Currently the Typst community uses Discord for design discussions. For example, the CONTRIBUTING.md states:
Before starting significant work on a feature or refactoring, please find/open an [issue] or start a thread in the [#contributors] channel on Discord to discuss the design. Don’t be shy! Typst is a complex project with a long-term vision and it’s frustrating to find out that your idea does not align with that vision after you have already implemented something.
I believe that it is a terrible idea to use Discord for design discussions, or any important discussion that has lasting value and should be preserved.
Discord is a walled garden and some people (myself included) will not use it. (There exists synchronous-communication chats that are more open by design, and are also open-source, notably Zulip. These days even Discuss offers a live chat.)
In a few years Discord will enshittify, people will move to the next shiny chat platform, and all the content produced there will be lost.
Years of design discussions, debates, considerations are going to be lost in a few years when people move off Discord. Troves of information for the early design of Typst will be lost.
The middle-term solution would be to migrate design discussions, and any other discussion with lasting value, to a web-open platform that is publicly accessible, searched, indexed, and accessible by default without login.
The short-term workaround is to archive the content of the discussions on Discord and make them publicly available, archived forever for posterity.
When could the Typst team fix this choice?
(Note: I understand that Discord is a good place to have a community, for example beginners asking entry-level questions, because there is a critical mass of people using Discord, as IRC in the old days. This is fine – or at least this is a perfectly reasonable choice of going where the users are. But there is a large difference between a support channel among others being available on Discord, and the place to discuss Typst design and its evolution being on Discord. By all mean, keep a Typst thing on Discord if you like, as long as important things for the community are discussed elsewhere.)
While I see your point, practically I think live chat archives are not a good place to record large-scale design discussions regardless of platform choice. Fast chat is nice to get a bit of quick back and forth about a topic, but to retain a discussion in a way where other people can properly catch up and take part in the conversation, in my opinion it’s necessary for information to be synthesized from time to time.
In practice, this kind of happens already: Results from Discord discussions frequently end up in GitHub issues or PRs over time, often in a more compact, digestible form. Since you aren’t on the Discord, I should also mention that in practice it’s not that a ton of design discussion is happening there in the first place. There are discussions from time to time (and there have been more in the past), but often it’s mostly just some coordination or discussion that feeds back into things happening on GitHub.
I’m open to clarifying and shifting the documented norms a bit to reflect what I mentioned above, but I’m not really convinced another live-chat platform is the solution. And I imagine making the chats from the Discord search-engine accessible is also not something people would be very happy about as it’s not what people expect from being on a Discord.
I agree that live-chat logs are not the best way to present information for posterity, and that summarization processes to consolidate contents in more structured forms (RFCs, etc.) are important. (I myself prefer longer-replies forums such as here because the structure gap is narrower.) But my experience in other communities is that this summarization only spontaneously happens for a small fraction of the content, and other important discussions remain in the initial form they happened, so archiving is complementary/useful.
Looking for references to Discord discussions on the current forum returns more than 50 matches, which I take as a sign that there a significant amount of content/discussions that will become harder to understand/follow if/when Discord later becomes unavailable. (One notable example is your own summary of types / custom elements in 2024, which suggests that the best place to learn about the current design discussion on an important upcoming feature is a Discord thread.)