The truth about Wikipedia's flagged revisions

According to blogs such as Citmedia and mainstream news outlets like CNN and the New York Times, Wikipedia will soon begin requiring that all changes to articles on living people be approved by an aristocracy of established editors before going live. The move has been criticized in almost all these stories. I was working on an article criticizing the move myself when I learned something interesting: the story is false. In this article, I will try to explain what is really happening.

Wikipedia’s open policy to editing has allowed impressive growth and responsiveness. In fact, it has proven so responsive that information—of celebrity deaths, for example—is sometimes posted before it happens. The resulting bad publicity and lawsuits have led to a lot of discussion in the Wikipedia community about how to address the problem. The current confusion appears to stem from a New York Times article that conflates two of the approaches suggested by the community.

Two approaches

The first is called “flagged protection“. When this feature is enabled for an article, edits are possible but they will not be visible to the general public until an established editor flags the article as free of vandalism. This approach—the one discussed in the media—has been around for quite a while. It was adopted by the German-language Wikipedia in 2008 and following some high profile vandalism in January 2009, Wikipedia founder Jimmy Wales strongly advocated its adoption on the English version.

The second approach is called “patrolled revisions“. It uses the same flagging system as the first but the flags are informational only; edits go live immediately but visitors can see whether the article has been vetted or not.

Both of these are provided by installing an extension to the current Wikipedia software called “flagged revisions“.

What really happened

The Wikipedia community agreed to a trial run which entails installing the software extension and combining the two above approaches. Currently, certain high profile articles are locked: no unregistered or new users can edit them. The trial will replace article locking with flagged protection. This will actually open Wikipedia further by allowing new and unregistered users to suggest changes where they could not before. The trial will also have patrolled revisions applied to articles on living people. The trial proposal document is titled “Flagged protection and patrolled revisions” which should not be confused with an earlier proposal entitled “Flagged revisions trial“.

As I’m sure you’ve guessed by now, the whole misunderstanding was just a mistake based on confusing names.

This case provides an interesting example of how fast misinformation can spread. The New York Times article was published on August 24th. The next day it was cited in numerous blogs and within one more it was being discussed on CNN. It also highlights how quickly the media (old and new) move on to a new subject. The Wikimedia Foundation issued a clarification—which initially contained the same mistake itself—on the 26th and my research has only turned up a single blog post that cites it.

This phenomenon will continue to pose a problem in the future. Hopefully we can find a way to alleviate it.

  1. A couple of corrections:

    1) Jimmy Wales should properly be termed “co-founder”, the other co-founder was Larry Sanger. See Larry Sanger’s page:

    2) “. This will actually open Wikipedia further …”

    The phrasing here is misleading. These changes were developed to deal with widespread vandalism, up to and including true libel. A small side-effect is that they will allow some loosening elsewhere. But that is not the main effect, or their intent.

  2. “This will actually open up Wikipedia further…”

    This phrase is not misleading, because the previous method of vandalism prevention was to lock articles, barring them completely to users not registered for a certain amount of time.

    That isn’t desirable for what the project is attempting to do, so this new method allows high-risk articles to be protected without instituting a universal ban.

    And in moving from the universal ban to a review process, Wikipedia is opened further, and that is the main effect, and that is their intent.

  3. Luke, the implicit fallacy being smuggled-in there is that the new method would only ever be used on the exact same articles as are locked now, instead of being applied much more widely. Whereas it seems pretty obvious that in practice, many more articles will be restricted. The whole point is to have a method which can be applied to more articles because the Wikipedia admins are very reluctant to lock articles. Hence the net effect will be more restrictions.

  4. I never intended so much to be read into that sentence. I was thinking of the contrast between what was reported and the actual (short-term perhaps) results of the change. The NYT and blog articles said the move was Wikipedia locking everything down. It is not the case. In theory, it’s a move in the other direction. Of course, it is also true that by reducing the severity of the locks they will also reduce the inhibition towards locking.

    One can also see the implementation of patrolling as a dry run for protecting an entire category. Right now, people are worried about delays, but if articles can be flagged in an acceptable time period (according to the WP community), then it will be easier to sell a broad protection scheme.

  5. Oh, I just noted the sentence to make clear the context for where I was replying. This is a case of truth being, though not exactly in the middle, definitely somewhat apart from two extremes. I agree with you that the media reporting has gotten the story wrong, by greatly exaggerating the extent of the restrictions. HOWEVER, Jimmy Wales and other Wikimedia people are spinning just as hard in the opposite direction, by trying to portray what is going to be more restrictions, as an up-is-down increase in openness. That aspect is laughable, and is as inaccurate in the reverse way. No personal criticism intended, I was only trying to outline that the other side is problematic too.

  6. I still think flagged revisions is a bad idea (we seem to have tamed vandalism without using it, and this will absolutely reduce edits by new users when they don’t have the satisfaction of seeing the effect of their work), but this post does a good job of explaining it in lay terms.

Comments are closed.