplication/. |
| CrocSwap math | Incoming. No CrocSwap-specific tree in the repo yet. |
| Base math | Incoming. Shared numeric and proof scaffolding will land here; nothing dedicated is checked in yet. |
What is here today:
- AREAS.md — survey of where Foundation-0 can make progress in mathematics.
- data/codetables/ — census snapshot and script for open entries on codetables.de (see that README for files and how to re-run the census).
Other mathematics spin-outs are listed under Projects below. Work is in the open: results should be backed by something a machine can re-check, with credit for people and models.
A strand that outgrows a single tree gets its own repository. The new repo branches from math0 (or from the commit where the strand lived) and links back here for history and shared rules.
physics0 is the existing pattern: physics work started consolidated, then spun into focused repos (for example fractals and gravity) while physics0 remains the hub.
| Project | Question | Status |
|---|---|---|
| incentre-lattices | Where does one rule (join the incentres of a square's four triangles) lead, in any number of dimensions? | Preprint on Zenodo; not peer reviewed. |
| Integer multiplication (consolidating here; see table above) | Which claims in the race to lower the integer multiplication exponent hold up, and how far has each been checked? | Tracker feed paused since 11 Oct 2026; vault material incoming in math0. |
| formal-claims-audit | Does the formal (Lean) statement of an AI-claimed result say what its paper claims? | Not started. |
| covering-min-modulus | Does the record covering system with minimum modulus 42 check out, and is there one with 43? | Not started. |
| gfq-codes | Which open entries in the tables of linear codes over GF(3) to GF(9) can be improved? | Not started. |
| rigid-realizations (with physics0) | Can all 880 complex realizations of the 10-vertex rigid framework be real? | Claimed in rigid-realizations PR #14 (af32f535), pending independent verification: L880 all 880 real (r_2(10) = c_2(10) = 880; planar asymptotic lower bound 440^(1/7) ≈ 2.385811^n); L880 at Bartzos et al. lengths ≥ 868; G160 ≥ 132; L136 not certified (spec targets r_2 = 136, all real, at the repo's rational lengths); Prism_6 and L344 certificates not yet in repo; r_2(12) ≥ 4,128 (L344 glued with Prism_6 on an L344 triangle) conditional on a Prism_6 certificate built on an L344 face. |
- Certificates anyone can re-check. A generator matrix, a list of congruences, a proof file. Where possible, two checkers that share no code.
- Formal proofs are read, not only compiled. A proof checker confirms a proof proves its formal statement. Someone still compares that statement with the claim in words.
- Plain status labels. Each result says whether it is claimed, a preprint, peer reviewed, replicated or formally verified, and who has (or has not) reviewed it.
- Maintainers first. Before submitting results to a table or a problem list in bulk, ask its maintainer.
- Credit. Prior work and table maintainers are cited where they are used. Each project records which AI model did what.
From data/codetables/:
python census.pyUse python census.py --csv entries.csv to export the full entry list locally. Details are in data/codetables/README.md.
AI agents start with Foundation0's guide for agents, then a project's AGENTS.md and its open issues. People can say hello on the forum.
Directed by Ahraz Arifuddin
Researched and verified by different model families across Gemini, Claude, Grok, OpenAI, and Muse