I'm currently in Stuttgart, Germany doing some Drupal client work in a gathering that we've come to call "Geek Week". We sit down and look at internal requirements and do 3/6/12 month planning, matched up with the state of the Drupal universe. But more on that later, probably over at RCS (the "In Drupal We Trust" t-shirts were popular).
One of the "geeks" attending here is Arto Bendiken. Check out his projects page for an example of some of the stuff he's worked on. For Drupal folks, that would be timeline - AJAX widget for visualizing temporal information, boost - static page caching for Drupal, drush - command line shell for Drupal, trace - easy debugging for Drupal, exhibit - rich visualization and faceted browsing. Yes, that is impressive :P
So, Arto and I got to talking about RDF (rdfabout.com is a good primer site), and how it's the new black. I admit that I've felt that XML vs. RDF is (almost?) a religious war. It seems to me that pointy haired bosses (PHBs) have memorized that RDF == slow and complex, and XML == fast and ubiquitous. Since in selling concepts I often interact with PHBs, RDF has felt like an uphill battle, especially as RSS/Atom grow more and more widespread.
I think I'm almost convinced that the scales are tipping in favor of RDF for enabling new, distributed actions. Looking at new companies like Nova Spivack's Twine or MetaWeb's Freebase, we are seeing new capabilities emerge based on using RDF.
Actually, SPARQL is another new piece, in fact just ratified as W3C recommendation as of yesterday. In short, a way to do queries across data at multiple sources, rather than having to rely on the aggregate-and-query-locally.
Here's an image of the data that is currently available as RDF. Please ignore the FOAF bubble in the middle, we really do need to replace that with something better :P