Improving Foundation-l

From Meta, a Wikimedia project coordination wiki
This list has been renamed Wikimedia-l

The Wikimedia Foundation mailing list, or "Foundation-l", is the primary forum for e-mail discussion of Wikimedia Foundation issues (and, by extension, various cross-project issues). As the Wikimedia projects have grown in size, so has the volume of list traffic, and with that growth come several distinct problems.

Discussion started at Use of moderation. See also: a similar page on Foundation-l at Strategy Wikimedia. See the conclusion of the discussion at [Foundation-l] Foundation-l open for business, with changes (Sat Nov 14 03:20:23 UTC 2009).


  • Overall volume is too high for many to keep up; people simply give up and unsubscribe, removing their voice from the debate.
  • A minority of posters dominate the discussion, giving disproportionate attention to their points of view (and substantially increasing traffic).
  • Another minority argues for the sake of arguing, again inflating the overall volume of the list while contributing little to the debate.
  • The mailing list's purpose is clouded by allowing far too much off-topic discussion. Foundation-l is for discussion of topics directly related to the Foundation - not anything you may wish other people are interested in.
  • At the Paris Commons meeting it became clear that people who are crucial to the Foundation REFUSE to write or respond to the list because they consider it disfunctional and not worth their time.
  • Other contributors are driven away by the haphazard and inconsistent moderation standards.

Possible solutions[edit]

Changes to foundation-l[edit]

  • Moderation of problematic individuals
    • Difficult to draw a line; perceived lack of fairness when an individual is judged to be just slightly more problematic and moderated, while another is allowed to continue posting.
  • Moderation of threads ("killfiling")
    • More "fair," but still controversial; individual moderation invariably accompanies it to ensure that instigators don't simply change the subject line and continue the thread.
  • Individual caps on the number of allowed posts per month
  • Presumably, someone would go on auto-moderation after they reach their post limit; the mod would need to have help keeping track of the posts. I'd be in favor of a 30-posts-a-month limit (1 per day, your choice whether to post all at once in a single discussion or spread them out) -- phoebe 23:08, 11 September 2009 (UTC)[reply]
  • I would not be against this, provided it is applied consistently to all posters, no matter who they are. As you can see from the stats page, individual posters have at times exceeded 100 posts per month.Wjhonson 22:37, 8 November 2009 (UTC)[reply]
  • LOL at "applied consistently to all posters". That'll be the day! Otherwise, I'd agree to this solution, too. Actually, I think it would be more effective still to effect limits of "30 per month, and 3 per day". I think someone posting 30 times in three days is really annoying. -- Thekohser 21:34, 9 November 2009 (UTC)[reply]
  • Appointment of volunteer facilitators for specific discussions
    • People participating in a facilitated discussion agree to respect facilitation
    • You can't resolve the list's problems by resolving the problems of single discussions.
    • I like this idea. Arguments are good, but on foundation-l they often just don't work. We get distracted from the productive debate by arguments about arguing, which is a waste of everyone's time. Good quality facilitation could help keep discussions on topic and make sure that both sides of an argument get equal time. The main problem I see is that email is a delayed medium, facilitation is much easier in real time, particularly in real life where you can easily catch the eye of the facilitator to let them know you have something to say. --Tango 21:53, 8 November 2009 (UTC)[reply]
  • Use wiki pages to document list rules and make moderation decisions through open wiki processes (similar to AfD)
    • AfD-like pages would multiply the number of battlefields.
    • I would advocate opening the list processes to this sort of community control. Rules set by the community, which the moderators simply enforce. Rather than the situation now where there is a perception of arbitrary enforcement.Wjhonson 22:39, 8 November 2009 (UTC)[reply]
  • Wikiinfrastructure to support and ease moderation
    • All users on foundation-l must have an User account on Meta, with automatic mailsignal when discussion page is changed
    • Document wanted behaviour rules on meta in the same way as on wikipedia (wp:et, wp:not, no chat, do not overload etc)
    • Warn unwanted bahaviour on the users discussion page (gives tracebility)
    • Block user when the bad behaviour does not stop after warnings
    • (and keep pages like this on meta to be a place for disussion on processes etc of foundation-l, ie keep them away from the list itself)
Anders Wennersten 07:57, 11 September 2009 (UTC)[reply]

