Revealed! Why the community is on crack.

The problem with Internet-based projects is that they form groups of humans, and a group is its own worst enemy. That’s a marvellous essay by Clay Shirky, who’s on the Wikimedia advisory board for good reason. When I read it I was just nodding my head and going “yep” over and over. A community (Internet or not) has a life cycle. It starts, it’s good for a while, it chokes itself or falls away. I’ve seen this time and time again.

On Wikipedia, the community is not an end in itself but has grown around a purpose. The English Wikipedia’s interesting community problems are an emergent phenomenon, not Wikipedia or Jimmy Wales doing something wrong.

(Not to mention the flood of people for whom this is their first online community, who haven’t experienced the cycle even once. We have enough trouble enculturating Usenet refugees and their robust interaction style.)

Larry Sanger is trying to work around this on Citizendium, as advised by Shirky’s main source, Wilfred Bion‘s Experiences In Groups: group structure is necessary. Robert’s Rules of Order, parliamentary procedure and so forth. The obvious risk is killing the best in favour of steadiness.

Shirky notes: “Constitutions are a necessary component of large, long-lived, heterogenous groups.” I’ve long spoken of Wikipedia’s fundamental policies — neutrality, verifiability, no original research; assume good faith, no personal attacks, don’t bite the newbies — as a constitution, and said that any process that violates them must be thrown out. The catch being there’s not yet a way to enforce that.

One thing Shirky strongly points out: “The third thing you need to accept: The core group has rights that trump individual rights in some situations. This pulls against the libertarian view that’s quite common on the network, and it absolutely pulls against the one person/one vote notion. But you can see examples of how bad an idea voting is when citizenship is the same as ability to log in.” You would probably believe the outrage when I applied the phrase “one moron one vote” to Requests for Adminship, the prime example on English Wikipedia at present of a group that’s being its own worst enemy. Worse than Articles for Deletion. (The reason people form into insular groups that defend one moron one vote is that the groups then attain local “core” status and feel they can get some work done. This is why new committees keep popping up.) The trouble is then squaring this with not being exclusionary toward the newbies.

(And you’ll see Shirky’s 2003 essay speaking of Wikipedia as a project that’s dodged that one. Whoops.)

The Tyranny Of Structurelessness by Jo Freeman is one of my favourite essays on emergent hierarchies: if you pretend there’s no hierarchy, one will emerge out of your sight and bite you in the backside. (I’m unconvinced its solutions, particularly electing everyone, are directly applicable here — just about every process on English Wikipedia even resembling a vote rapidly turns into an insular committee or a lynch mob.)

Some consider cabalism on English Wikipedia the source of all problems. Unfortunately, with 4330 frequent editors and 43,000 occasional editors each month, no-one is going to know everyone. So people will cluster with those they do know just to get anything done.

The people who do work on a project will usually ignore idiocy until it gets in their face. In the Linux world, the lists resolutely ignore the baying fanboy cat piss men, and Linus Torvalds remains project leader by acclaim. The LambdaMOO solution in Shirky’s paper may be the best option: the wizards return and lay the smackdown. Let’s start with shooting all rules that violate the above six constitutional basics. So who are the wizards?

How to keep the community focused on the point of the exercise? What level of control does one apply to keep the project on track without killing off the liveliness? How would you apply Shirky’s findings?

Is there a sociologist in the house?

(Other useful responses on the social networking site. From people I mostly know from Usenet.)

10 thoughts on “Revealed! Why the community is on crack.”

  1. Isn’t it too late to create a “constitution” in Wikipedia?

    The population is too large and you simply have no way of a significant majority to agree to any major change in the status quo.

    I believe you have only two viable options:

    a) create a fork and lay down the ground rules from the start (the Citizendium way)

    b) try the [[boiling frog]] / [[stone soup]] approach: introduce small steps that slowly, gradually move the project towards your goal, but every step is below the pain threshold of the mob. This is somewhat similar to what’s happening with the fair use rules or gradually limiting the editing rights of anonymous editors. This can take several years to implement though.

  2. I think you’re underestimating the addictive powers of wikicrack. A harsh back to basics will shake off hardly anyone, I think, for more than a week or two. Look how bloody hard it is to ban people and have them actually go away.

  3. But who has the authority to implement any significant change to the status quo?

    The ArbCom? Jimbo? The Foundation? The Cabal? :)

    When was the last really significant and sudden change put through that affected how the English Wikipedia operates or governs itself?

    Alas, Wikipedia has no wizards.

  4. I largely agree with Shirky’s ideas and find his writing insightful. The main rebuttal to the application of his ideas to Wikipedia is that Wikipedia is now large enough and unique enough that there is no model to use as an example. You can extrapolate data and experiences from smaller wikis, but yet there is clearly a tipping point. There are ideas that do not scale, and conversely things that work better in larger communities. Or you can extrapolate data from large, non-Wiki communities, but the wiki and the nature of MediaWiki in particular lead to different dynamics than might be found in meatspace or in other forms of computer-mediated communication.

