IDEA: Suggesting a SRFI for Geiser and similar projects?

Hi,

I wanted to suggest proposing a SRFI to make supporting Scheme in text editors easier. Sort of like LSP but simpler and more focused on Scheme/based on S-Expressions. I couldn't find any previous discussions on a conversation like this, but if this has already been considered and rejected, my bad.

Although I have never wrote a SRFI before, I'd gladly take over that part. The reason I came here first, before consulting the SRFI mailing list, is that I think people involved with projects like Geiser have more experience with what would be needed in such a "standard". Or what is currently difficult, but could be made easier with a protocol approach.

My idea has been using a features-like approach to let different implementations define what capabilities they expose. So I'd guess that most implementations could manage auto-completion and loading files, and only some would support jump-to-definition and macro expansion, while debugging would certainly be too much. A tool like Geiser would use this information to implement completion-at-point, xref, etc., with minimal configuration needed for various implementations (from Geiser's side).

So, any opinions on this? It is feasible or necessary, or would it not be worth the effort?