I agree with all of Arthur's points of view.

> What is the summary regarding how we will proceed
> to get the reference documentation about SRFIs (and R^7RS?). >
> I think micro-data was chosen?

IMHO it's be nice to be able to extract both the full text and the
metadata (procedure names and argument lists, etc.) of all documents.
This can be done bit by bit - doesn't matter if the full effort takes years.

> They told me that usually, the devdocs is generated from existing
> documents that live in the web. So, I am wondering what to do from
> here, since I don't think SRFI documents qualify as user documentation.

Yes, they just scrape HTML. Same approach as Kapeli Dash (a Mac desktop
app similar to devdocs.io).

> I think getting the documentation on devdocs.io would be something
> extra. It would be best if we could get together something as snappy
> and easy to use as devdocs.io using Scheme :)

Fully agreed.

If you feel inspired to submit some Scheme documentation to devdocs.io
already, I don't see any problem with it :) I'd like to serve it from
the API at some point, but it will take 6 months or more to get ready.
If 6 months is too long for you, just go ahead and do it quicker :)

Generally, for everything in this project, if somebody is feeling
inspired, they can do a quick job with some hacky scripts and we can do
a clean implementation in full Scheme later. At least I can't think of
any aspects that can't be fixed later. If starting with full Scheme is
better for morale, I can try my best to do all my stuff in Scheme from
the start.