    Other notable sociologists with interests in online communities have thought, spoken, and written about Wikipedia, but there is no silver bullet. Sunir Shah’s MeatBall is one example. There are others.

    The lack of any sort leadership, like the Wizards of the MOO, is by choice. The first step in creating such a thing would be to convince Jimbo. A second step might be to evaluate the role of the Foundation, which is essentially limiting its activities to financial, legal, and fundraising matters.

  5. I don’t think that rebuts the application of it to Wikipedia – I’m seeing the same problems here as everywhere, over and over. (Also read examples in the comments on the LJ link at the bottom of the post.) What it does rebut is applying the specific solutions. The problems are the same. Yay, we get to fuck it up from scratch!

    I understand Jimbo only read “A Group Is Its Own Worst Enemy” recently. I sure wish it had been earlier, but now is much better than never.

    Personally, I’d like to install the devs as Ultimate Authority. They are by structure anyway. There’s ranting by the Cunctator on the unfairness of this on the Nostalgia Wikipedia. Against Magnus, who I’d hope Cunc considers less evil now. Of course, the devs live by knowing a poisoned chalice when it’s described to them.

  6. David, the problem is not “cabalism” but the self-definition of some cabals as the “worthwhile editors”, which is accompanied by a definition of others as not worthwhile. Others apply the discussion you make here too firmly, believing that an “inner group” has the lock on rightness, and outsiders to that group cannot ever be right where there’s conflict.

    You talked about constitution, and two things strike me about that. First, that the constitutive articles are poorly formed and conflict with one another, which are problems. When a person with a particular POV sees that all POVs should be fairly represented, they find it hard to understand why presenting their POV fairly is not permitted, and why the power structure supports that. They aren’t aware that the “old boy network” trumps NPOV, so that editor X, clearly a POV pusher, friendless and not clued in on how to work the system, gets his arse banned, while editor Y, clearly a POV pusher, with lots of friends and clued in, is commended for fighting off “trolls”. Second, a ton of other “policy” documents have spun off from the original consitutive articles, and have created a huge bureaucratic nightmare. There is a huge leap from “you should try to get on” to “you should not breach policy 321B on civility”. On a talkpage the other day I saw a person suggest that the talkpage for KMRFC4 should not have been deleted under section G8 of CSD. Someone else said, no, it was deleted under IAR. And the first guy goes, yeah, okay, if it was under IAR.

    Dude, an admonishment to ignore all the rules if they get in the way is now a statute — a rule in itself!

    In my view, and not being a KoolAider, I know it counts for little, you need to recast the constitutive articles. They are nonsense as they stand and allow too much room for (mis)interpretation. Particularly NPOV. It conflicts with the sources policy really badly. Either you are presenting all views, or you are presenting the orthodox views of the Western world. Britannica has no problem stating that it does the latter. You also need to stop pretending that you have a small open-source workgroup. Give up the idea that you write policies by consensus. Restate the policies that you actually want followed and bin the rest. Shift as many “decisions” as possible into policy and away from OMOV: for instance, adminship should be conferred at an edit or time limit and removed in prescribed circumstances, or as I’ve suggested several times, rotated among active editors; articles should be deleted only on strict criteria, not on the judgement of editors, and not by voting them in or out.

    Also, you very much need to do something about “group ownership” of articles. Some articles — article sets even — are strongly biased to one POV, simply because a group of editors can control them. 3RR was a terribly ill-conceived policy because rather than cool down editwarring, which it does not, it has created a tool for group ownership. If one person opposes another, all that either needs do to impose their view is find one other person to help them group-revert the article. We all know articles like this, and some of the editors involved are not widely moaned about because the moaners are “trolls” but because they are doing something worthy of moaning.

  7. I’ve just reread Cunc’s essay. He doesn’t attack Magnus as a person, he talks about the essential powerlessness of the non-developer. Unfortunately, I’ve still not heard a cure for this that doesn’t involve learning to program.

    (And, on Wikimedia, learning to program really well, because our lack of resources makes it nightmarish to do anything interesting, e.g. replacing categories tags you can run queries on while we live on MySQL.)

  8. I’m not sure there is a solution for these — or any other set of problems — on Wikipedia for one reason that may surprise some:

    It’s amazing just how easy it can be to keep one’s head down, work in a specific area, & fail to notice the rest of the project burn down to the foundations.

    Maybe I’m the Wikipedian equivalent of the old American Mountain Man, who would feel crowded whenever he saw the smoke of a neighbor’s fire; I know I’ve intentionally focussed my efforts on the more esoteric topics of Wikipedia. Although I doubt it: I’ve encountered a few vandals & counterproductive would-be contributors. Yet whenever I find myself stumbling over the smouldering embers of the latest flamewar, trying to figure out just what happened so I can determine if this conflict was something important, I’m amazed at how poorly any news travels amongst Wikipedians.


Leave a Reply

Your email address will not be published. Required fields are marked *