TopicsNextMeeting - math-comp/math-comp GitHub Wiki

Topics for the next meeting

Time:

  • France CEST(UTC + 2) 14:00
  • Japan JST 22:00 (when winter time in France)

How to join

https://rendez-vous.renater.fr/zw4q6-f243e-tsmlq

(if the meeting has not started yet, please wait a few minutes)

2026-11-17 meeting

  • Chair: leading the meeting, to be decided at each meeting for the next one, should send the reminder or cancel if no topics
    • by default: Cyril
  • Secretary: taking notes, alternative secretary can be decided at the beginning of the meeting
    • by default: Reynald

Topics

The topics of the next MathComp meeting are taken from this list. Please append your topic:

  • pick a secretary

  • (cont'd from last meeting) list our needs and limits w.r.t. AI (in particular to help with future PRs) Proposal

    As in Rocq and its satellite projects, AI assistance is permitted in math-comp projects. Contributors and maintainers remain responsible for their work:

    • Understand what you submit. You must review, understand, and be able to explain every contribution, including AI-assisted code, proofs, documentation, and reviews. We want to keep the project in the hands of humans.
    • Respect reviewers’ time. Submit focused, tested changes with concise explanations. Avoid bulk-generated PRs and walls of generated text. Maintainers may decline contributions whose review cost outweighs their value.
    • Be transparent. Disclose how you used an AI, not which agent or model you used.
    • Respect legal obligations. Do due diligence to ensure that your AI-generated code is not reproducing / directly derived from an existing code (or that its licensing terms are respected) and that you are allowed to include it in the Rocq project under its own licensing terms. Do not submit generated material whose origin or right to reuse is uncertain.
    • Consider environmental costs. Use AI deliberately and proportionately. Increased output is not a goal in itself; prioritize useful, maintainable work over volume.

    Satellite projects may adapt these guidelines to their maintenance capacity and technical needs.

    Human understanding, accountability, and sustainable development practices remain essential.

    Compared to the Rocq prover AI policy we may want to add the following:

    • If you translate a math-comp project, you are required to cite the original authors as the main authors of the development and you need their explicit consent to redistribute the code.
    • Moreover, if you extend a translated development, you are required to translate back your extensions.
  • Bits and Apery still depend on Algebra Tactics, but they shouldn't. Who will do the porting?

    • TODO
  • Decoupling norm codomain types

  • Analysis: rework/replace PseudoMetric with an actual pseudometric?

  • PR triage

    • #1627
      • provided Florent attends
    • #1647
    • #1623 and #1616
    • "That's enough! The end is near." ^_^ #1656 (start port of near to MathComp)
      • real-closed PR: #93
      • related MathComp issue: #1655
    • #1654 (quotient of left modules)
    • #1657 (using preorder for intervals)
    • analysis#2094 (open_nbhs)
    • analysis#1720 (Mertens theorems)
    • analysis#2086 (Rintegral_gt0)
    • creation of an algebra_classical file in classical in mathcomp-analysis
      • to include lemmas such as these
    • multinomials#116 Replace the infrastructure of mpoly with monalg
      • There are still some rough edges, but it's done, and it compiles. freeg.v is no longer needed.
      • It also eliminates the dependency on Bigenough, although the proofs are no longer written in the spirit of Bigenough. A smarter bigop mechanism that automatically infers the range from the body could further improve the situation.
      • The porting of CoqEAL is in progress.
    • multinomials#139 The evaluation function mmap in monalg.v should be generalised, but it's unclear how to do so.
  • bullet, indentation, controlling number of resulting goals... (again)

  • rewrite syntax for goal selection? (rewrite only main goal, all but main goal,...)

  • change the link for the visio (renater's service being discontinued next December)

  • MathComp sharing day on Nov. 30th?

  • pick a chair for next meeting

- any sensible way to introduce a disjoint union "operator"? (Re: https://github.com/math-comp/analysis/issues/1914)