July 10, 2006
What a disappointing species we are. Stick us in a virgin paradise, and we create great honeycombed bureaucracies, vast bramble-fields of rules and regulations, ornate politburos filled with policymaking politicos, and, above all, tangled webs of power. Freed from history, freed from distance, freed even from our own miserable bodies, we just dig deeper holes in the mire. We fall short of our own expectations.
Witness Wikipedia. For some of us, the popular online encyclopedia has become more interesting as an experiment in emergent bureaucracy than in emergent content. Slashdot today points to Dirk Riehle's fascinating interview with three high-ranking Wikipedians, Angela Beesley, Elisabeth "Elian" Bauer, and Kizu Naoko. They describe Wikipedia's increasingly complex governance structure, from its proliferation of hierarchical roles to its "career paths" to its regulatory committees and processes to its arcane content templates. We learn that working the bureaucracy tends to become its own reward for the most dedicated Wikipedians: "Creating fewer articles as time goes on seems fairly common as people get caught up in the politics and discussion rather than the editing." And we learn that the rules governing the deletion of an entry now take up "37 pages plus 20 subcategories." For anyone who still thinks of Wikipedia as a decentralized populist collective, the interview will be particularly enlightening. Wikipedia is beginning to look something like a post-revolutionary Bolshevik Soviet, with an inscrutable central power structure wielding control over a legion of workers.
It will be interesting to watch how those workers respond as they confront the byzantine bureaucracy that's running the show. Will they continue to contribute, or will they become alienated and abandon the project? As Angela Beesley remarks, "The biggest challenge [for Wikipedia] is to maintain what made us who and what we are: the traditional wiki model of being openly editable." Kizu Naoko singles out "lack of involvement" as a major threat to the project: "we need to go back to the first and foremost challenge: To keep the openness of the wikis that makes it easy for people to join." The fate of Wikipedia - and perhaps the general "participative" or "open source" organizational model of online production - appears to hinge on how the tension between openness and bureaucracy plays out.
There was one passage in the interview that was of particular personal interest to me. Some time ago, I proposed the Law of the Wiki: "Output quality declines as the number of contributors increases." At the time, I was heavily criticized by leading members of the wiki community, including Wikipedia founder Jimmy Wales and wiki-preneur Ross Mayfield, who argued that the opposite was true - that the more contributors an entry attracts, the higher its quality becomes. So I was gratified to find my Law of the Wiki confirmed by the interviewees:
Dirk Riehle: What about the 'collective intelligence' or 'collective wisdom' argument: That given enough authors, the quality of an article will generally improve? Does this hold true for Wikipedia?
Elisabeth "Elian" Bauer: No, it does not. The best articles are typically written by a single or a few authors with expertise in the topic. In this respect, Wikipedia is not different from classical encyclopedias.
Kizu Naoko: Elian is right.
There you have it: Experts matter. And they matter more than the "community." Indeed, "a single or a few authors with expertise" will trump the alleged wisdom of the crowd. Now, there's something to build on.
This has been a proved model for our 5,000+ year history. The problem has always been the feeling of exclusion or not having a voice. Utopian models exist, just not in the real world.
Our founding fathers created a wiki, representative democracy, where everyone (supposedly) has an equal voice. But even that has been corrupted over time.
I've been trying to opt-out as a Wikipedia subject, so far, unsuccessfully:
Posted by: Seth Finkelstein at July 10, 2006 01:54 PM
Nick, you oversimplify. The substance of the best articles are indeed written by a few experts. They are also copy edited by hordes of amateurs and re-ordered and flow improved by a smaller set of editors. These people are often experts on some other topic. With the exception of highly controversial topics, the more editors that an article has, the quality is generally better due to better copy editing.
I agree with you that emergent bureaucracies exist, but I don't think they are seriously damaging the wiki spirit.
Posted by: Kingsley Joseph at July 10, 2006 02:31 PM
This is a textbook example of Pournelle's Iron Law of Bureaucracy.
This paradox needs more discussion. The wisdom of the individual needs to be more clearly identified from the wisdom of the crowd.
Posted by: Paul Cox at July 10, 2006 03:20 PM
"With the exception of highly controversial topics, the more editors that an article has, the quality is generally better due to better copy editing."
Kingsley, On what evidence do you make such a dubious claim?
Posted by: tim Swan at July 10, 2006 03:46 PM
Really, Kingsley, that's ludicrous. I've worked as a copy editor, and I can assure you of one thing: one decent copy editor will do a better job on any given piece of writing than will 10,000 mediocre copy editors.
