Most people think that tagging on the Web is pretty easy and fun. Give 'em a blog or a Web page and a field named "tags," and they'll start stuffing in text with wild abandon in the hopes that their content will be easily found by people who are desperately searching for information and opinion on feline hairball cures or cycling in the Ozarks or whatever their particular hobby is. Alas, all these folks are doing is polluting the Web.

Tags arose out of a need for a way to classify Web page content and blog entries that the big search engines, such as Google, couldn't find or ignored. Tagging also appealed to people because it was a democratic technique that was fast, easy, and had a perceived payoff. If that payoff ever existed it was back when the blogosphere was smaller and tagging hadn't gone mainstream. Today I doubt there is much of a payoff any more.

The trouble is rot has set in, and tagging has developed a few significant problems that are making it progressively less valuable. This is not to say tagging is, per se, a bad thing, merely that its popularity and the lack of standards have ensured that its utility value will continue to degrade. This degradation ensures that tagging will turn into a bigger source of content "noise" as every day passes.

The first problem with tagging is semantic vagueness. For example, does the tag "china" apply to the country or crockery? While you might hope that the distinction between the two would be evident from examination of related data, such as other tags used for the same item, specific words used in the item or in the rest of the site hosting the item, the effort required to resolve the context wipes out the value of tagging in the first place.

A second problem is the format of tags isn't standardized. This means that issues such as how white space is handled, which characters are legal, and which characters have special meanings and what those meanings are go undefined.

As an example, consider whether the tag "wind and rain" should be treated as three separate terms, two terms with a "stop" word or as a single phrase. In the latter case the question of whether "wind & rain" would be equivalent is also important and, as things stand, completely undefined. This means different applications parsing the same set of tags will index the same data in different ways and extract different conclusions.