Slashdot indirectly pointed me to a very useful paper that was presented last May at the 15th WWW Conference. In Exploring Social Annotations for the Semantic Web, Xian Wu, Lei Zhang, and Yong Yu describe an approach to improving the disambiguation of tags within social bookmarking service. The "co-occurrences" of tags logged by users with similar interests for a given url are analyzed to form knowledge categories and to derive the semantics of the free-form assertions made by users.

Representation of Semantics:

• We represent semantics of an entity (a web resource, a tag or a user) as a multi-dimensional vector where each dimension represents a special category of knowledge.
• Every entity can be mapped to a multi-dimensional vector, whose component measures the relativity between the entity and the category of knowledge. If one entity relates to a special category of knowledge, the corresponding dimension of its vector has a high score.
• The total knowledge of users, tags and resources are the same, we can represent them in the same multi-dimensional space, which we call conceptual space.

Source: Slides from the WWW2006 Presentation

As more objects in web apps become taggable by users, this approach can be used to determine which object properties users are interested in annotating, searching, and viewing. The UIs for lists and views could even be customized per user to emphasize the aggregated data of other users with similar interests.