The notion of the semantic web and the schema efforts to enable it are worthwhile, and by their openness on the right track, but since language and hence knowledge is a game of incompleteness and ambiguity, the schema efforts are likely to fail due to their grand scheme nature. People will not comply. Reading some comments about XML support in the most common client data tools on the planet (apart from the browser) namely MS Office it is comforting to know that they are at least getting some of it right, working up from data instead of down from metadata. This is the only thing that could possibly work. And this is the reason why Google is such a huge success. The ambient contradiction is, that the story is about the XML-enabling of Office (i.e. a huge push down from metadata, not up from data).
From a client perspective, however, there is no question that the direction I indicate is the important one. The really interesting thing is that one will expect to be able to import old non-standard data to XML (proprietary of course - they are still not the good guys, just less shady).

Next step up: RDF actively deployed and used by e.g. Google. The first application of this is already out there of course - being the many interlinked weblogs about web services and their many cross supscriptions and structured cross linkage.