In “Metadata for the masses,” Peter Merholz presents an interesting idea, which is build a classification scheme from free-form data that users apply. He points to Flikr’s “Cameraphone” category, which would probably not exist if there was only a pull-down list.
He also points up problems: Many categories for one thing (nyc, NewYork, NewYorkCity), one category that means many things (“Flow, for instance, can either mean optimal creative experience, or the movement of a fluid,”), and categorizations that are wrong.
I think there’s a tie here to memes, or ideas which encourage you to adapt them. If I see a tag which strikes me, is evocative to me, or I see as useful, I’m likely to use it myself. If I create a tag which I find evocative, but no one else does, (say, “Bastiat-ic”) its unlikely to get picked up. I am a big fan of evolutionary, or memetic systems like this, and am sorely tempted to try to include it in my project, but the goal of that project isn’t actually to create a taxonomy, its to create a useful naming scheme. I think a taxonomy is part of that, but others who get a say in the final analysis disagree, and so I’d like to focus on getting a taxonomic name space, rather than a cool evolutionary method for creating it.
(Via Nudecybot. Oh, and its too bad that there’s no RSS on Merholz’s page. I’d like to see their essays, but not their “appearance dates and other news.”)