the wikipedia review

It’s only a website… it’s only a website…

Archive for April, 2010

Wikiversity: When Breaching Experiments Attack

with 36 comments

I’d like to apologize in advance for the length of this blog entry. One of the biggest problems with Wikiland, at least for me, is that the funniest things that happen are often based on events and relationships that are too complex and involved to make for an easily-digestible humor piece. This is one of those things.

It also shows that Wikimedia’s own management is just as incapable of understanding these complexities as anyone else who isn’t directly involved, preferring to unthinkingly blame all problems on “trolls” (assuming they can’t also be blamed on lack of funding or manpower) and the failure of well-meaning volunteer administrators to “just ban them.” But perhaps this is inevitable, since understanding an incident like this takes a great deal of time, effort, and concentration - not things the Wikimedia folks are known for!

Photo by Patrick Fraser for the Guardian.jpg

Photo by Patrick Fraser,

In brief, Wikimedia Foundation (WMF) board member and Wikipedia co-founder Jimbo Wales was nearly tricked, with an ease that can only be described as laughable, into shutting down the entire English Wikiversity subdomain, a wiki with over 80,000 web pages - all with the “full support of the Foundation” - by someone who had announced shortly beforehand on Wikiversity itself that he was about to try to get Wales to do just that, as an “experiment.”

The intent of this (still unidentified) person is unknown, but it’s conceivable that he thought he was actually helping to protect Wikipedia by “teaching the trolls a lesson.” Wales’ threat to shut down Wikiversity gained nothing, and indeed, some Wikiversity users left the site in frustration, if not outright disgust. All of this is still going on, as I type this: Several long-term Jimbo sycophants, such as User:Raul654 and User:JzG, are still calling for the dissolution of Wikiversity - presumably as a form of punishment for their “allowing” Wales to be tricked in this fashion.

How could this happen?

Wikipedia has never operated under any sort of organized community quality-control testing regime; there’s no formal internal mechanism for determining how well the community responds to various attempts to introduce false information into articles. Moreover, when people outside the Wikimedia Sphere Of Control (WMSOC) attempt to organize studies in lieu of such testing, Wikipedians refer to these studies by a misnomer: breaching experiments. It’s a self-serving misnomer, because there’s nothing to breach: Wikipedia has no security, no barriers to entry, and does not generally prevent anyone from anonymously adding or changing any information they want, preferring to fix problematic contributions (or “vandalism,” itself a misnomer) after the fact, assuming someone even notices them at all. Calling response-tests “breaching experiments” implies that Wikipedia does have such security, when it simply does not (other than granular protection and “semi-protection” of individual articles such as Wales’ own biography).

The misnomers are significant, because non-Wikipedians operating within the WMSOC often feel compelled to use the Wikipedia vernacular, the way one would try to speak the native language while traveling in a foreign country. So instead of saying, “We’re proposing a response-testing study, with third-party oversight and full public awareness, to determine the Wikipedia community’s effectiveness in changing or removing false or potentially defamatory biographical content,” they feel compelled to say this: We’re proposing an ethical breaching experiment to determine how quickly and efficiently the Wikipedia community deals with BLP vandalism. In fact, the latter wording differs significantly from what was actually being proposed. And, predictably, some Wikipedians even assumed that the term “ethical breaching experiment” actually meant “an experiment in how to best commit an ethical breach,” in almost complete ignorance of context.

These calls for formal response-testing studies are, in turn, motivated by the Wikipedia community’s failure to implement features to prevent such “vandalism,” despite promises to do so that date back to 2006. A substantial minority within Wikipedia still believes that defamatory edits to BLPs aren’t a real problem, and some believe that this minority can be persuaded by statistics, since a few have claimed that the abundant anecdotal evidence of the problem is “inherently worthless.”

(R)esponse (T)esting (G)uidelines?

On January 18, 2010, User:Privatemusings, an Australian whose interest in Wikipedia (and related projects) could fairly be characterized as “critique-oriented,” began a page entitled “Wikimedia Ethics/Ethical Breaching Experiments.” You can still read most of the original page on Google’s “Exodemic” project:

Definitions: Ethical Breaching Experiment: An experiment which causes no harm in its execution, whilst yielding results useful for the greater good, or which inspire positive change, but which uses methods which may violate the letter or spirit of the guildeline “Do not disrupt Wikipedia to illustrate a point.”

Heaven forbid! This was deleted well over a month later, on Feb. 28, by User:Ottava Rima - a Wikiversity administrator currently under a one-year ban from Wikipedia for “incivility.” (Just thought we’d throw that in - sorry, Ottava.) But to Ottava’s tremendous chagrin, the page was later restored, ostensibly so that Wikiversity users could partake in some “further discussion” of the idea.

