[Mailman-Users] Automatic subscription based on e-mail subject

[Stephen J. Turnbull]

R. Diez writes:

 > Your comments are surprisingly unfair for someone in a mailing list
 > for mailing list software.

How would you be a good judge of fairness?  Have you been developing
mailing list software for twenty years and reading the requests and
problems of users daily for that period?  We developers have a record
of considering what others have to say, and their requests, at least
somewhat fairly.  If we didn't, Mailman wouldn't be the most popular
open source discussion list solution.

Specifically with respect to the comments themselves, consider that in
many cases mailing lists deliberately direct replies to themselves,
making it fairly inconvenient to reply to author.  (I find it a little
bit shameful that Mailman explicitly supports that abuse of
"Reply-To", but the demand for it is indeed overwhelming.)  Most lists
I know of do have a "what starts on the list should stay on the list"
policy even if not enforced by Reply-To munging.  That indicates that
these communities desire a coherence that is harmed by the kind of
behavior you described if it becomes frequent.  I also have some
experience with how such features affect the communities that use them
(see below).

 > Let's take me as an example.

But that's the whole problem, you see.  In deciding how to improve
Mailman, we need to consider not only what some individual posters
want, but also communities, their dynamics, and how major changes like
this affect them.  We think about how mailing lists work, and
therefore what they can and cannot do well.  The service you want is
provided much better by existing forum software and by issue trackers.
There has been some ambition to fill that niche with Mailman features
(less now that Barry has retired), but web-based tech already exists
that does it well, and I don't see how both can be supported well at
the same time.

 > I asked about a way around a perceived limitation,

And I responded that it is not a limitation, it is a feature of the
kind of community that mailing lists support best.  *This* community
does support the mode you requested, you know, just not via Mailman.
There's a tracker at gitlab.com.  I would expect that these days most
communities that develop software do.

I've thought about implementing this in Mailman and came to the
conclusion that you can't have both.  No pushme-pullyou software does
both well because "push by mail" is basically asynchronous while "pull
by web" does support useful synchronicity.  Both groups of users get
frustrated because they don't get the experience they expect.

I've also seen this in practice in groups that move, or try to move,
from lists to forums (I don't know of examples of the reverse).
There's a general turnover of active posters, with much more
specialization in thread participation, and a departure of experts who
are overwhelmed by repeated questions and proposals and are
disappointed in the decreased information density.

In some cases that may be a desirable effect (even for the
"disappointed experts", who waste less time).  But our mission is to
support the kind of community those experts (and many other users)
apparently want.

 > But am I spamming?

I don't know in general, because it depends on the community and I
don't know where you've posted.  In many of the mailing lists I
participate in the answer would be yes, if you posted to them in that
mode.

Here, inasmuch as what you want is technically beyond what Mailman can
currently support, there's no problem with posting the question as
long as you accept the answer "no, Mailman can't do that and no, there
are no plans to support it soon," and don't ask for a personal reply.
That is useful to us as we can gauge the amount of support for the
request from other users.  (Zero, so far.)

 > Is this discussion not welcome here then?

You got multiple responses, obviously discussion is welcome.  But I
see no enthusiasm for your proposal from other users.  Continuing
discussion here doesn't seem profitable.

If you want to become a developer and contribute some of the effort
required to provide the features in Mailman 3, subscribe to
mailman-developers at python.org and clone the HyperKitty repository from
gitlab.com/mailman/hyperkitty.  It has some of what you want, but
currently it's the tail of the dog.  To support the features you want
it really would need to wag the dog, which is quite possible.

Or you can just subscribe to mailman-developers and advocate for what
you want.  But given the current lack of manpower and the many more
important tasks that need to be done, it will quickly become tiresome
if you don't contribute substantial effort yourself.

 > Maybe you are implying

No, you're taking insult where none is intended.

 > Other projects have benefited from a bug report or a small patch I
 > sent to their mailing lists. I was never actually subscribed to any
 > of those.

Since those are for the benefit of those projects, they're not
spamming by my definition, though in many cases on this list we prefer
they be directed to the tracker.  Still, we accept those here for that
reason.

 > If all this actually bugs you,

I don't respond when I'm actually bugged.  I just disagree, and don't
think the behavior you want supports our mission.  If there were
demand for it, I'd consider implementing it despite my misgivings (see
"Reply-To munging" above).  But there isn't, at least not so far.

Regards,

Steve