Semantic Web

The Semantic Web is a vision of Tim Berners-Lee, the original architect of the World Wide Web, which enables computers and users to cooperate in processing data on a global basis. The premise of the Semantic Web is that the current World Wide Web does a great job of indexing documents, but that there is very little opportunity for systems to help us digest this information.

With the Semantic Web in place, a user could ask a question and the Web would cooperate to find a meaningful answer. The search would not be limited to finding keywords or to known schema, but the query would understand the concepts behind the question and could compare them at a conceptual level to concepts encoded on Web sites.

For this to work, several technologies must be developed and deployed. Many of these are under development in academic and research groups.

Resource description framework (RDF) is the cornerstone technology. RDF is a way to express models of data or knowledge. It is a simple "triplet" of information, consisting of a subject, a property or predicate, and an object. What makes it unique is that any of these three can be a universal resource identifier (URI), which grounds the model in a definitive definition of the thing modeled.

RDFS adds a few modeling primitives to RDF to allow object-oriented modeling. OWL, which is the official successor to DAML + OIL, adds the expressive power needed to model ontologies in the RDF syntax. (See Chapter 14 for definitions of these acronyms.) With OWL we now have a way to express ontologies in a standard format. Many products and tools are being marketed to help with the process of building ontologies in OWL (and other) formats, such as Prot g from Stanford.

It is understood that there won't be one overarching standard ontology on which everyone will agree; instead, there will be many ontologies, each for a specific domain or perspective. Each site then will determine which ontologies the site will use to express its semantics, and it will "commit" to those ontologies.

Perhaps the largest hurdle is marking up Web sites and databases in these new languages and expressing them in terms of the ontologies that they have committed to. This is a major area of research and commercial activity, but at present there is no clear winner in terms of approach or tools.

Finally, query tools will need to evolve to capture the intent of a query and traverse the various sites that have committed to the same ontology or others that are reachable transitively.