Proposal: accept typst submission

[GiggleLiu]

I really love typst. I hope JoSS could consider adding support to typst submission, it features

1. Good support for programming languages.
2. Easy to create a template (if this proposal is to be considered seriously, I can contribute a template for JoSS). It has professional styling, comparable to latex. But much faster, smaller, easier than latex. It should not be difficult to create a new stack based on that.
3. It is an open source software.

[jedbrown]

Looks like Typst recently implemented PDF/A support, which was a blocker earlier. I like Typst, but I have mixed feelings about the cost/benefit of JOSS adding support. JOSS articles are relatively short and Markdown is mostly sufficient. Typst would have the benefit of

1. being an entirely self-contained system (the pandoc toolchain is somewhat complicated and many authors use some LaTeX features, which don't always interact well with markdown);
2. having a responsive, collaborative web environment (typst.app) for authors to write their papers (which I assume would still go in the repository for review).

[boisgera]

Looks like Typst recently implemented PDF/A support, which was a blocker earlier. I like Typst, but I have mixed feelings about the cost/benefit of JOSS adding support.

The cost/benefit probably isn't great, especially if the plan is to accept typst but keep the existing support for markdown. (For the record: I like typst too.)

JOSS articles are relatively short and Markdown is mostly sufficient. Typst would have the benefit of

1. being an entirely self-contained system (the pandoc toolchain is somewhat complicated and many authors use some LaTeX features, which don't always interact well with markdown);

2. having a responsive, collaborative web environment (typst.app) for authors to write their papers (which I assume would still go in the repository for review).

Point 1. would be a benefit only if typst was to replace the current markdown+pandoc toolchain, right?

[jedbrown]

Point 1. would be a benefit only if typst was to replace the current markdown+pandoc toolchain, right?

I meant point 1 not in terms of distribution (that's a factor too, but pretty moot if we stick to containerized), but in terms of consistent documentation and diagnostics/error messages. There is no single point of documentation for markdown+latex. For example, our official example paper uses \autoref to refer to a figure inserted with Markdown image syntax and \label. That "works" in the obvious way if one is familiar with both markdown and latex, but there is no authoritative documentation for such mixing and error messages are not good. If authors opted into Typst (presumably by putting paper.typ instead of paper.md in the repo when they submit), they could choose a consistent environment.