Monday, January 18, 2010

On Citizendium

Once upon a time, a rival arose to take on Wikipedia.  It was a scrappy upstart led by the same Larry Sanger who co-founded Wikipedia where editors had to use their real names and experts were given additional authority (solving two of the most commonly raised objections to Wikipedia).  Ars Technica proclaimed: "There's no ifs about it: Citizendium is being created to outdo Wikipedia" and Larry Sanger himself declared "Citizendium will soon attempt to unseat Wikipedia as the go-to destination for general information online."

The earliest days  of Citizendium were accompanied by a certain fanfare.   Within its first year, it drew extensive stories in The Guardian, the Financial Times, the Irish Times (curiously, Citizendium seems to have drawn more press in the UK than the US), but by 2008 Citizendium was already dropping off the map; the press had lost interest.  In the month of December 2008, less than 200 editors made a single edit, and the figure has remained roughly stable since.  The graph below shows the number of editors making at least one edit to Citizendium per month from October 2006 until December 2009:

Image Credit: Aleksander Stos (Citizendium).  
 for details and licensing information.

As of this writing (more than 3 years after its inception), Citizendium has 13,051 "live articles", defined as either articles started on Citizendium or those which have been taken from other sources (e.g., Wikipedia) but altered significantly in three or more places.  The encyclopedia ranks at around 60,000 in the Alexa rankings and is growing at a rate of roughly 15 articles per day.

At a comparable point in its history (March 2004), Wikipedia had 239,255 articles (it reached 13,051 articles in just under 10 months).  It ranked in the Alexa top 1000 (somewhere 600), and was growing at somewhere near 500 articles per day.  By all these measures, Citizendium can not hope to compare to Wikipedia, and significantly, at present it is stagnating as the graph of editors shows and a graph of edits per day shows even more clearly:

Image Credit: Aleksander Stos (Citizendium)
for details and licensing information

The only argument that can be made on behalf of Citizendium is that it is of higher quality than Wikipedia (as in terms of reach and quantity it falls far short).  This is a hard claim to evaluate, but it is dubious at best.  It may, in fact, be the case that the average Citizendium article is better than the average Wikipedia article, but this is meaningless, given that Wikipedia includes so much more.  It would be useful to directly compare Wikipedia articles and their Citizendium counterparts, but it is far from evident how one might do this.  There are clearly Wikipedia articles which are far better than their Citizendium counterparts and vice versa.  A random sample of Citizendium articles which I took for the purposes of comparison was not encouraging for Mr. Sanger's latest effort (assembled via the Citizendium "Random Page" link).  I'll let you form your own conclusions

  1. "Belgian Cuisine": Hardly a gem of an article.  It appears to have been taken nearly word for word from on November 17, 2007 from a Wikipedia version dated November 7, 2007.  Since then, the Citizendium article has not changed at all, but the Wikipedia version has at least been supplemented with a few pictures and a citation.  Neither is what anyone would call a "good" article, but the Wikipedia one is at least slightly superior.
  2. "Certificate revocation list": A single sentence (which so far as single sentences go is not a bad one), but it cites no sources.  The Wikipedia article does not cite any sources either (to be fair, it uses external links to fulfill a similar purpose), but is substantially longer.  I will not pretend to be an expert on this subject, but the Wikipedia article was more informative, and followed closely with what I was able to find elsewhere.
  3. "Maple Syrup": A much finer article than the previous two.  The Wikipedia article is longer and conveys more information,  but is not quite as well-written.  On the whole, I find the Wikipedia version superior, though I invite you to form your own judgements.
I could continue with this exercise, but I think after even just three articles it is clear enough that Citizendium articles do not, any sense, tend to be significantly superior to their Wikipedia counterparts (if anything, Wikipedia might be able to lay claim to this title).  Citizendium, I must admit, has certain advantages over Wikipedia (most notably its freedom from obvious vandalism), but on the whole it must be judged a failed endeavor.  

What does the failure of Citizendium mean?

