The following commment by Rod Petrovic appeared in a discussion on Digg.com regarding Google CEO Eric Schmidt's comment on the Web 3.0:
Web 3.0 is the Semantic Web and I believe it was correctly described by Schmidt: a "cloud" of data, which can be used by a large number of small, ubiquitous applications.
I have nothing against describing clearly identifiable trends on the web by version numbers, pet names or any other ID. I totally understand it: Web 1.0 was a web with few content providers and many users. Web 2.0 is a web where users are content providers, too. This was made possible by an evolving technology (AJAX, RSS, web services and APIs) and, consequently, application providers who use that technology (blog engines, social networks, etc.).
On the current web, application providers use RSS and APIs to occasionally "open doors" to their data, while they still keep it inside . On the Web 3.0 - the Semantic Web, data will be outside - distributed (see RDF and OWL W3C standards, for example).
I know this concept might be hard to grasp at first, so let me give you an example: If I have a friend in my LinkedIn social network and I open a Facebook account, as well, I'll have to add that same friend again, since LinkedIn's data is owned by LinkedIn and Facebook's by Facebook. On the Semantic Web, the data about me, my friends and our relationships will be distributed - open for all applications. It will be somewhere on the web in a form of RDF/XML and all applications will be able to access it. Application providers, like LinkedIn or Facebook, will then offer me innovative ways to use this omnipresent data. I can use one, two or all applications, on different platforms and for different types of services - it doesn't matter. What matters is that I have access to the same data from any application and I'm not being blackmailed into using any of them just because "all my friends are there" or "all my data is there".
Web 3.0 will be a mother of all mashups and a father of all social networks.