The last time I looked, Paolo Valdemarin and Matt Mower - both of whom I not only respect but really like - were expecting users to categorize what they're writing according to some accepted taxonomy; I'm pessimistic about getting users to take that step. But if it works, it'd be fabulous and I totally support it.
I must say first that I really like David too. I'm looking forward to meeing him again somewhere in the world hopefully soon.
Then I'd like to say that when we first presented him the whole idea of ENT and K-collector in Vienna at BlogTalk 1.0, we did notice the skeptic look in his eyes and we have been working to solve this problem.
While at the very beginning we were expecting all users to use our tool to add topics to what they were writing (even back then we were trying to help users by suggesting appropriate topics for their posts), the system has evolved since then in a few ways.
What we are doing is separating the process of topics creation from the actual association of topics to content. While the possibility of adding topics to posts still exist and it's the best way of doing it, there are also other ways of associating content and metadata,
The first one is TopicMatching, which allows us to add topics to posts streaming through our servers automatically. Visiting our open w4 site, you will be able to see content from about 100 weblog grouped using topics created by about 15 "active users" (users with the k-collector client). In other words we are leveraging on the work of a few people to categorize a vast amount of content.
We are also investigating other ways of creating topics on our site. For example by analyzing words used in search engines that brought users on our sites. There are also other options we are testing which will allow to improve the results.
It's true, we have not resolved the whole problem, but we are definetly listening :-).
[Paolo Valdemarin: Paolo's Weblog]