One of the clearest lessons of Citizendium is that Wikipedia is an anomaly.  Larry Sanger, co-founder of Wikipedia with the benefit of several years of hindsight, was unable to repeat the trick despite possessing numerous advantages over the situation of Wikipedia in 2001 (an established name for himself, far more early press attention, a better financial picture in the early years, a larger base of initial contributors, etc.) and only one disadvantage over the situation of Wikipedia in 2001 (the presence of Wikipedia itself).  If Sanger could not succeed despite all of this with Citizendium, it seems clear that he and Wales had no idea what they were doing the first time around (a fact clearly born out by the focus on Nupedia in the early history of Wikipedia).  The failure of other "Wikipedia killers" such as Google Knol (which had even more advantages than Citizendium) tends to confirm that there is something special and anomalous about Wikipedia.

What is it, then, that makes Wikipedia so special?  It is tempting to say that Wikipedia simply enjoys a privileged status because it was the first online user-generated encyclopedia, but this is simply not true.  The "Interpedia"was planned as early as 1993, though it never left the planning stages.  Richard Stallman's GNUpedia was also planned much earlier than Wikipedia (in 1999) and went live around the same time as Wikipedia.  It however, fizzled and was eventually more or less consumed by Wikipedia.  It is also worth mentioning the Stanford Encyclopedia of Philosophy (a highly successful, though specialized internet encyclopedia founded in 1995, though not "user generated" in the same sense as the other projects).  So, while Wikipedia may have been the first user-generated online encyclopedia to "take off", it was not truly the first; it did something right that these other encyclopedias did wrong.

Thus, I present the following as uniquely important aspects of Wikipedia's unparalleled growth, reach, and success, not shared by Citizendium (ignoring those aspects that are shared, such as wiki software)
  1. You don't have to give your real name: Forcing registration with a real name, in my opinion, is the single greatest contributing factor to the failure of Citizendium.  Because you don't have to give your real name on Wikipedia, it becomes an escape for many editors (and famously something of a game).  Other successful online communities from Second Life to Yahoo! Answers allow a similar pseudonymity.  The Wikipedian is, by this pseudonymity, allowed to create a new persona, and in editing, escapes the constraints of his daily life (Perhaps, the enormous number of college and high school students found on Wikipedia has something to do with these effects; after spending all day being taught, it must be fun to escape into a world where you are the knowledgeable and important one)
  2. You can rise to prominence through on-wiki actions alone:  On Citizendium, becoming an "expert" editor requires real world credentials, which have nothing to do with your editing of Citizendium as such.  On the other hand, on Wikipedia, anyone can become an administrator, bureaucrat, or member of the Arbitration Committee, simply by virtue of his on-wiki record, whether he be a tenured professor of theology with multiple graduate degrees or a 24-year old college dropout.  Keeping everything internal to the Wikipedia system gives contributors more incentive to involve themselves heavily.
  3. You can write about anything you want (more or less): The incredibly permissive "notability" guidelines on Wikipedia mean that you can show up to write about just about anything (or anyone) you want.  If you're particularly interested in something (or someone) incredibly obscure, you can write an article, get it featured, and have it on the main page, spreading this knowledge to others.
  4. There is an "enemy": Because "anyone can edit", Wikipedia is constantly beset by vandals and trolls, creating an enemy for Wikipedians to fight off (certainly this adds to the game-like nature of the site).  The sense of having a common foe, as social scientists have long documented, creates a sense of community and purpose, which keeps people editing.
  5. The system is easy to manipulate: Because you don't have to give your real name and editorial oversight is sorely lacking, a skilled editor can manipulate Wikipedia easily.  Doing so, however, generally requires an established reputation on Wikipedia.  Thus, for example, when Inc. Magazine gives tips for Wikipedia marketing, it includes advice to "spend a little time adding information to those subjects where you have expertise".  A surprising amount of good content on Wikipedia, then, comes from people who write it only in the pursuit of their own, other goals on the site.
What I imagine you have noticed about this list by now is that the five items on it are also five of Wikipedia's greatest weaknesses (in terms of developing a reliable reference work).  A reliable reference work generally requires a system with named authors, expert editors, and so forth.  The fact that you can write about nearly anything creates many headaches in the BLP area (As Alexander Lih has noted, "inaccuracy or vandalism problems are difficult to stop for people who are 'notable but not extremely famous'".  Wikipedia's liberal inclusion policy open the floodgates to biographies of literally hundreds of thousands of such people).  A perception of an enemy is also not a particularly helpful thing for your average reference work, and it goes without saying that it shouldn't be easy to manipulate a reliable work.

