Friday, January 19, 2007

Encyclopaedia Britannica Versus Wikipedia, Versus Hyperlinks:

Months ago there was a lot of web excetement about a comparison between Encyclopaedia Britannica and Wikipedia. An outside organization asked experts to read some (unidentified) articles from each and count mistakes. E.B. was stung by the results of the study, but they took their time, and they put together a strong rebuttal. They argued, in various ways, that many of the mistakes attributed to them were not mistakes. I read the rebuttal carefully, and one thing really struck me that I want to share with you.

In many cases, E.B. argued that an item was not a mistake because of some material in a different article. "X" was not a sin of omission because they covered it THERE, "Y" was not a wrong implication because they plainly discussed implications of Y in this other volume, etc. Now the obvious point to make is that Wikipedia has hyperlinks in profusion, so that it's not very important to decide WHERE a piece of information goes. There'll usually be a link to and from it in that other article. E.B. can't have hyperlinks in its text version, nor could it drown readers in printed references to other articles (in comparison, hyperlinks are much tidier than explicit references).

Now here's a related point: the challenge that E.B. accepts, to place each piece of info in the correct article, is a very difficult challenge. I imagine there are a few experts at E.B. who spend much time agonizing over correct categorization, and doing it very well. But we, the great unwashed who read E.B., lack their expertise in categorizing, so often we will not know where they would know to look. This is a general problem with categorizing, that it's self-defeating to create a sophisticated set of categories that must be used by many other people who will never have time to learn the categorizing rules.

"Tagging", a technique in wide use on the web, illustrates what's wrong with formal categories. You might think that the people who run the photo web site Flickr would set up a few hundred categories, and then let you tag each photo for the category it belongs to. They might even use an expansion of the Dewey Decimal System for this purpose. But people will not know such category systems well enough, and refuse to use them out of frustration. Poeple will also come up with new exciting ideas that do not fit the existing categories. Flickr allows us to create our own "tags" and use them to categorize our pictures. The results are best described as chaos, but they tend to suit our immediate needs rather well. I have some experience with controlled categories, and I'd call the result of imposing them chaos as well!

No comments: