Harry Chen blogged about the article Flying machines, desktop software and Web 2.0 by Jay Larock, who compares web and desktop software.

Larock gives three reasons why we will continue using desktop computers and desktop software (abbreviated by me):

1. it takes time to port software and make the shift
2. we are not always online, but the desktop still runs
3. some software works better when eating your local cpu power

I copy all of these arguments, and add a fourth: some people don't trust free services on the web, who may censor your work, suddenly go out of business, or be hacked, and therefore some people keep a copy of their data on their own harddisks and enjoy desktop apps.

Harry then comes to this conclusion:
We shouldn’t ask the question whether desktop software will survive in the age of Web 2.0 (yes, they will survive). But instead, we should ask: how can Web 2.0 applications (and Semantic Web applications) complement the existing functions of desktop software, so that the users can be made more productive?

This question was asked differently by me in 2003 and answered in this thesis:
http://www.dfki.uni-kl.de/~sauermann/papers/sauermann2003.pdf

"If the goal is to have a global Semantic Web,
one building block is a Semantic Desktop,
a Web for a single user. "

After this, Stefan Decker and Martin Frank published their "Networked Semantic Desktop" paper, and you find several implementations that bring Semantic Web technology to the desktop, www.dbin.org, www.openiris.org, gnowsis.opendfki.de.

And lots more that are under the "radar". So, the questions is good, but using the keyword "Semantic Desktop" you easily find an answer. There are many articles about it. If you have more questions, ask the people@semanticdesktop.org.

So yes, its indeed a good idea to combine web 2.0, semantic web and desktop computing :-)