Replacements/alternatives to foundation-l, superseding it[edit]

  • "I think we should stop using this outdated technology altogether and instead switch to a web-based forum, where comments can be postmoderated (i.e. removed after posting), and unproductive threads can be moved or locked." (from foundation-l, Tim Starling)
    • I think that it would decrease participation: email inside of inbox is much more user friendly than browsing through forum threads. Let's wait for Google Wave. --Millosh 06:23, 9 September 2009 (UTC)[reply]
Wikimedia Foundation mailing list is archived using, see here. Some people asked about creating an forum as an alternative to the Wikimedia Brasil mailing list and it was pointed Google Groups can archive a mailing list and you can use it as a forum interface as well (we are using only for archiving purposes tho, navigate through our mailing list version on Google Groups). At this discussion platform, people can vote (from 1 to 5 stars) for each topic and each comment, returning an average as a result. It's also powered by Google search engine for browsing through topics, which is quite good, IMO, for back reference. Depending on configuration, you can also allow people using this Group only as a forum interface and people could decide if they want continue using emails or post directly on the Group page (with an interface similar to the Brazilian version I linked). There are other Web 2.0 technologies which I believe can improve communication. I agree with Tim the mailing list is outdated for its purposes and trying Google Groups wouldn't require too much effort and technical skills. I think it's worth trying --Tom 14:24, 9 September 2009 (UTC)[reply]
Henning Schlottmann just informed Foundation-l is also at I think it's a good solution, it still doesn't use the potential of the community for ranking topics and comments, but it seems enough for me. --Tom 14:01, 10 September 2009 (UTC)[reply]
Another cool feature of gmane is that you can read the mailing list as a blog or using a NNTP news reader. See a topic being discussed in blog format . More information here. --Tom 14:30, 10 September 2009 (UTC)[reply]
Gmane may provide a more manageable interface to the list for some, but in exactly the same way that a mail client with thread grouping features does. It's still, fundamentally, a mailing list; none of the content issues are addressed. Austin 18:34, 10 September 2009 (UTC)[reply]
What about using the software used in Slashdot or reddit, integrated with email (for those that would claim to be a pain to change from email to another place)? --Tom 19:36, 10 September 2009 (UTC)[reply]
Gmane is not very reliable, sometimes days worth of messages are missing. --Tgr 01:36, 16 September 2009 (UTC)[reply]
Nabble offers a forum-like interface to existing mailing lists with many of the features an average forum software offers (posting, flagging, moderation, filtering of posts based on collaborative rating, better search than in pipermail). --Tgr 11:26, 11 September 2009 (UTC)[reply]

Obligatorily pimping Extension:LiquidThreads, which is ready for test deployment. Werdna 11:41, 14 September 2009 (UTC)[reply]

I just saw this mentioned on the September 14th edition of the Signpost, and I was going to make the same suggestion as in this section. Just set up a phpBB board (or other alternative) in addition to the mailing list. This is one area that I've felt that WMF has failed in for a long time now, and eschewing web forum software is a large cause of the problem. I see the other comments that seem to indicate some sort of fundamental dislike of web forums, but I simply don't find those arguments to be compelling, and the commercial success of vBulletin, phpBB, and others seems to be a clear indication that such views are not at all common. LiquidThreads is another obvious alternative (and I'm personally ecstatic to hear that it's ready for test deployment), but I think that using a dedicated forum software solution would be more appropriate in this narrow area. Wiki software is terrific, and LiquidThreads is going to be an outstanding improvement, but even with LT wiki software is simply not designed for discussion.Ohms law 01:05, 16 September 2009 (UTC)[reply]

I second this. The mailing list is used as a forum. As mentioned, set up a phpBB forum _and_ remove the mailing list. Jtico (talk) 15:36, 17 September 2009 (UTC)[reply]

I am in favor of setting up the LiquidThreads extension on Meta, and using it as a replacement for or alternative to foundation-l. Discussion guidelines and policies would be developed here on Meta as wiki pages. This also meets the need to have a single forum (with a set of boards) that serves the world-wide Wikimedia movement.--Eloquence 00:18, 13 November 2009 (UTC)[reply]

Even though LiquidThreads is rough around the edges, I second this proposal.. It will get better because we use it. We truly need a functional way of getting messages out ... (I wanted to announce the start of the Russian Wikiversity for instance).. GerardM 00:23, 13 November 2009 (UTC)[reply]


