Tragedy Of Complexity

Some think WikiWiki is a self-stabilizing system, but it isn't. While it hasn't suffered from a direct analogy to TragedyOfTheCommons, it is falling apart because of the unrestricted and unorganized growth of information. New users are coming to WikiWiki now and can't find anything, and as a result they don't connect with the community and they leave. That is a shame.

Some comments from TragedyOfTheCommonsCantHappenHere:

No one knows the entire PageDatabase any more so people just rewrite the same crap over and over again, because everyone wants to write on wiki but no one wants to read wiki.

This illustrates one common problem, but I don't agree with the cause. People write the same stuff over and over again, because they can't find the other stuff. (related: See InformationOverkill)

The Tragedy of the Commons occurs when people using a shared resource don't treat it as well as they treat their own private stuff, so it ends up going to heck. No one has a short-term individual incentive to treat it nicely (since their one act of treating it nicely won't have a material impact on how well-maintained it is), so it rots. Enlightened self-interest is, I think, enlightened precisely to the extent that it manages to avoid this Tragedy by seeing beyond each short-term individual interest. -- PeterMerel

This isn't enough. The TragedyOfComplexity is happening here because people are treating WikiWiki exactly like they treat their own stuff. Is your book collection organized by author, or by the DeweyDecimalSystem? Mine isn't. Back when I had twenty or so LaserDiscs? I kept them organized, but now I have > 200 and to find anything I have to dig through the whole pile. My DVD collection is not quite as bad, but it's not as large yet.

CoryDoctorow? captured some of it eloquently in his paper on meta-data: http://www.well.com/~doctorow/metacrap.htm.

In the paper, he discussed the problems of people properly marking up and organizing their work, in terms of meta data. These problems included:

WikiWiki doesn't not suffer from all of these (some are not applicable, like metrics, and others depend on the user base remaining enlightened), but it's not enough.

Better tools could probably raise the bar considerably (like more complex searching, and LikePages based on more than the first or last word of the title). But ultimately, WikiWiki will either survive because more and more people work on stitching information together and keeping its scope manageable; or it will diffuse into nothingness.

-- DerekWoolverton

Better tools will have a significant impact on Wiki. Just look at the difference between searching on, say, Google and searching for a phrase or mutually inclusive pair of words here. Wiki probably needs a whole raft of new facilities just to keep up with the content being provided by users; otherwise, Wiki will sink under its own weight. This is inevitable.

So what folks are saying is, WikiDoesntScale. I'm relatively new here, so I have no memory of TheWayThingsUsedToBe?; I do remember what happened to UseNet when it was "discovered" first by the general public (which introduced new blood and new viewpoints; but certainly upset lots of apple carts and added noise to the signal), and then when it was discovered by spammers (which has ruined it in many ways as a medium).

But scalability is a problem for all systems. If you have 10 books in your collection, it doesn't matter how you sort them. IF you have 100 books, you probably ought to have a simple, one-level sorting system; author or subject. If 1000 books, you probably want to sort by both, and consider software to catalog them, as well as adding a cache (a shelf for recently read books or favorites). 10000 books and you should implement the LibraryOfCongress or DeweyDecimal? systems to catalog your books. 100000 or more and you need a librarian; 1000000 and you need a staff of librarians and a database administrator.

Obviously, hiring librarians and database admins to maintain a collection of 10 books isn't cost-effective.

In the case of Wiki (and the web in general), we've gone from a mere collection of web pages, statically-linked in a haphazard fashion, to statically-maintained indices, to dynamically-generated (but arguably dumb) indices. The web has gone further, to more intelligent search engines and so forth.


Post your solutions here I simply DateStamp and then AlphabetizeEverything. -- ChrisGarrod


CategoryWiki CategoryComplexity?


EditText of this page (last edited October 31, 2005) or FindPage with title or text search