This is a brief article (Library Journal – Googlizers vs. Resistors) with discusson of ethnoclassification (bags-of-keywords) vs. controlled vocabularies to describe resources.
Stephen Downes pointed out the article, with the insightful comment:
In a world where tomorrow is different from yesterday, controlled vocabularies (are) not only inefficient, they’re dangerous. In a controlled vocabulary, not only can you not classify ‘podcasting’, you cannot even register its impact. Some librarians are getting this message. Others are not. Oh, and judging the effectiveness of a non-taxonomy based search by what Google can (only) do today? Mistake.
I’m seriously thinking of biting the bullet and finishing my MSc, with a thesis on ethnoclassification vs. controlled vocabularies…