T-spec meeting 2024-02-01

--- title: T-spec meeting 2024-02-01 tags: ["T-spec", "meeting", "minutes"] date: 2024-02-01 discussion: https://rust-lang.zulipchat.com/#narrow/stream/399173-t-spec/topic/Meeting.202024-02-01 url: https://hackmd.io/idaAR8RbSzus5XyKdeBgoQ --- Attendees: Connor Horman, Joel Marcey, Eric Huss, Lukas, Mara, TC, Monadic Cat, pnfelix ## Potential Agenda - Feedback from the lang team - Connor's XML example - Next week's meeting ## Lang Team Feedback [Raised](https://github.com/rust-lang/lang-team/issues/250) in a triage. Certain axes that they care about that is emphasized in the specification. The members decided to have a design meeting with an outcome being a document that goes beyond feedback of sample chapters but heads into what t-lang is looking for into the spec. t-lang wants more direct guidance from t-spec on what type of feedback to provide. pnfelix will manage the document to provide the t-lang team with input from the t-spec team. e.g., emphasize differences between versions of Rust Joel suggests that proposed outline of the specification may be helpful in helping guide the conversation. The t-lang design meeting may not be for a few weeks. ## XML Example Connor made an XML example of the input that can be easily translated into HTML. Example: <https://github.com/chorman0773/rust-lang-spec/tree/arrays-xml-rewrite> <https://github.com/chorman0773/rust-lang-spec/blob/arrays-xml-rewrite/spec/lang/exprs/index.xml> Used VSCode as the editor. Integrated the schema and was able to get some dev tool help on writing it (e.g., autocompletion). Eric prefers markdown with simple extensions as needed. TC described Akin's Laws of Spacecraft Design to highlight that writing tooling as part of writing a specification means we would now have a program for writing tooling to write a specification. Mara: Easier to be flexible with types of tooling with less authors. pnfelix: Agree that enabling the primary authors is a consideration, but the safest way to protect the future, is to create something that works with the bottom line XML. Folks can migrate to the current snapshot of the XML and then do whatever they want tooling-wise. TC: How unhappy are we with mdbook and with the Rust Reference? Looking at this from resources available, we don't have ten editors, so how should we balance starting fresh with actually creating content? Eric: I've been thinking along similar lines. I'll show an example of what can be done with an extension to mdBook. pnfelix: Do you anticipate a specification outline to be different than what the Rust reference already provides? And can they be merged? Joel: If we are planning to go down this route of basing things off of existing content and tooling, should we consider the Ferrocene specificaiton in this discussion? Connor: At what point does it become a diminishing return to structurally change the format of the Rust reference vs. writing content. pnfelix wonders if the Reference and specification ultimately remain two separate entities. Eric says that once we have a happy specification, the reference could potentially be deprecated. ## Next week's meeting Will begin 5 hours later than usual to accomodate JF Bastien from Woven by Toyota (in Japan) who would like to talk about safety critical Rust and how it might relate to the specification.