  • One thing is to tag the messages. This is an alternative to the spliiting of the list. The tags need to be thought out but for instance these could be: strategy, management, funding etc. One message can carry one or more tags. I would for instance ignore the funding issues since I am not interested in them for the time being, but would read the strategy tagged messages. From the name it is sometimes impossible to understand what the message is about, and also in the same thread they can change topics unexpectedly.--Yaroslav Blanter 16:06, 9 September 2009 (UTC)[reply]
    But generally I prefer moving to a web interface (like forum or smth).--Yaroslav Blanter 16:07, 9 September 2009 (UTC)[reply]
    It seems that only automatic keyword-based categorization of posts is available. I don't know if that can function, but the use of topics is indeed enticing. --Nemo 07:00, 11 September 2009 (UTC)[reply]
  • Encourage people to do a bit of decision-making on the read side, and provide some information on how to make it easier to do so. With some of the suggestions above, I see absolutely no reason they need to be done at a centralized, list-moderator level. Think some threads are a waste of time? Don't read them! Think some posters always post nonsense? Don't read their posts! Most email software makes it easy to sort by thread or by poster, and you can mass-delete all the posts you don't want right away. More advanced users might consider killfiles. --Delirium 09:51, 11 September 2009 (UTC)[reply]
  • I dispute that anyone on Foundation is arguing "for the sake of arguing". This paints with a broad brush discussions evidently based simply on "I Don't Like It". It also begs the question, and casts prejudice. I would suggest this part of the discussion be removed from the heading as it's counter-productive to discussion. Wjhonson 22:34, 8 November 2009 (UTC)[reply]

Related suggestions[edit]

  • Start an announcements list, as has been proposed before/elsewhere, that is solely for announcements related to the projects & the Foundation. Provide this as a way for people who are only interested in news to keep up with the projects without having to slog through Foundation-l. Have open posting (anyone from any project, any language, can submit something), but have heavy moderation -- no discussion or off-topic posting tolerated. CC these posts as well to the current foundation-l, so those who wish to discuss them can do so separately. Call the new list 'announcements-l' or similar. -- phoebe 19:26, 10 September 2009 (UTC)[reply]
    You can consider chapters-reports an example: clear subject of the list, low traffic, very very high signal to noise ratio, discussions on Chapters or Internal list.
    This may be a good idea, but we do not have so many announcements on the foundatio-l, and these do not usually get many follow-ups, so I do not think this suggestion (which may still be good to implement) can reduce the foundation-l traffic considerably.--Yaroslav Blanter 08:00, 11 September 2009 (UTC)[reply]
  • please please please are there better options than pipermail for the "official" list archives? I have spent a lot of time digging through the list archives, and they are totally unwieldy and a pain to use. please? -- phoebe 19:32, 10 September 2009 (UTC)[reply]
    Phoebe, there are two alternatives, or In you can even see discussions in blog format. See comments above. --Tom 20:20, 10 September 2009 (UTC)[reply]
    Ah, I think such alternative should be linked here. Maybe also with a link to a page we can create explaning how to use newsreaders as another way to navigate through topics. --Tom 20:26, 10 September 2009 (UTC)[reply]
    Yes, there are alternatives, but Phoebe commented about making the "official" list archives better, which is definitely a better solution than using other mediums. Cbrown1023 talk 20:45, 10 September 2009 (UTC)[reply]
    Precisely. I know about gmane et al, but simultaneously improving our own hosted solution would be ideal. In my experience posts do get missed in gmane, which makes it or similar unsuitable as a substitute for the official archives. -- phoebe 06:15, 11 September 2009 (UTC)[reply]
    And markmail lacks a threaded mode, so it's useful to find a single message, but not to read a whole or a part of thread. --Nemo 07:12, 11 September 2009 (UTC)[reply]
    OK, Phoebe, I didn't get at first what you meant, since you said you spent a lot of time on mailman archives and I think gmane is easiear for searching and back reference. Sorry. I talked to Cbrown1023 about a possible solution and I will send an email to Wikitech list when possible. After a quick search, I found a Python script which can be improved for archiving and searching, and there is as a mailman wishlist requesting some of these features. --Tom 13:20, 11 September 2009 (UTC)[reply]
  • see also [1] and followup -- phoebe 06:00, 13 September 2009 (UTC)[reply]

Moving forward[edit]

Foundation-l is currently on emergency moderation (blocked to all traffic except official announcements), and I don't plan to open it back up for general discussion until there's at least some idea of how to improve things moving forward.

