What follows is a summary of the SemanticWeb architecture as described by Tim Berners-Lee's 2001 Scientific American article.
The function of the Semantic Web relies on knowledge representation: computer-accessible structured collections of information and sets of inference rules that can be used to conduct automated reasoning.
XML allows users to add arbitrary structure to their documents, which can be used by computers in sophisticated ways. However, the writer of the computer program must know what the creator of the page has used each structure for - XML alone does not capture information about what the structures mean.
RDF enables users to use MetaData to describe data on the Web. Machines can infer relationships between resources because their metadata declares that they describe the same concept. The triples of RDF therefore form webs of information about related things.
See Also: DublinCore
In order to interconnect the Semantic Web, computers need to be able to discover common meanings between different metadata definitions.
The third important technology for the Semantic Web are therefore collections of information which describe common meanings and relationships between resources on the Web, called Ontologies.
Typically, ontologies for the Semantic Web have a taxonomy and a set of inference rules. The taxonomy defines classes of objects and relations among them, and inference rules allow computer programs to deduce new facts about the objects. Ontologies can be used to relate the information on a page to associated knowledge structures and inference rules.
The RdfSchema is one such language which provides a means for formalising ontologies. RdfSchema extends the ResourceDescriptionFramework model by enabling a collection of resources to be described according to a simple class hierarchy.
The real power of the Semantic Web will be realised when autonomous computer programs called agents are created that collect and reason over Web content from diverse sources, exchanging data and working cooperatively with other agents.
The exchange of proofs is an important aspect of agents functioning, and agents will be able to convert their internal reasoning into a common representation language (such as RdfSchema) so that other agents can check their reasoning.
Digital signatures are another important feature of the Semantic Web. Digital signatures are encrypted blocks of data that computers and agents can use to verify that the attached information has been provided by a trusted source.
Tim Berners-Lee wrote the Semantic Web Roadmap  in 1998, and it continues to be a good place to start for those who enjoy theory.
-- Last edited October 27, 2002
Sponsored by Eastgate