This week might just have been the spark for the next phase in Wikipedia’s evolution. The public embarrassment for the online encyclopedia that “anyone can edit” was the Washington Post publicizing short-lived errors in the articles for Ted Kennedy and Robert Byrd, two US senators that had to excuse themselves on inauguration day for health and other reasons [ref]. To be fair, the WaPost article exaggerated the importance of this run-of-the-mill vandalism in Wikipedia that existed only for a number of minutes. There were probably a few hundred or a thousand people who might have seen that error, but likely no more. Far less harmful than an errant AP news or wire story.
But this was reason enough to cause Jimmy Wales to “sigh” and push for the implementation of Flagged Revisions, a feature that would show the casual readers of Wikipedia a version of an article that had been approved (ie. flagged) at a very basic level by editors. Even before this incident, there was a fairly well publicized straw poll to gauge interest among the frequent Wikpedians. (I say “fairly well publicized” because the announcement showed up primarily when people loaded their Watchlist, which I don’t often do. I was informed by other Wikipedians by chance.) Wales said on his talk page:
This nonsense would have been 100% prevented by Flagged Revisions. It could also have been prevented by protection or semi-protection, but this is a prime example of why we don’t want to protect or semi-protect articles – this was a breaking news story and we want people to be able to participate (so protection is out) and even to participate in good faith for the first time ever (so semi-protection is out).
We have a tool available now that is (a) consistent with higher quality (b) will allow us to allow more people to edit it a wider range of circumstances and (c) will prevent certain kinds of BLP harm.
Flagged revisions has roughly a 60/40 for/against at the straw poll. I was dismayed that it took place on a rather obscure location: a subpage of a talk page of a policy page, and didn’t quite get the due diligence of a widespread announcement. Nevertheless, there are some good points raised by folks in reaction to the poll — for substantial policy changes in Wikipedia, how realistic is it to get the overwhelming consensus of 75% or more? Things in Wikipedia have almost always moved forward with the will and prodding of Wales, as a benevolent dictator, without such high consensus numbers. This was the case with the Seigenthaler incident that led to anonymous users not being able to create articles, and also the biography of living persons policy that has become a solid part of the editing guidelines.
Noam Cohen at the NY Times has a good writeup on this as well, describing the dynamics of the debate. In the end, there is going to be a bit of a crisis about jurisdiction, as Wales has requested:
“To the Wikimedia Foundation: per the poll of the English Wikipedia community and upon my personal recommendation, please turn on the flagged revisions feature as approved in the poll.â€
Why is this an important turning point? It’s arguably the first major change to en.wikipedia that seeks the authority of the nonprofit Wikimedia Foundation, with a strong hierarchical management layer of executive director (Sue Gardner), deputy director (Erik Moeller) and chief technology officer (Brion Vibber) to act on the request of Wales and the will of the community per the straw poll. This is quite different than in the past, with Brion Vibber and other technical team members making a field decision, based on Jimbo’s recommendation and testing the winds of the community.
It will be interesting to see if this dual-appeal to official authority (WMF) and authority based on historic social capital will win out and become the norm.
Meanwhile, it’s good to see Brion Vibber approach this carefully and with sensible goals. His response:
I just want to note that we would not turn FlaggedRevs on here on enwiki before working out some very specific parameters for the test first. Keeping an eye on workflow and seeing what can be streamlined or taken out would be very much part of our attention.
This is a notable word of caution by Brion — flagged revisions will be no panacea. It has the risk of submerging and pushing the problem of edit warring and inaccurate edits into another unknown, and perhaps harder to track, domain if editors are not used to seeing how flagged revision modifications are reflected in the recent changes list or their watchlist. Are the tools ready to track and flag? Is the community “tight” enough so that flagged revisions can be turned on with the same concept of what being “flagged” or “sighted” means?
For English editors who voted “yes” — you should really look at the German experience to know what you voted “yes” on exactly. On the Wikipedia Weekly podcast episode 56, we had the pleasure of interviewing Phil Birken of the German Wikipedia who was in charge of turning on and monitoring flagged revisions there. Take a listen at the Wikipedia Weekly web site.
Andrew, the poll had also been listed at the Community Portal for pretty much the whole time it was running. It sort of points to the limits of the extremely fragmented lines of communications on so complex a wiki as English Wikipedia. But we do have the technical capability to make publicity a little more uniform, which might be a good step at this point.
Good post. I think it’s great that you point out Brion’s role in all this, I’ve always been impressed with his attitude and execution to the the project and the software. One important point I’d add:
Feel free to call me out on this if it sounds like a way out-there conspiracy theory, but no one seems to have noticed that on the day after The Washington Post published a piece railing against us for our treatment of political figures (based on the exact same kind of vandalism we’ve experienced for years, nothing new there), they launch a wiki project of their own (WhoRunsGov) about the very same topic.
