Citizendium dumps Wikipedia
January 20, 2007
Larry Sanger has changed his mind about forking Wikipedia. As you'll recall, Sanger, who cofounded Wikipedia with Jimmy Wales, has been working on creating Citizendium, an expert-edited alternative to the motley online encyclopedia. The original plan was to "fork" Wikipedia - in other words, make a copy of Wikipedia's contents and then re-edit the articles to bring them up to snuff. The theory was that borrowing the existing articles would provide a jump-start for the fledgling knowledge base.
But that hasn't worked so well. The handful of early contributors to Citizendium have, says Sanger, become "disheartened by the fact that their first obligation seems to be to edit mediocre Wikipedia articles." The prospect of "cleaning out the Augean Stables" put a damper on their enthusiasm. It was, apparently, a little too Herculean of a task, even for experts.
So, as an experiment, Sanger is doing an unfork. He's erasing all the as-yet-unedited Wikipedia articles from Citizendium and letting people begin with a clean slate. "If we start over," says Sanger, "then we can create our own more distinctive culture, and we can take more pride in our articles and in the processes we develop. In short, we can be ourselves. And putting yourself into a piece of work is what gives you passion in creating it."
Sanger wonders if there's a larger lesson here. Maybe the forking that works for software projects is unsuited to media projects: "Content is different from software. Software does stuff, and that’s something objective that either works or doesn’t work. So you can take software and make it work better. By contrast, content expresses ideas, and people just might not want to start their work by seeming to express other people’s ideas."
Nick, we surely can Google for it ourselves, but a direct link to Citizendium in the article would be welcome.
Posted by: Sergey Schetinin at January 20, 2007 11:52 PM
I wonder if it's also that there are different types of person - some, clearly those attracted by Citizendium, who want to make something new and others who are content to edit, fiddle, fix and clean up. There are many excellent editors in this world, but perhaps Larry Sanger just hasn't managed to reach them.
Posted by: Bill Thompson at January 21, 2007 06:13 AM
I don't know the "right" thing for Citizendium, but to say that (paraphrasing here) "forking works for software not for content" isn't right. In fact, software is itself content and the systems world is crazily complex today because of the identical concerns that Sanger has. For example, large companies don't have one ERP system, they have dozens of functional equivalents. Companies of any size don't have one of anything... and they have thousands of some things (think Microsoft Access or Excel-based systems).
And software companies don't always fork-and-re-use, we often rewrite for the exact reason Sanger says: some engineer wants his or her "own voice" (that's a bad reason) and sometimes we re-write because it's ultimately cheaper and better to start over after learning something (that's a good reason).
The content-creation world can (or at least might) learn something from looking at this other electronic content, which has been worked on in wiki-like ways for decades (source control systems manifest not dissimilarly from wikis).
All human-generated intellectual property is content to someone. And duplicated, non-forked content creates the following dilemma: if we do not retire the original, then we increase complexity by increasing the number of silos that have some useful purpose, but are not wholly useful (if they were wholly useful across the enterprise or world, we wouldn't have needed to branch). Aside from this confusion, we increase cost because we maintain two systems. (In the Citizendium vs. Wikipedia debate, the global cost is increased because there will be dual maintenance of similar information).
There is no answer that I propose, other than to suggest that although highly collaborative technologies offer much for the future, the notion that hive intelligence always tends toward correctness or goodness is wrong on a global level... although probably right on a local level (I could also say "enterprise" and "workgroup"). We need to have a better understanding on the role that context plays in content-creation and refinement. This context is often hidden, defined by job, class, geographical or other invisible demarcations.
I am less interested (for now) in the wasted time anonymous people spend updating these web sites. I spend a lot of time researching and listening to old music. That's a hobby. I assume for many people that contributing to wikipedia or whatever is a similar interest to others.
But I am directly interested in how these collaborative technologies will evolve at corporate and governmental customers of ours. And from what I have seen, we still have a lot to learn about how these should be governed, from the outset, to achieve the correct amount of control over the user-generated-content model (to insure maximum "correctness"), and the correct amount of user-generated to overcome the control model (to insure maximum scale)!
We're still working on the problem in the software world... I assume we'll still be working on it in the content world many years from today.
Posted by: Phil Gilbert at January 21, 2007 09:40 AM
As I said in those posts, some Wikipedia articles have two major problems: one is that they're flat-out wrong (and consequently require either a thousand tiny changes or a couple of big ones); the other is that they're in Wikipedia (meaning that the thousand tiny changes would probably be challenged one by one, and the two big changes would almost certainly be reverted). The Citizendium fork removes the second problem but not the first.
The problem isn't finding 'excellent editors' - it's editing material that ought to be scrapped.
Posted by: Phil at January 22, 2007 06:48 AM
Does forking really work for software for the type of thing Sanger is trying to do?
Sanger wants to take the lessons learned, and try again from scratch to take a big leap forward. It seems to me that software innovations usually take the same path (from scratch) to achieve the same goal (leap forward). The main difference is that when software leaps forward, it is usually built on a better platform. A platform that provides elegant new facilities that allow the new project to avoid the plumbing misery of the old project*. Because the new software project can do more with less effort, and doesn't have to avoid things that were too hard in the past, it can truely innovate.
This is very different from "having better (software) experts", which off course does matter, but only goes so far. It is the progress in the platform, and the ease of use of that progress, that changes the game.
For Citizendium, the platform may be something different than software or technology. But it can be sure that if better experts are all it is banking on, then it will either not succeed as intended, or be surpassed by something built on a cleverer approach to the problem. Some Darwinian principle perhaps?
(*) For example, from text interface to windows, from custom windows with text to HTML, from HTML to dynamic technologies.
Post a comment
Thanks for signing in, . Now you can comment. (sign out)(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)
"Riveting" -San Francisco Chronicle
"Rewarding" -Financial Times
"Ominously prescient" -Kirkus Reviews
"Riveting stuff" -New York Post