It's clear that the current situation is unsatisfactory. Consensus is that the signal:noise ratio is too low, with the only (albeit vocal) dissent coming from the most prolific posters. Suggestions of "just ignore me" and "get a new e-mail client" simply aren't good enough—mailing list posts do not exist in a vacuum, and what might otherwise be a productive discussion on a topic quickly spirals out of control with all the tangents introduced by those who feel the need to weigh on every single comment.

Some have asked me to simply be more heavy-handed in moderating the list and ensuring that discussions are productive. Others have (rightly) called for consistency—there's a reason I'm reluctant to moderate people at my own personal whim, and only do it in extreme cases. It's really not fair for me to impose moderation just because a thread annoys me, or because I've received complaints.

Right now I'm focused on immediately practicable solutions, and I want to take a straw poll. Please weigh in below:

Impose a periodic (monthly/weekly) post limit[edit]

Each person can only send a certain number of messages (10 per week? 50 per month?) over a certain span of time. Because our current list software can't enforce this by itself, this would by default be a soft limit—a moderator would be notified when someone approaches the limit, and if it goes too far the violator would be put on moderation, with every post needing to be approved manually.

  • I'm in favor of this, and would make the cutoff 30 posts/calendar month (perhaps with a 3/day addendum as suggested above). Looking the top poster statistics, only 4-5 people consistently go over this limit in any given month. -- phoebe 22:57, 10 November 2009 (UTC)[reply]
  • I think this is a good idea; I'm not sure what the optimal limits might be, but a daily and monthly limit would help a great deal I think. Nathan T 18:55, 11 November 2009 (UTC)[reply]
  • Trying to improve discussion by stifling it doesn't seem a good approach to me. I don't see any problems with the number of posts. If there are problems it is with what the posts contain. This proposal is just assuming that the posts of those that post a lot are less valuable than other posts. While there may be an increase in quality due to posters being more selective, I don't personally expect that to happen to a significant degree. Also, this proposal doesn't account for different lengths of posts. Should a quick one line reply be treated the same as a 1000 word monologue? I would say not - the time required to read an email is roughly proportional to its length. I think this proposal would result in people just combining their replies into one long email, which I find harder to read than separate replies to each email. --Tango 19:14, 11 November 2009 (UTC)[reply]
  • I am in favor of this. My expericnce is that isit is good to make a distinct difference between a maildiscussion and IRC chat. And instead of the spontanous instantanous replies of a chat to force more though-over replies on maillist, with limiting even to one reply per user and day (and thread).Anders Wennersten 19:51, 11 November 2009 (UTC)[reply]
  • I'm in favour. Sure, there's downsides. But I think the downsides of unstifled conversation outweigh them. --Bodnotbod 20:06, 11 November 2009 (UTC)[reply]

Establish rules and consistently enforce them[edit]

Currently, the only (de facto) rule is "try to keep discussions productive and civil, or you'll be moderated." This is obviously an extremely subjective rule, and although it can easily be more heavily enforced with the addition of a few more moderators, some would prefer a more codified set of dos and don'ts. Unfortunately, it's going to take a while to reach consensus on rules.

This would certainly be desirable. Perhaps just a strongly enforced rule about staying on topic and branching off to a new thread if you want to move onto a different topic would help. --Tango 19:15, 11 November 2009 (UTC)[reply]
Usually the rule would be Common sense and real destructive entries will cause the user to be taken off the list, about as it is stated today. If the list is too untangly that common sense do not work as in this case, it will not be helped by more detailed rules - it will then only start disussions of the intepretation of the rules.Anders Wennersten 19:56, 11 November 2009 (UTC)[reply]
The problem with common sense is that it isn't common, it is different for different people. --Tango 22:18, 12 November 2009 (UTC)[reply]
I would welcome any change because foundation-l is dysfunctional. I would suggest having it enforced for now by your obvious good intent while working on a solution. The problem is that without a functioning foundation-l there is no channel for this subject. GerardM 08:11, 12 November 2009 (UTC)[reply]

Moderate the most prolific posters and wrangle them into the notion of pith[edit]

This is actually the most common proposal I get, usually privately. I don't like this idea, and it seems a lot like a bill of attainder; moreover, I just don't have the time and effort free to put into this.

