Yet another article (in Wired this time) about tags and folksonomies and yet another article that fails to mention any potential problems associated with a taxonomy uncontrolled by people (a folksonomy).
Tags work very well on a personal basis, you tag the data that's of interest to you, da_way-youWant. But what happens when the sites need to harvest the potential of those tags to empower the users?

The article states: We know that Flickr hosts 23,081 images tagged with "cat" or "cats" and only 17,463 with "dog" or "dogs". and the list of popular tags on del.icio.us reveals, for example, "blog, blogging, blogs" and "photo, photography, photos".

Take an end-user (that might also be using the service to save his bookmarks, but not necessarily) looking to use del.icio.us or flickr to find good links/pics on a given subject (their being-saved-here implying a certain degree usefulness --or prettiness in the case of Flickr-- as opposed to Technorati); he is presented with duplicate keywords... Moreover, he cannot splice (is it the right word?) those keywords to find links tagged by a selection of keywords. He cannot refine his search. Should he go back to Google?

Tagging systems don't seem to be future-proof yet. Aren't sites like Google Suggest (widget that could be reused and enhanced to show, for similarly spelled tags, the recommended spell) and the rudimentary del.icio.us tag stemmer proofs that we already have a big mess and are in the process of creating another one in parallel?
I sure do hope that there is more to tags than pretty weighted lists and a rehash of the meta keywords tag.