September 15, 2022 - Agenda

https://meet.jit.si/d8a744b413e339d95fba907b81fb249609520161

Editable Agenda: https://cocalc.com/projects/7586529a-3b95-46f7-bfdf-6bf154d03321/files/2022-09-15.md

Most discussion during the meeting was between: William Stein, Dima, Matthias, John Palmieri, David Roe

William

1. Wondering how or if I should get an Apple aarch64 macOS server.

• The Pari devs aren't asking for hardware, since they do NOT want to learn to or put any effort into maintaining macOS. They just want a remote shell to a well maintained "M1 apple server in the sky" where they can build and test Pari.

• There's also Isuru Fernando who works on the mamba-forge distribution of sage for M1.

• I could buy a Mac Mini (16GB ram max) or Mac Studio (32GB ram min) and host it at my house (I have excellent internet with unlimited data), then give accounts to anybody who needs them. One time cost of around $2K. Tricky part is ensuring a stable domain name as my ip address probably changes.

• Find some cloud hosting provider for M1's and pay a subscription fee. This one is a candidate https://www.scaleway.com/en/hello-m1/

• If anybody has thoughts, email me.

2. Has anybody ever tried to port Sage to WebAssembly? I.e., imagine this (which is 100% client side in your browser and has numpy, scipy, pandas, etc.) but with the SageMath kernel?

   1. start with sagemath-objects

   2. GAP? worry about not easy C, e.g., sbrk, low level garbage collection or use boehm...

Dima

Say a few words about state of migration to GitHub...

• in touch with open source relations people at GitHub, e.g., get free "Enterprise license".

• issue author mapping requires something more regarding licensing

  • maybe this is going to be in beta soon, e.g., auth tokens for users that target a repo

• voting process?

  • project needs to be "forward looking" rather than "backward looking"

  • personal pref of longterm devs is at conflict with this.

  • definitely should be a vote; should be no sooner than 2 weeks from now; when to vote depends on what the "sides" want.

  • William: longterm should we work out a more precise "governance system"? (modeled on Python, Jupyter, etc., or just what we do now but written down)

    • no urgent need to formalize such a thing since for a long time our current process has led to good outcomes.

    • need more new developers to provide balance

    • add need for new developers to proposal

    • encourage drive-by contributions! optics is good.

    • the release manager concerns are addressed, which is critical (e.g., CI)