  • I think the past few years has shown anyone paying attention that there are two or maybe three people mostly responsible for the problems we've seen; it makes sense to address them directly at the same time that the structural issues (such as post limits) are addressed. Nathan T 18:56, 11 November 2009 (UTC)[reply]
    I'm afraid you are viewing this rather naively. Problems rarely have a single isolated cause. Getting rid of the more prolific posters would certainly minimise the problems in the short term but that doesn't mean those posters are the primary cause. Perhaps first we should decide what the problems actually are. I don't think the sheer volume of posts is a problem since they are so easy to ignore. Is the problem a difficulty stay on topic? (That's certainly an issue we should look at.) The aggression and hostility in some posts? (I think such posts are rare enough not to be a real problem - the moderators should be able to deal with them on a case-by-case basis.) The more important topics not getting the attention they deserve? (That's a highly subjective issue, so it is difficult to say if it is a problem or not.) --Tango 19:32, 11 November 2009 (UTC)[reply]
Here are some problems with the "prolific" posters: They are argumentative, and view all discussions as adversarial debates. They endlessly debate sideline minutiae. They make disparaging comments about other list contributors or people associated with the Foundation. They make repeated posts to single conversations, often replying to each new message, thereby dominating the subject and making it harder for others to participate. When they join a thread it often drops off shortly thereafter except for their own posts, because others are unwilling to engage in the type of exchange that these people seem to value above all others. I'm not surprised that you disagree with the methods for improvement suggested by others; its no secret that many of them are aimed at forcing you to change your posting style, since you have refused to acknowledge subtle and less-than-subtle prods to do so on your own.Nathan T 19:11, 12 November 2009 (UTC)[reply]
"Argument", "discussion" and "debate" are all synonymous terms to me. The way you reach good decisions is to share your ideas then listen to other people's reasons for disagreeing with them. You then either realise they are right and change your mind, or you tell them why you think they are wrong, and then they either change their mind or tell you why they thing you are wrong, and you repeat until you either reach agreement or decide you aren't going to and agree to disagree. Unfortunately, on foundation-l there are very few people that seem to actually understand that process. People consider any disagreement to be a personal insult and refuse to give a rational defence of their position. I agree that sometimes foundation-l gets bogged down in unimportant minutiae, we need to do something about that. I think the simplest way to solve that problem would be for other people to drag the conversation back in the right direction by example - make points or ask questions about what you think should be the main topic under discussion. I don't see why it is hard for other people to participate just because some people are already participating. There are no restrictions on how many emails a thread can contain and it's not like an in-person conversation where you have to wait for a quiet moment to speak. I have explained my reasons for disagreeing with the "prods" you mention. Nobody has made a significant effort to convince me otherwise (complaining again and again doesn't count - you have to listen to my reasons and tell me what is wrong with them). --Tango 22:29, 12 November 2009 (UTC)[reply]

~*Could be an idea but an implemetation do need an elaborate set of procedures for warnings etc and close supervision.Anders Wennersten 19:58, 11 November 2009 (UTC)[reply]

Replace the list with a web forum or newsgroup[edit]

Move the group to a technology better suited to threaded discussion, with more fine-tuned control. There are various technologies available, many of which allow both web and newsgroup access. Proposals about specific software are welcome.

  • it is a possiblity but then I see the maillist as closed down. For me for example I would not follow this webbfora, beacuse it is not part of my daily routines as it is to overlook my mailbox(es).Anders Wennersten 19:59, 11 November 2009 (UTC)[reply]
  • I don't favour it. I'd like discussion to go to my inbox, albeit with reduced traffic. --Bodnotbod 20:09, 11 November 2009 (UTC)[reply]
  • webfora have their own dynamic, some things are a lot easier (e.g. actively moderating on-topicness, splitting into subfora, ...). However, it would require splitting the discourse into yet another place when one of the problems the Wikimedia community is dealing with is a very fragmented space to talk: there's the ml, irc, Skype, the wiki's themselves, meta, ... . Changing the technology will not change the underlying problem we're having but it can limit the influence that problem has. Henna 21:13, 12 November 2009 (UTC)[reply]

Split the list[edit]

Create another list specifically for announcements and "official" traffic, with posting rules yet to be determined; allow foundation-l to remain just as it is. Alternatively, create a "foundation-moderated" list with stricter rules.

