Re: [Haskell-cafe] PROPOSAL: Rename haskell@ to haskell-announce@
[cc-ed to haskell@, as this discussion is about haskell@]
There are four things sent to the haskell list@
1) Calls for papers 2) Annoucements 3) Oleg's stuff (which are really announcements of a library or technique) 4) Off topic stuff
I'm initially only proposing to mop up category 4, which I am pretty sure the haskell@ people don't want to see.
i still keep all haskell[-cafe] email in one folder, and tend to reply on the list a post comes from;-) but iirc, those who proposed the split were not looking for a pure announcement list: (*) the idea was to have a low-traffic window into all haskell developments of wider interests, while also establishing a free-form forum for general discussion/newbie questions. so everything that would be of interest to all haskellers, including those too busy to follow haskell-cafe, would go to haskell, everything else would go to haskell-cafe. but even those topics starting out on haskell are meant to migrate to haskell-cafe after a few posts at most. in other words, people were meant to subscribe either to haskell or to haskell+haskell-cafe, and posting to haskell was meant to be a flag able to raise a topic briefly over the general din in haskell-cafe. see also the welcome messages: http://www.haskell.org/mailman/listinfo/haskell http://www.haskell.org/mailman/listinfo/haskell-cafe (note that the second is slightly misleading: *everything* is off-topic on haskell@ after a few exchanges, i think; note also that crossposting was explicitly ruled out)
I did wonder whether this discussion should take place on the haskell@ list or the haskell-cafe@ one - the great ambiguity of the lists.
this discussion should have started out on haskell@. and since it is entirely about changing haskell@, it probably should have stayed there as well. it is not very helpful to discuss changes to a specific list on another list!-) however, it has become a problem, and i don't know whether everyone on haskell-cafe is really subscribed to haskell as well anymore. so posting announcements only to haskell might miss the majority of haskeller on cafe, the crossposting some have started to resort to is expressly discourage in the haskell-cafe 'charta', newcomers don't know where to post or where to subscribe, threads sometimes linger on haskell@ instead of migrating to haskell-cafe, etc. as i indicated, i wasn't a fan of the split, but it had its merits: the free discussion on haskell-cafe is nice, and the intention of haskell@ as a highlights channel for busy haskellers was understandable, even if it doesn't seem to work any more. if my interpretation of the split intentions (*) is accurate, my proposal to alleviate the current confusion would be: - do not assume everyone is on haskell@ as well - do not prohibit cross-postings anymore, but give clear instructions about when to use them - for most haskellers, haskell-cafe has become the main channel, so do not post only to haskell@ - for most specific discussions, there are specific lists, so there seems little need for haskell@ as a general discussion forum => all posts meant for either haskell or haskell-cafe go to haskell-cafe => threads that are likely to be of interest to everyone on both lists may be highlighted on haskell@, either by crossposting the first (and only the first) email in the thread, or by posting a brief discussion announcement/summary to haskell@ that way, there'd be no confusion about where to post (always haskell-cafe), or whether to crosspost (permitted, but only for thread starters) and no limitation to formal announcements and cfps only (whenever something of wider interest starts on -cafe, interesting techniques or discussions, or events, or announcements, also send a brief invitation to haskell@). and busy haskellers can see on haskell@ whenever there is a thread on haskell-cafe that they might want to read/join. does this sound workable? claus
On Mon, Sep 24, 2007 at 01:00:40PM +0100, Claus Reinke wrote:
in other words, people were meant to subscribe either to haskell or to haskell+haskell-cafe, and posting to haskell was meant to be a flag able to raise a topic briefly over the general din in haskell-cafe.
Do people think that is working? Like I think I said in an earlier mail, apart form the announcements and CFPs etc, I couldn't see anything that distinguished those messages sent to haskell@ from most of the -cafe traffic. Also, I think HWN now does a good job of bringing the current issues to the haskell@ readers.
(note that the second is slightly misleading: *everything* is off-topic on haskell@ after a few exchanges, i think; note also that crossposting was explicitly ruled out)
I don't think cross-posting works well on the lists at all; unless people do some magic they get 2 copies of all the mails, and you can end up with some subthreads only on one list and some only on another.
however, it has become a problem, and i don't know whether everyone on haskell-cafe is really subscribed to haskell as well anymore.
I think it's reasonable to assume that people subscribing to a list called foo-cafe will also subscribe to lists called foo or foo-announce. And if for some reason someone did choose not to, they won't want to be CCed with things sent to those lists.
and busy haskellers can see on haskell@ whenever there is a thread on haskell-cafe that they might want to read/join.
Which they would do by polling the web archives or something?
does this sound workable?
It sounds bizarre to me. Thanks Ian
Forgive me, but I would much prefer a newsgroup to a mailing list. True, I could unsubscribe now and just browse the mailman archives - but for posting, I'd have to temporarily re-subscribe, which is awkward. (Indeed that's the only reason I'm not doing it.) We already have comp.lang.haskell. - Best regards, -- -- Johannes Waldmann -- Tel/Fax (0341) 3076 6479/80 -- ---- http://www.imn.htwk-leipzig.de/~waldmann/ -------
Hi
Forgive me, but I would much prefer a newsgroup to a mailing list.
True, I could unsubscribe now and just browse the mailman archives - but for posting, I'd have to temporarily re-subscribe, which is awkward. (Indeed that's the only reason I'm not doing it.)
