(This is an update to a post originally published in November 2006 – but still relevant today)
The debate about the various merits of taxonomies vs. folksonomies will probably continue long after I’ve departed this mortal life; it’s the lifeblood of dyed-in-the-wool information professionals. Based on experience I’ve had in implementing enterprise search solutions, users presented with either a taxonomic organisation of content vs. doing a keyword or free-text search for what they are seeking, the vast majority of users will choose a free-text search. The reason being that users don’t want to spend valuable time trying to understand the taxonomy, and particularly where the new breed of search engine is able to return relevant results AND cater for the serendipitous nature of some search queries. Interestingly, Verity (since bought by Autonomy, which was itself bought by Hewlett Packard….are you keeping up?!) had developed a collaborative taxonomy facility for their K2 search engine, where common terms could be identified for taxonomy labels. It sounds to me that they had recognised the limitations of the inflexible top-down taxonomy approach and were heading towards the realms of folksonomies. A strategy since adopted by nearly all of the current batch of enterprise search engines.
David Weinberger once said:
Folksonomies are not only frequently more useful than top-down taxonomies; they better reflect the bottom-up, messy, ambiguous, inconsistent, social nature of meaning, despite Aristotle and the tradition his genius spawned.
I couldn’t have put it better myself!