O'Reilly Network: The Semantic Web: It's Whom You Know [Apr. 19, 2002]
Tim Berners-Lee defines the Semantic Web as "A new form of Web content that is meaningful to computers" and believes it will "unleash a revolution of new possibilities". This is a very broad definition and includes any kind of machine-machine communication, any kind of "Webservice".
Andy Oram's article at O'Reilly Network focuses on using semantic web technology for knowledge management purposes, i.e. for "intelligent" retrieval of information from the overwhelming amount of content that is available on the Web.
The semantic web people are trying to formalize semantic and by that means make it machine-processable. Is this the right approach? Andy Oram points out that formalizing semantics basically means reducing semantics to syntax. And the more you delve into formalizing a semantic system, the more complex it gets. By the point you have formalized enough to make the system somewhat useful, it's so complex that it's hardly possible to handle it any more.
While Tim Berners-Lee's semantic web tries to offload the filtering work to machines, Andy suggests that the filtering can only be done by people, and we should then leverage technology to access this filtering work that has already been done. Google is doing this by evaluating the links that have been set by people between web sites, for its page ranking system. Imagine e.g. you could configure Google to personalize the page rank computation by taking your personal amount of trust in differerent web sites or even authors into account. Amazon's "customers that have purchased this book also purchased these ..." is another example of machine-reaping of semantic filtering work that has been done by humans. It was a human who was interested in subject A, bought book B and then decided that book C may be interesting as well. The machine did not understand anything of the content of the books and why the people bought them. Weblogs are another powerful source of filtering work that has been done by people.