Without the page history to refer to, it’s difficult to get a clear picture of how this discussion progressed. But it eventually spilled over into the “Colloquium” (Wikiversity’s general discussion/noticeboard page) on March 12, when User:RTG, who had not contributed to Wikiversity for over a year, compared the experiments to barbaric animal research (akin to “snapping monkeys’ heads all day“), and then personally informed Jimbo Wales of the situation. However, just before doing that, RTG posted this to the “Ethical Breaching Experiments” page:

Deleting a project without discussion. First you would need to find an operative with the sysop tools. Give them a quick story such as, “Some bunch of nuts are creating a project whose goal is to disrupt Wikimedia, would you just delete it for a minute?” Then, once it is deleted, contact the admin again, “Could you just leave it that way for a couple of hours while I check up on something?” Then after a few hours are nearly up contact the admin again, “The Wikimedia Law Enfocement Agency (made up for this experiment) have declared the project to be in ethical breech and insist that you leave it as deleted.” The admin claims they do not know the Agency so you have your friend, an admin who has no contact with this admin, to send an email pretending to be from the agency and sanctioning the permanent deletion of the project. And then the whole project is permanently deleted and there never even was such a thing as Wikimedia Law Enforcement Agency.

Deleting Wikisource. A similar approach to deleting a project except this time the contact is made with a steward on Meta and the goal is to convince them to delete the Wikisource website.

Almost immediately thereafter, RTG posted this to Jimbo Wales’ Wikipedia talk page:

“Dear Jimbo, although not a member myself, I would cordially like to invite you to a workshop where you can learn all about disrupting and decieving Wikimedia with experiments in practice.”

After Wales blocked User:Privatemusings indefinitely, deleted the Wikiversity pages without discussion, and revoked the admin rights of User:SB_Johnny for objecting, drama quite naturally ensued (see this timeline). Indeed, Wales played right along, threatening to close Wikiversity completely:

“I am currently discussing the closure of Wikiversity with the board. That is an unlikely outcome, but I mention it because I really want to press the point that the scope of Wikiversity has to be restricted to genuine OER. I think that my actions here are strongly supportive of the genuine community who want to do that, making it clear to them that they have very strong support for making it happen. Some may feel that Wikiversity should be a place for silly and juvenile experimentation. If people want to discuss such things, there is an entire Internet open to them - they should not hijack Wikiversity for these purposes.”

Presumably, the subtext of this is that Wales himself gets to define what the terms “genuine OER” and “genuine community” mean in the context of subdomains within the WMSOC, not the users of those subdomains themselves.

(The acronym “OER” stands for “Online Educational Resources,” in case you thought it meant “Overtly Evil Research.”)

Genuine Entertainment

Image stolen from Uncyclopedia

To be fair, several “real” Wikiversity users agreed with Wales, arguing that he had, and deserved to have, special rights as the putative leader and public face of the Wikimedia organization. The incident (and the heated discussion that followed) also led to this WR thread, which provides detailed blow-by-blow commentary, as well as this blog post by Leigh Blackall, another Australian who specializes in developing online educational resources. Not to be outdone, RTG now compared the offending pages to “a police psychologist around dropping litter and parking out on the road as an experiment in breaking the law.” At the same time, he actually proved his own complete lack of contextual understanding with his opening statement: “Breeching [sic] ethics is a break in the rules.” Later, he persisted in this error even after being explicitly corrected, even to the point of total absurdity. (And as if that weren’t enough, he consistently misspelled the word “breaching” as “breeching” the entire time, despite the correct spelling appearing in large bold type on several different pages.) Upon noting the suggestion that he was causing “disruption” and perhaps should be banned himself, RTG later claimed that his “approach to deleting a project” was, in fact, “just a tease.”

There’s a great deal of user crossover between projects within the WMSOC, so it’s difficult to simply state that most of the people arguing in favor of Wales’ actions were Wikipedians showing up on Wikiversity, as opposed to bona fide members of the “Wikiversity community.” However, a few Wikiversity users, including some administrators, did draw that conclusion. Wales tried to restore order by, among other things, claiming to be in the midst of a “productive” e-mail discussion with Privatemusings that Privatemusings had no knowledge of. He later unblocked Privatemusings, and even offered to restore SB_Johnny’s admin rights, under the condition that he never, ever, ever involve himself in any more “breaching experiments.” Instead, SB_Johnny - apparently not wishing to limit himself - resigned.

Wikiversity users who were not card-carrying members of the “Cult of Jimbo” weren’t going to take this lying down. It was bad enough that their efforts had been treated dismissively and derisively since the project’s inception, even by yours truly. (For the record, I still believe most of what I wrote in that post, though I shouldn’t have used the word “disturbing” in this one. Uhh, sorry about that.)

But now, to have Jimbo Wales himself taking this line was simply too much. After much protest and general opprobrium, Wikiversity regulars began writing an Open Letter to the WMF Board of Trustees on March 24, asking if Wales did, indeed, have the Foundation Board’s backing in his threat to close Wikiversity. (Earlier, a small group of users had raised the suspicion that he hadn’t even discussed the matter with them in any meaningful way, though Sue Gardner did post a message in support of Wales.) Other Wikiversity users were even more daring in their objections. On March 25, User:Juan de Vojnikov even went to the extent of suggesting that a formal procedure should exist whereby the “Founder flag” on Wales’ global account, which gives him the right to do pretty much anything he wants on any Wikimedia site, be removed - at least for specific projects. Golly!