Posted by: Nick Carr at July 10, 2006 03:59 PM
>>one decent copy editor will do a better job on any given piece of writing than will 10,000 mediocre copy editors
True, but more to the point, we are talking about substantive content, not jots and tittles. Wikipedia is laden with articles that are nothing more than assemblages of random factoids, with no narrative flow whatsoever. Where is the Wikipedian who is willing to invest the time and energy to make these mashups into coherent essays on the topic?
Posted by: marianc at July 10, 2006 04:29 PM
"I've worked as a copy editor, and I can assure you of one thing: one decent copy editor will do a better job on any given piece of writing than will 10,000 mediocre copy editors."
And I will readily cede authority to you on this. My unscientific survey of the last 3 featured articles on wikipedia show about 15-20 editors for that last 50 changes, in patterns that support my argument. Take a gander at the edit log of today's featured article, Microsoft. I speak from the experience of participating in both good and terrible article crafting in the Winkipedia - it's not much like a conventional editorial process. More editors does seem to correlate to better quality.
Marianc, give those articles time. They will either be deleted or spruced up.
Posted by: Kingsley Joseph at July 11, 2006 01:38 AM
But at least we still have illusions of freedom:
"Freed from history, freed from distance, freed even from our own miserable bodies."
Your post is full of history. Distance is still a factor. Freedom from your miserable body is called death.
But at least we're free!!!
Posted by: Clyde Smith at July 11, 2006 04:59 AM
POV disclaimer: I am a Wikipedia administrator.
The big challenge I see now is that it's just spectacularly easy to get completely lost in the 1.25 million articles that now exist.
Aside from the WikiProjects, there's no cohesive and officially-sanctioned effort that puts together "good" editors to adopt and watch over articles in a given area of knowledge and guide the gradual improvement process forward. Thus it's all too easy for articles to take a giant leap backward in one quick edit... and then get swallowed up into the morass. For high-profile articles, this doesn't work... but what about the 1.2 million articles that aren't high-profile?
I also feel that it's absolutely impossible to know for sure what's out there right now because there's so much of it - and because of that lack of systematic organization, there's no way to know what's fallen through the cracks.
This doesn't mean Wikipedia is doomed - in fact, for generally non-controversial stuff like airplanes, bands, etc. I find more people I know than ever using it as a quick reference. What this means is that there are a lot of growth-related management challenges that need to be addressed. Quite clearly, editorial control of a "freely editable" encyclopedia gets more and more difficult as more and more pages are added because it's so easy to create pages that just disappear.
I do want to note that I share your dismay at the endless reemergence of bureaucracy once we seem to have a good thing going. I was involved with numerous left and anarchist projects in the 80s and early 90s and it was always so harsh to experience the recreation of power structures to which we were supposedly in opposition.
It's kind of like watching new media replicate old media, for example, in the recent Rocketboom disaster.
Posted by: Clyde Smith at July 11, 2006 07:10 PM
At the risk of self-promotion, here's a link to a post partially spurred by your post... http://blog.lombardicto.com/2006/07/software_kills_.html
An interesting, if controversial, example is Al Queda. As Hoekstra points out, while we're "busy reorganizing to develop an agile response organization" (a corporate oxymoron), Al-Queda is focused on the virtual, skills-based side of things. Implicitly (or who knows, maybe explicitly) they use metadata to "route" tasks. They don't need to know who reports to whom... the operational ownership for this particular process are probably completely masked... but certainly there are attributes of other people that determine who Terrorist A routes a particular task to. Separately, I imagine, a cell leader manages the "HR side" of things.
Highly distributed task management, but very localized mentoring and HR management. This is the model of the highly adaptable organization.
Posted by: Phil Gilbert at July 13, 2006 01:20 PM
Your Law of the Wiki can be expanded to mailing lists as well. You only quoted part of your law from your 2005 post, but I think this part is an important section as well:
"Making matters worse, the best contributors will tend to become more and more alienated as they watch their work get mucked up by the knuckleheads, and they'll eventually stop contributing altogether, leading to a further fall in quality."
You can see this on professional mailing lists. A few people with a low threashold for posting quality start populating the list with content. People with higher standards decide that 1. reading those contributions is not worth their time; 2. making any contributions of their own is not worth their time. And thus the quality of the list goes down not only because of the bad material that gets posted, but also b/c less and less good material is added to the discussions.
Posted by: eszter at July 14, 2006 07:16 AM
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