Last Thursday W3C's SPARQL 1.1 Federated Query extension went into Last Call status -- meaning that the working group thinks it has done everything necessary to complete the specification, and wants others to review.
This may seem a little obscure to many developers, since SPARQL queries RDF and RDF itself is far from ubiquitous. But Sir Tim Berners-Lee himself had already affirmed SPARQL's importance five years ago -- and as semantics become more and more important in the development of HTML5, the gulf between HTML and the Deep Web shrinks correspondingly.
The Federated Query extension is important because it lets SPARQL blow past unnecessary barriers to universal semantic interlinking. The problem is, basically: SPARQL can search RDF data, but what RDF data is available to a single SPARQL query? If only one database, then SPARQL's effectiveness is limited by the accidents of web architecture. But a simple metasearch won't do the trick, because then the whole semantic advantage of RDF is lost.
Federated search is the general solution -- like Google Scholar, or WorldWideScience.org (awesome sites themselves, by the way). But combine federated search with RDF and the promise of legitimately universal search begins to peek over the horizon.
If you're interested, you might want to work your way back from the SPARQL 1.1 Federated Query extension to SPARQL 1.1 itself, and from there to W3C's general RDF portal, their RDF primer, or their full list of RDF standards.