Looks like yellow journalism from where I’m standing, and frankly I’m surprised that Jimbo would be so reactionary. He’s weathered the storm of much worse criticism before…
Yeah. I somehow managed to only find out about the poll two days before it ended.
That said, any poll that gets several hundred votes on en:wp is likely significant.
Sage, yes I will admit that I do need to change my “habits” on where to keep on top of news in the community. WikiEN-L and Foundation-L just doesn’t cut it these days, and I abandoned the Community Portal years ago after it split into a half-dozen specific groups, which kind of took the idea of the “Village pump” and broke it down into “Metropolitan area borough-specific water spigot.”
Of course there was no way for it to scale.
Something we talked about on podcast before about “forums” is that if there was some easy way to just search the discussions and talk pages, that might help a lot.
There is now, however, the [[Wikipedia:Centralized_discussion]] page which did indeed list the straw poll, and we should use more and more. It is probably the closest thing to a fair notice, “Hear ye, hear ye” board we have.
It seems there is also a new movement for “flagged protection” which seems like a workable King Solomon approach to implementing a version of flagged revisions.
Steven, you know I had not made that connection, but I did find it curious that they started a wiki-like project like Whorunsgov.com. Rather than some type of conspiracy, I do think it was likely in the front of the journalist’s mind that his own company was in the midst of launching a more mainstream site for exactly these type of profiles.
Flagged Revisions for Biographies of Living People have been heavily pushed, alongside Semi-Protection of BLPs, by Wikipedia Review for a long, long time. The poll was the culmination of a long campaign and some of the key protagonists were long time Reviewers.
At the root of this campaign is to end Wikipedia’s unethical practice of compiling information on living people to sit at the top of google searches, which allows anyone with a grudge to anonymously defame the victim within seconds. Flagged Revisions is step towards ending that insidious and disgraceful feature of Wikipedia.
This practice happens repeatedly, as evidenced by numerous studies. Damaging edits from drive-by “vandals” have been proven to stick for days, weeks and months. Long enough to be carried to many mirror and scraper sites.
This is the real issue. Forget about all that “consensus” garbage and the “feelings of your community”.Free editing on BLPs should never have been allowed in the first place.
I was going to post this on-wiki but this is a better place to rant…
A trial is fine. Finding a way of protecting BLPs from plausible and persistent vandalism is a good idea. But I’m irked at the growing BLP absolutism at Wikipedia. People are seriously considering getting rid of BLPs, or imposing draconian editing restrictions (i.e. real names only) as a way of solving the problem. There are people out there on Wikipedia Review suggesting that the world would be better off if Wikipedia was destroyed, because that would get rid of BLPs. (I miss the days when the critics’ sites were more concerned with ”opposing” censorship). I’m especially irked at the idea that those who are lax on BLPs have no ethical compass. The New York Times or CNN surely have led to a lot more hurt feelings than Wikipedia. Google also has a huge potential to “do harm” to living people. Do these organization engage in such ethical handwringing about their very existence? Much less blogs or gossip columns or tabloids (if you think that MMORPG-playing nerds with no social skills violate BLP). Not that those should be the model for wikipedia.
Wikipedia’s strength is its openness, not its 100% accuracy. The only way to deal with the problem, ultimately, is to say “no big deal, it’s just Wikipedia” and “nobody should trust Wikipedia completely”. There’s no way to make Wikipedia absolutely perfect without destroying it. Do we want the Wikipedia that makes the internet not suck, by giving the world access to information on every conceivable topic (including much information that was nonexistent, or in even more unreliable form, or very difficult to find on the Internet)? Or do we want a perfect Wikipedia that garners no bad press, and is absolutely immune from lawsuits, and doesn’t hurt anybody’s feelings, but only has a fraction of the content?
The image of Wikipedia you get from reading Wikipedia Review, or the press, or OTRS complaints only represents a small fraction of Wikipedia’s true impact in the real world. Most of that’s just drama. But more and more, its all that Wikipedia insiders care about, even the ArbCom and Jimmy Wales.
@Will
“imposing draconian editing restrictions (i.e. real names only) as a way of solving the problem.”
Wikipedia has fostered and promoted a clearly unethical practice of allowing anonymous unaccountable people to manage the most prominent information available on named victims. This is unethical by any standards in any society. A WP biography will be read by everyone with even a passing interest in the person. The people who read this information are told that Wikipedia is an “encyclopedia.”
People are either promoting the continuation of this unseemly, damaging practice, or they are campaigning against it. Take your pick.