The paradox of Wikipedia, then, is that the same strengths that drive it success as a website and online community are the very weaknesses that hold it back as an encyclopedia.  When Sanger, in Citizendium, tried to remove some of the weaknesses of Wikipedia qua encyclopedia, he failed to see the importance of Wikipedia qua community and website.  The result was failure.   The genius of Jimmy Wales (an expression that may cause some of my wiki-skeptic readers to grate their teeth) is the realization, conscious or not, that Wikipedia is significant as a community, not just an encyclopedia, and the decision to nurture that community (sometimes at the expense of the encyclopedia).

What does this mean for the future of Wikipedia?  Well, to go forward, I for one believe that Wikipedia must blend the ingredients of its success qua community (although of course, the research of Ed Chi shows that this community may be in decline) with the elements of conventional encyclopedic success in a way that does not threaten the success Wikipedia has experienced thus far.  The obvious way to accomplish this is to return to the original vision of Wikipedia in the days of Nupedia as a feeder project. Continue to let a thousand flowers bloom and a certain anarchy prevail on Wikipedia, but then, take its key articles, run them through a review by experts (in the tradition of Citizendium) and publish these polished articles somewhere in a stable form.  Let the folks on Wikipedia keep editing away in the background if they like, and when a superior product evolves, polish it and publish it as well.  

Well, as they say, that 's all the time we have for tonight.  I think that's quite a lot to chew on for you, and I welcome your thoughts and comments.


Anonymous said...

You aren't saying Ireland is in the UK are you? Also, if you are writing for a general audience you may want to spell out what BLP means.

Anonymous said...

Obvious conclusion. A step beyond Wikipedia will be a hybrid. Imagine Wikipedia with a class of editor called Expert. The expert is admitted as such by other Experts, subject to appeal, and it would be topic-specific. Experts would be required to identify by real name. Articles may be created by anyone, and edited by anyone, but review by Experts would be flagged as such. (i.e., flagged revisions using experts as a high-level validation of a revision). It actually would take no new software, it could be done with protected lists and bots. Bingo. Wikipedia plus Citizendium. Yes, there would be conflicts to resolve. And real dispute resolution, as distinct from "ban the point-of-view pusher" is mostly missing from Wikipedia, but it's known how to do it.

RDH(Ghost In The Machine) said...

Excellent and insightful entry!
But I think Citizendium was doomed from the start by some fundamental flaws.

It essentially beefed up and combined the constipated vetting system for articles that failed on Nupedia, with the community policing that has largely failed on Wikipedia.

Suck out all the fun and replace Jimbo with Larry (who in spite of his promise to step down after two years, remains Cit's god-king, for what its worth) and you've got Citizendium.

Mr. Sanger is a man of impressive intellect and character, but he does not know when to let go and cut his losses or suffer fools gladly. Both of which Jimbo does very well.

I agree with you fully, though; the best way to build a serious, reliable, quality online encyclopedia is to adapt some sort of assembly line approach, with the WP approach at the start, Veropedia's in the middle and Cit's at the end.

Anonymous said...
This post has been removed by a blog administrator.
Daniel Mietchen said...

Thanks for the feedback. I am one of those who are active at Citizendium and prepared a comment in my text editor but your form did not allow to paste it in (for use of the UL tag), so I put it up on my blog.

Limey said...

Note: I removed a post containing an attack against a particular member of the Citizendium community after it was brought to my attention. Also, please see my new post on this topic: More On Citizendium.

Post a Comment