I believe you can post from the gmane archives, and that they also offer a newsgroup feed of this list. I think the general consensus (and my personal view) is that a mailing list is the preferred mechanism for most people. Thanks Neil
We already have comp.lang.haskell. - Best regards, -- -- Johannes Waldmann -- Tel/Fax (0341) 3076 6479/80 -- ---- http://www.imn.htwk-leipzig.de/~waldmann/ -------
_______________________________________________ Haskell mailing list Haskell@haskell.org http://www.haskell.org/mailman/listinfo/haskell
Johannes Waldmann wrote:
Forgive me, but I would much prefer a newsgroup to a mailing list.
So do I, that's why I'm reading and posting through gmane :) You still have to be subscribed (to haskell@ but not haskell-cafe@) in order to post but there is a mailman option to not send you e-mail (sounds silly for mail-only but is extremely useful for read-through-gmane). Regards, apfelmus
in other words, people were meant to subscribe either to haskell or to haskell+haskell-cafe, and posting to haskell was meant to be a flag able to raise a topic briefly over the general din in haskell-cafe.
Do people think that is working?
i don't think it is working anymore. haskell-cafe works (mostly;-), haskell sometimes works, sometimes seems a distinction without a difference, and more and more often causes confusion (where to post? who is on what list?)
Also, I think HWN now does a good job of bringing the current issues to the haskell@ readers.
yes, if HWN was more, well, weekly,-) it would nicely cover that job. there is the secondary issue that we'd actually want to alert the haskell@-only readers to discussions early, so that it doesn't take a week before they join a discussion on what to do with mailing lists!-) but i have the feeling that those who are likely to join discussions have taken to at least browsing haskell-cafe as well?
(note that the second is slightly misleading: *everything* is off-topic on haskell@ after a few exchanges, i think; note also that crossposting was explicitly ruled out)
I don't think cross-posting works well on the lists at all; unless people do some magic they get 2 copies of all the mails, and you can end up with some subthreads only on one list and some only on another.
one might exaggerate the split intent as: any message to haskell@ should have its reply-to set to haskell-cafe. but i agree, neither split threads nor crossposts are nice, but they are a reality. the current welcome message discourages crossposts. nevertheless, they are used, for instance, for HWN, and for this present thread, because we are no longer sure of the original assertions, as expressed in the original split: Welcome to the Haskell Cafe http://www.cse.unsw.edu.au/~dons/haskell-1990-2006/msg07680.html we either need to guarantee that haskell is a sublist of haskell-cafe (so crossposts are never needed, because noone is subscribed to cafe only, and any accidental crossposts could be filtered from cafe; all threads are archived in full in the cafe archive, no matter which parts appeared where), or we have to find another way to make this list combination work. personally, i could live with (as proposed) haskell-cafe + hwn + haskell-announce + hcar but i wanted to point out that the post-split haskell@ was not meant to be limited to announcements. it is just that 'low traffic, stay-in-touch-only' has proven to be too vague a charter to work well. which is why we're having this thread, i believe?-) claus
Claus Reinke wrote:
so everything that would be of interest to all haskellers, including those too busy to follow haskell-cafe, would go to haskell, everything else would go to haskell-cafe. but even those topics starting out on haskell are meant to migrate to haskell-cafe after a few posts at most.
in other words, people were meant to subscribe either to haskell or to haskell+haskell-cafe, and posting to haskell was meant to be a flag able to raise a topic briefly over the general din in haskell-cafe.
see also the welcome messages:
http://www.haskell.org/mailman/listinfo/haskell http://www.haskell.org/mailman/listinfo/haskell-cafe
(note that the second is slightly misleading: *everything* is off-topic on haskell@ after a few exchanges, i think; note also that crossposting was explicitly ruled out)
I haven't been able to see how it makes sense to subscribe to haskell@ but not haskell-cafe -- because if a discussion interested you, and went in-depth, you might want to be able to read the whole thread! (It is possible to read the archives, but that doesn't help if you're inspired to _reply_ to the thread...) Isaac
From: Isaac Dupree I haven't been able to see how it makes sense to subscribe to haskell@ but not haskell-cafe -- because if a discussion interested you, and went in-depth, you might want to be able to read the whole thread! (It is possible to read the archives, but that doesn't help if you're inspired to _reply_ to the thread...)
This is exactly what I do. I subscribe to haskell so as not to miss anything important, and when something I'm interested in moves to haskell-cafe, I use gmane. I think most of the people complaining about this issue do the same. Barney.
(...) I subscribe to haskell so as not to miss anything important, and when something I'm interested in moves to haskell-cafe, (...)
Wrong assumption. Even if you know the author it makes more sense to wait and see how others respond to his newest input. This means that you should check up Cafe first to see if it is worth moving to Cafe;-) Another false believe is that best things appear first in Haskell@. On the contrary, long threads in Cafe indicate that participants know what to expect of each other and pick up a thread on the same principle some pick up new Stephen King novel. Before reviews on the amazon.com publishers had to result to tricks like labelling a collection of science fiction stories by various lesser known authors with Isaac Asimov name even when his only cotribution was a preface. As I said ealier - really big traffic will need intelligent screening and lebbeling. And to what pathology it leads when unattended see comp.soft-sys.matlab. Cheers, -Andrzej Jaworski
participants (8)
-
Andrzej Jaworski -
apfelmus -
Barney Hilken -
Claus Reinke -
Ian Lynagh -
Isaac Dupree -
Johannes Waldmann -
Neil Mitchell