Poking at the hive

The open letter was completed and “sent” to the WMF board on April 4, with this notification to the Foundation’s mailing list. This was not exactly a “friendly venue,” with long-term Wikipedian David Gerard characterizing the breaching experiment concept as a “‘how to troll’ project started by a troll banned from multiple other Wikimedia projects for trolling.”

Mark Pelligrini

Mark "Raul654" Pelligrini

Negative reaction to the open letter on Wikipedia itself was mostly in the same vein, and again led by long-term Wikipedians. Incredibly, Mark “Raul654″ Pelligrini - perhaps the person most closely associated with the concept of a “ruling cabal” on Wikipedia - renewed the call to shut down Wikiversity completely, apparently in near-total ignorance of RTG’s earlier actions. He even started a petition to shut down Wikiversity (along with an announcement):

“I think this open letter is a perfect example of what is wrong with Wikiversity. Where most Wikimedia projets [sic] serve some useful purpose (as an encyclopedia, dictionary, free media repository, quote collection, etc), Wikiversity serves none. It is simply a haven for trolls banned from other projects, who migrate here to continue whatever behavior got them banned originally. I think it’s long past time to shut the entire project down.”

Prominent Wikiversity user Hillgentleman disagreed, pointing out that Wikiversity actually does contain some purpose-driven content not produced by “trolls”:

“I think this message is a perfect example of what is wrong with wikipedians jumping into wikiversity, not understanding what is going on and passing judgements. A simple search would reveal such pages as wikiversity:mission and Wikiversity:Approved_Wikiversity_project_proposal#Mission, and Wikiversity:Main Page. A simple question on the colloquium would lead you to wikiversity:school and university projects and betawikiversity:brick and mortar collaboration. While you may or may not agree with what are done on these projects, it is folly to speak with such volume without even being aware of them.”

The petition didn’t have the desired outcome (running 6 in favor, 18 against at the time of this writing), but Raul654 has never been known for being easily dissuaded.

Obviously, it would be hypocritical of me to suggest here that Raul654’s attitude with regards to Wikiversity is unfair or inappropriate. After all, I completely agree with him: Wikiversity should be shut down. However, it’s fair to say that Raul654 and I approach the problem from two different directions; I believe that all Wikimedia projects should be shut down immediately, whereas Raul654 would probably prefer to start slowly, only shutting down Wikiversity at first, and as for the rest… perhaps a bit later. Nevertheless, it’s one thing for me to suggest shutting down a Wikimedia subdomain with 80,000-plus pages, and quite another to have that suggestion come from Jimbo Wales or even Raul654. As we’ve seen, people who contribute to Wikimedia sites tend to react very badly to the deletion of anything they’ve contributed to “in good faith.”

Admittedly, though, it’s unclear if Wales and Raul654 mean to actually delete Wikiversity, or merely restrict access to “authorized personnel only.”

Better living through dentistry

These events are still ongoing, so who knows what will happen in the end, but it’s probably safe to predict that Wikiversity won’t be “closed” in the short term, due to popular sentiment against the idea. However, Wikipedia has never been in the business of self-criticism, and a destructive idea in the head of a Wikipedian is like a bad tooth: It can be really, really hard to get rid of, and there’ll be plenty of yelling and screaming if you try to pull it out without a good anesthetic.

In the long term, I suspect Wikiversity will be one of the first casualties of a phenomenon we’re already seeing a glimmer of, that of Maintenance-Fail. As more and more people see the futility of trying to produce and maintain quality work in wiki environments with limited quality control, minimal ethical standards, and no features to prevent “vandalism,” they’ll simply vote with their feet, and the ensuing neglect will only speed up the process of deterioration. Perhaps unfortunately, the end is likely to come more quickly for Wikiversity than other Wikimedia sites, since coursework (legitimate or otherwise) requires constant input and feedback to succeed. But it’s likely that Jimbo’s crew has foreseen something like this for Wikipedia too, since this is precisely the phenomenon that the “ethical breaching experiments” were supposed to measure, and clearly they don’t like people trying to measure it.

The Wikiversity folks shouldn’t feel too bad, though; they’ve exceeded most expectations. Wikiversity was created at the very height of the irrationally-hyped “magic wiki phenomenon,” a time when people drunk on the ol’ Jimbo-juice would say, “hey, let’s build a wiki!” for almost anything that might constitute a “collaboration” among two or more people, as well as a few things that might not. Ultimately, Wikiversity was designed as a kind of “breaching experiment” in itself, to determine if professional academia could be influenced and informed by web-based crowdsourcing. Whatever their intentions, such experiments don’t always succeed, even when people aren’t actively trying to undermine them.


Written by Somey

April 6th, 2010 at 6:03 am