Is it possible to find out how many people are subscribed to coq-club? (I'm mostly just curious. Also curious about download numbers for the various versions of Coq.)

FWIW, you can get the number of installations per Coq version for the last month from this page: https://opam.ocaml.org/packages/coq/

How are those counted?

Good question!

I wonder what fraction of these installs are on CIs

@Jason Gross did you look at this paper that does a breakdown of statistics of Coq-Club circa 2016? https://arxiv.org/abs/1609.07127

One can query the GitHub API to find out how often sources were downloaded, which is likely more accurate than the OPAM link

https://somsubhra.com/github-release-stats/?username=coq&repository=coq and more raw: https://api.github.com/repos/coq/coq/releases

nearly all our CI across coq-community these days does not download/build Coq

That's a really interesting paper, thanks!

I think from 2018 and onwards, it's not even interesting to compare emailing lists across communities any longer (or even analyze in isolation), since our communication platforms have completely fractured: GitHub, Discourse, Zulip, Stack Overflow, Reddit, ...

Jason Gross said:

Is it possible to find out how many people are subscribed to coq-club? (I'm mostly just curious. Also curious about download numbers for the various versions of Coq.)

At the time of writing, there are 1676 subscribed email addresses (probably with a few redundancies though).

Even if coq-club somehow keeps a reference role, its place among the different media, as @Karl Palmskog says, decreases.

Last updated: Jan 27 2023 at 01:03 UTC