Austin 22:46, 10 November 2009 (UTC)[reply]

Interesting. On the flip side what about a completely unmoderated list. It's at least worth a try.Wjhonson 03:38, 11 November 2009 (UTC)[reply]
A completely unmoderated list is what we have now. The list admins step in once or twice a year when things get completely out of hand, but it takes quite a lot for that to happen. This is the format people are dissatisfied with. Austin 04:40, 11 November 2009 (UTC)[reply]
But I'm here referring to the split issue. Two lists. One moderated, the other not. I don't see a downside to that. People can pick whether they want tea conversation, or the rumpus room (that's a joke).Wjhonson 23:21, 13 November 2009 (UTC)[reply]
Who in his right mind would join the talkies if not only for amusement ? GerardM 11:29, 14 November 2009 (UTC)[reply]
An announcements list would probably be good. I suggest very strict posting rules (replies should be limited to adding factual, neutral information or correcting factual mistakes). There should be no discussion at all on the announcements list. If there is, it would result in discussion being split between two lists which gets very confusing (even if emails to the announcements list are automatically forwarded to the discussion list, it would get confusing). There would also be a problem with what happens when people reach their posting limits if you try and restrict discussion to a few posts each - people wouldn't have a right-to-reply which results in lopsided conversations. For example, a majority could overwhelm a minority since there would be far more posts available to them. --Tango 19:45, 11 November 2009 (UTC)[reply]
Yes this could very well be the right solution, the number of posts on the list since it has started to be moderated has been baout just right.Anders Wennersten 20:01, 11 November 2009 (UTC)[reply]

Fwiw, if you wanted to commandeer an existing list, Foundation-news-l has never been used - though it was intended to send out html versions of Quarto (remember that?!). I'd be happy to hand it over. :-) (Probably best to ping me by email if so.) Cormaggio @ 19:32, 11 November 2009 (UTC)[reply]

Change to a curated list[edit]

Instead of moderators, whose job is to say no to the bad, I propose a curated list, where a somewhat larger group acts like magazine editors, who say yes to the good. Announcements would be treated like they are now. Essays would be pitched to the list curators; if any one of them felt a topic was worth of attention, they could shepherd the essay to on-list publication. Following essays on the same topic would be handled by the same curator. Direct replies would be discouraged (as they tend to be more interesting to the participants than the audience). However, the curator would have the option of following up with summaries of public reaction or other topic activity. There's more to it, but that's the basic notion.

William Pietri 19:59, 11 November 2009 (UTC)[reply]

  • Why shouldn't the list contain things that are of interest to just the participants? I think a series of essays in response to each other is better suited to the blogosphere. --Tango 20:08, 11 November 2009 (UTC)[reply]
    • Sorry, that was unclear. I mean that when you have person A and person B talking to mainly to one another, it's easy for them to forget the large silent audience. Certainly, the long back-and-forth discussions are the ones people complain about the most. As with large physical meetings, I think it's more effective for those discussions to happen separately, with participants reporting back to the group. William Pietri 08:58, 12 November 2009 (UTC)[reply]
      • In a real life meeting, people gather together in groups to discuss different things. If a particular discussion is only of interest to two people they will quickly find that they are sitting apart from everyone else. If it is interesting to other people, they will gather round and either get involved or just listen. I don't see why the same can't be true of a mailing list. This comes back to the point I've been making since the start - it is really easy to not read an email. --Tango 22:32, 12 November 2009 (UTC)[reply]
  • The downside of pursuing this idea, it seems to me, is staffing. So, on balance, I'd tend towards the simpler solutions, much as this is an intriguing prospect. --Bodnotbod 20:20, 11 November 2009 (UTC)[reply]
    • That's a risk, but I think it'll be easier than moderating or rule-making. Also, I think curating would be less burdensome than moderating; moderators have to stay on top of things, but curating a thread is something you can do when you have time and interest in the topic. I think that, plus the positive rather than negative nature of the role, would get us a bigger pool of volunteers. William Pietri 08:55, 12 November 2009 (UTC)[reply]