Discontinuity and knowledge management

I think many of the problems we people have, both in traditional management of our information and (often implicitly) in work on the Semantic Web, is Dawkins' notion of the discontinuous mind¹. His scenario is of a lawyer who presented him with a scenario: if we are descended from apes, and species cannot interbreed, at some point a parent would have been unable to breed with its child. Fallacy in evolution? No — the lawyer failed to realise that the attribution of a species is performed by an analytical observer. Evolution is continuous; we can each breed with our children, but over 10,000 generations we cannot.

How does this apply to the subject? Well, consider filing an email. It's from my supervisor, touching on work for a conference and going for a drink. Does that go in a folder for him, a folder for the conference, or some social folder? The answer, of course, is all of them² and none of them: the idea of classifying a piece of text into arbitrary categories, based on some characteristics of the content or meta-data, for later retrieval is rather strange. I would suggest that filing email is imposing discontinuity, and therefore quite naturally doesn't always work.

The same kind of thing goes for people who rely on type information on the Semantic Web. It's not what you think it is.

¹ Dawkins discusses this a little in this article, but the following example was discussed in A Devil's Chaplain (ISBN: 0618335404).

² See User Interfaces for Supporting Multiple Categorization, Quan et al, 2003; my bibliographic entry.

Posted at 2004-09-27 00:30:22 by Richard • Link to Discontinuity and …
Comments, trackbacks.