The Coq mailing list

["Benjamin C. Pierce" <bcpierce AT cis.upenn.edu>]

Saving up the announcements and releasing them simultaneously once a week would make them less intrusive.

This seems like a pretty good stopgap measure.  

As I said in the discussion on the types list (which probably most people here read, or at least ignored :-), I think the best way to handle this issue is simply to ask posters of announcements that are not obviously related to the list topic to include a little paragraph explaining the connection.  The only downside is that this requires a step that most mailing list software doesn’t support automatically: rejecting a message with an explanation and an invitation to resubmit after revision.

   - Benjamin

---

From: Hugo Herbelin <Hugo.Herbelin AT inria.fr>
To: Derek Dreyer <dreyer AT mpi-sws.org>; Perry E. Metzger <perry AT piermont.com>
Cc: coq-club AT inria.fr
Sent: Saturday, June 30, 2018 5:51 AM
Subject: Re: [Coq-Club] Is it time...

Hi,

In the absence of more comments about how to make the policy about
announcements evolve, let me tell that I will release four
announcements from non-subscribed senders (ACL2, SYNACS, CSL, EUTypes
summer school).

As for the discussion:

From Derek Dreyer:
> Personally, I can say for myself that I am not particularly interested
> in any of the announcements posted on coq-club -- and I delete them
> immediately -- because I expect that any sufficiently interesting
> announcement will also be sent to Types-announce.  (That may or may
> not be the case, but it is my attitude, and I imagine I'm not alone.)

I believe that I'm roughly in the same situation, so this view seems
reasonable to me. For instance, about the 4 announcements above, would
someone be missing one if there had been not forwarded?

From Derek Dreyer:
> Nevertheless, I would argue that independent of that point, it would
> be great to see announcements on a separate list because those posts
> are of a fundamentally different nature than the technical coq-club
> posts.

Under the assumption that coq-club continues to redistribute general
conference / jobs / school announcements such as POPL, LICS, CSL, ...
that's a possibility. But then we need a clear policy for subscribers
too. Also, I feel that what matters is more the degree of connection
with Coq than the fact that it is an announcement or not. For
instance, I don't see why we should suscribe to a different list which
shall give a lot of general announcements which we can get elsewhere
only to get the few announcements for the workshops, courses or jobs
directly related to Coq. If, at the end, the second list is only for
general announcements, why to have it at all since the general
announcements are available also on more dedicated lists such as
TYPES-announce.

From Perry E. Metzger:
> I run several large (thousands of subscribers) mailing lists. I've
> found over time that simply restricting postings to people who are
> actually subscribed to the list has an amazingly salutary effect.

Yes, it seems so.

> I would suggest automatically rejecting postings from
> non-subscribers.

This is also a possibility which would give a good compromise. If ever
the situation becomes unmanageable for moderators (which is not the
case currently) that should probably be the solution.

* Summary *

I did not have the opportunity to discuss the question with Matthieu
and other developers, but, to summarize, wrt releasing announces from
non-subscribers, I feel that it should:
- either be the status quo (but avoiding forwarding too off-topic
  conferences like ICPRAM last time),
- or downright renouncing to forward any announcement which has not an
  explicit link with Coq (typically leaving the role of distributing
  general announcements to TYPES-announce).

In the latter case, the policy should extend to subscribers.

Does my summary make sense? In any case, unless the discussion evolves
with a different outcome, I believe that I shall continue in the
meantime as it is.

Best regards,

Hugo