Scheme Review Lassi Kortela (23 Nov 2022 18:16 UTC)
Re: Scheme Review Jakub T. Jankiewicz (23 Nov 2022 20:14 UTC)
Re: Scheme Review Lassi Kortela (24 Nov 2022 07:39 UTC)
Re: Scheme Review Retropikzel (25 Nov 2022 09:53 UTC)
Re: Scheme Review Lassi Kortela (27 Nov 2022 07:15 UTC)

Re: Scheme Review Lassi Kortela 27 Nov 2022 07:15 UTC

Moi!

> I would also love to send some ideas for other people to discuss and
> even more I would like to read and learn from others people code and
> discussion. I think with time and good archiving it can become very
> valuable resource of knowledge.

Fully agreed.

SRFI's problems can be traced back to the fact the name says
"Implementation" but in practice it's a free-for-all venue.

What if most proposals started informally in a free-for-all forum, and a
subset of the finished proposals were promoted by authors of Scheme
implementations into the document series that is the current SRFI? At
the moment this is politically impossible, but it would almost certainly
be a better model in the long term. This would give control of SRFI to
implementers while Scheme users would still write the documents.

> Personally I would be glad if github was not the platform for this,
> in addition to not being open source there are social media elements
> and other features I dislike.

And Microsoft has an incentive to add more.

> Could codeberg (https://codeberg.org) be suitable? It uses gitea
> (https://gitea.io/en-us/) underneath and should have similar features
> to github regarding discussed use case.

R7RS-large is using https://codeberg.org/scheme/r7rs/issues

We have a self-hosted Gitea instance at https://gitea.scheme.org/.

Does Codeberg have important features that Gitea does not have?

The review process should probably have a website (similar to the
current SRFI site, auto-generated from git repos).

https://github.com/pre-srfi has no website, and it fails to get people
excited like SRFI does. The GitHub issue tracker (which seems almost
identical to Gitea/Codeberg) is good platform for discussion. But we
need some kind of reminder/announcement system of new drafts like SRFI has.

> I do agree tho, that something "easier" than mailing list would lower
> the barrier of entry. For example I have been programming for over 15
> years and consider myself quite a nerd but this is, I think, my second
> or third time sending anything to mailing lists. Mostly I'm worried I
> mess something up. Nice GUI would solve this "fear". :)

Point taken. SRFI uses email since it's the least common denominator,
but few people truly like email.

The ability to preserve the discussions would be important. For that
reason, I'd prefer something self-hosted. We're slowly building out
scheme.org and we're at liberty to install any software there that helps.

> I am very envious of clojuredocs (https://clojuredocs.org/). It works
> by having index of all the Clojure core functions etc. and allowing people
> to post their own examples of it. If there was place to discuss scheme code
> and ideas. Maybe in the future it could be expanded to have all the rNrs
> procedures
> so people could show their own examples. This would be very handy for
> newcomers
> to read.

Yes - I use clojuredocs too. Arvydas is building a similar site at
https://index.scheme.org/ but it doesn't (yet?) have usage examples.
Maybe you can convince him to support them :)

The original "documentation with user-submitted examples and
clarifications" site was probably PHP's back in the day. PHP was a total
mess, so it was a godsend.