Few days ago a quick thread about integrating semantic web technologies in GNOME took place in the gnome-devel mailing list.
The main question was if it's planned for GNOME a project like Nepomuk which API is going to be implemented for the KDE desktop-environment.
Apart from the decision if incorporate Nepomuk in GNOME (no decision has been taken), I've noticed that there seems to be a misunderstanding about what is and what can be done with a RDF-Store.
A RDF-Store is not a database; it certainly permits to store RDF triples but it doesn't have to be intended as a way to store informations in a schematic way as we usually do with databases.
Firstly RDF is a model to organize semistructured informations, and, most important, it's possible to create reasoning (via OWL) on top of it, that's a thing that can be hardly done with a classic database. This means that from a pool of data we can get more informations than what we've stored on it.
It's also much easier to integrate different sources of RDF triples together than joining data from different databases.
So, I think, that storing desktop-oriented informations using Semantic Web technologies is a more proper and far-sighted way to organize users data.