February 09, 2010
Catching Up With Wikipedia
By Eric Goldman
I recently posted the final published version of my article Wikipedia’s Labor Squeeze and its Consequences. In the course of updating the draft, I reviewed the news coverage of Wikipedia from the second half of 2009, and I thought I would share some of the more interesting tidbits that caught my eye.
Flagged Protection v. Flagged Revisions
The biggest news from the second half of 2009 was the August announcement from the Buenos Aires Wikimania that the English-language Wikipedia was rolling out Flagged Revisions for living people's biographies. It was first reported by Noam Cohen at the New York Times and then repeated in countless articles. I think the announcement took everyone by surprise because it seemed to come from out of the blue. Several other Wikipedia versions deploy Flagged Revisions, but before Wikimania, the consensus had been to try Flagged Protection and Patrolled Revisions--not Flagged Revisions--on the English-language Wikipedia.
[Some of you may be wondering: what's the difference? A lot! With Flagged Revisions, most user edits are invisible to logged-out readers until a more trusted editor approves them. Flagged Protection is a variation of Full Protection and Semi-Protection. Edits from less trusted users to protected articles are invisible to logged-out readers until a more trusted editor approves them. Flagged Protection substitutes for Full Protection, where less trusted users cannot make any edits to the protected articles whatsoever; and semi-protection, where only some users can make edits to those articles. The percentage of articles currently fully- or semi-protected is very low--I believe less than 1%--and presumably Flagged Protection would be used equally sparingly. In contrast, the media announcements indicated that Flagged Revisions would apply to all living people's biographies, a significant minority of Wikipedia entries. Patrolled Revisions is more procedural than substantive; it's merely a way for editors to communicate with each other that they have verified previous edits.]
The fact that Wikipedia apparently leapfrogged Flagged Protection to adopt the much more restrictive Flagged Revisions seemed like an ominous development and perhaps an indication that the vandals and spammers were winning more than we thought. The public angst over Wikipedia's announced move was deafening. People seemed to be genuinely shocked that Wikipedia might make a wholesale move from an open edit site to something substantially more restrictive.
However, this angst was misdirected. Despite multiple efforts to get clear information from Wikipedia (including this unbelievably confusing blog post), it turns out that the English-language Wikipedia isn't adopting Flagged Revisions at all but instead is proceeding with its trial of Flagged Protection. I think Farhad Manjoo properly captured our collective frustration in his TIME article entitled Jimmy Wales Quietly Edits Wikipedia’s New Edit Policy: “In several interviews, including many with TIME, officials at the Wikimedia Foundation, the nonprofit that manages Wikipedia, explained that the user-edited online encyclopedia would soon impose restrictions on articles about living people.” I really don't understand how or why Wikipedia successfully misdirected us for weeks, but they could have easily avoided a lot of confusion with a clearer announcement and more prompt corrections.
So the bottom line is that, despite the August drama, Wikipedia is not flipping its default to closed editing yet.
Nevertheless, as I explain in my Wikipedia Labor Squeezes article, Wikipedia inevitably will adopt more restrictive editing policies. It's just a matter of time. Accordingly, I will not be surprised when Wikipedia announces more restrictions. In contrast, judging from the reaction to the August announcement, I expect most people will be shocked when they next hear of Wikipedia's next effort to deploy a more restrictive editing policy.
In late August, Wired reported that Wikipedia was adopting Wikitrust, a tool that automatically color-codes edits to an entry to show the projected credibility of those edits. Thus, at a glance, a reader can tell which parts of an entry are more likely to be accurate and which parts should be scrutinized more closely. Amazingly, though, this was also a botched announcement (read the update to the Wired article). Wikipedia is only conducting a trial of Wikitrust.
A Couple Other Interesting Factoids
From New Scientist: ""Occasional" editors, those who make just a single edit a month, have 25 per cent of their changes erased, or reverted, by other editors, a proportion that in 2003 was 10 per cent. The revert rate for editors who make between two and nine changes a month grew from 5 to 15 per cent over the same period."
From NY Times: Wikipedia contributors are 80% male, 65%+ single, 85%+ childless and 70% under 30 years old. As I explain in my article, this is a group that will experience significant life changes, which in turn will reduce or eliminate the time they have to contribute to Wikipedia. Will sufficient numbers of replacements emerge?
From an article entitled The Singularity is Not Near: Slowing Growth of Wikipedia
- "the number of active editors and the number of edits, both measured monthly, has stopped growing since the beginning of 2007"
- "The rate of reverts-per-edits (or new contributions rejected) and the number of pages protected has kept increasing. Occasional editors experience a greater percentage of reverts per edits in comparison to the more prolific editors."
The Wall Street Journal Article
On November 23, the Wall Street Journal published an A1 article entitled I spoke with the reporter in June but my remarks only made the sidebar. The lead paragraph thesis is that "unprecedented numbers of the millions of online volunteers who write, edit and police [Wikipedia] are quitting." Citing research by Felipe Ortega, the article says that the English-language Wikipedia suffered a net editor loss of 49,000 in Q1 2009 (compared with a net loss of 4,900 in Q1 2008). It's worth reading the entire article. It stirred the pot quite a bit.
Erik Moeller, Wikipedia's Deputy Director, responded to the article with (among other things) a different cut at the numbers:
Studying the number of actual participants in a given month shows that Wikipedia participation as a whole has declined slightly from its peak 2.5 years ago, and has remained stable since then. (See WikiStats data for all Wikipedia languages combined.) On the English Wikipedia, the peak number of active editors (5 edits per month) was 54,510 in March 2007. After a more significant decline by about 25%, it has been stable over the last year at a level of approximately 40,000.
Consider this blog post my 4+ year check-in on my prediction in December 2005 that Wikipedia would fail in 5 years. In that post, I didn't tightly define what I meant by "failure," and frankly my "failure" rhetoric has been unintentionally and unnecessarily inflammatory. In all cases, I'm not rooting for Wikipedia's failure (however defined).
However, I remain baffled by the folks who are so enraptured by Wikipedia's mystique that they believe the site will defy gravity. Whatever you take away from the data points I cite in this post, I think it's undeniable that Wikipedia is changing in material ways. Bright minds might disagree about whether those changes are good or bad. From my perspective, Wikipedia's evolution has followed a fairly predictable path. Like many UGC websites, contributor activity peaks and then declines, and the transition from first generation contributors to second generation contributors naturally has some bumps. More structurally, Wikipedia has followed an entirely predictable evolution of progressively tighter editorial policies, and I anticipate even tighter editorial controls are come (to be accompanied by shocked public outcries each time).
As for my prediction, I'm waiting to see what develops this year, especially at Wikimania in August. Whether I'm right or wrong, I'll post my 5 year assessment of my prediction in December.
TrackBack URL for this entry: