>     PS2 Speaking of archives, but otherwise totally unrelated: It has
>     occurred to me that some recent discussions about SRFI drafts were
>     based on intermediate Github versions and not on published drafts.
>     This can make it hard to reconstruct the reasoning in the future. It's
>     like the difference between citing a published paper and refering to
>     some internet page.
>
> This is one reason that I try to publish drafts as quickly as possible.
> I try to assume that anything not in the official SRFI Git repos or
> mailing lists will disappear someday.

Good point. GitHub should have a feature to export a dump of the issues
and pull requests. We should look into automating dumps on a schedule at
some point, just in case.

Probably the SRFI work based on GitHub PRs is either minor maintenance
without substantive changes (to avoid noise on the mailing list when
only correcting typos, grammar and the like) or done out of simple
fatigue (our workload per Schemer is quite high :)