Berners-Lee and the Semantic Web Vision

He kicked off by explaining what he means by the two words "semantic" and "web." Underlying the Web was the philosophy of a navigable space, with a mapping from URI to resources. He stressed that a URI was an identifier for a resource, and not a recipe for its retrieval.

Berners-Lee said that in the context of the Semantic Web, the word "semantic" meant "machine processable." He explicitly ruled out the sense of natural language semantics. For data, the semantics convey what a machine can do with that data. In the future, Berners-Lee anticipated that they will also enable a machine to figure out how to convert that data, too. He described the "semantic test," which is passed if, when you give data to a machine, it will do the right thing with it. He also underlined that the Semantic Web is, like XML, a declarative environment, where you say what you mean by some data, and not what you want done with it.

Tim Berners-Lee addresses XML 2000 delegates.

Encouraging the audience to share his excitement at the vision, Berners-Lee related how difficult it was ten years ago when he was demonstrating the Web for the first time. Viewers seeing him progress from one document to another by clicking on links were nonplussed -- it's when the system scales that the advantages may be reaped. Shown a small example of linking zip codes between databases, we were asked to imagine the possibilities of this system scaling globally.

Having outlined its scope, Berners-Lee explained each of the elements in the Semantic Web architecture. He explained the importance of RDF/RDF Schema as a language for the description of "things" (resources) and their types. Above this, he described the ontology layer. An ontology is capable of describing relationships between types of things, such as "this is a transitive property", but does not convey any information about how to use those relationships computationally.

Running through several of the layers in Berners-Lee's abstraction is the importance of digital signatures. Although public key cryptography has been around for some time, it has not really taken off. Berners-Lee observed that one contributing factor to this was that it was too coarsely-grained, with the choice of either trusted or not trusted. An infrastructure must be in place where a party can be trusted within a particular domain. Once such granularity is possible, digital signatures can be used to establish the provenance not only of data but of ontologies and of deductions.

On top of the ontology layer sits the logic layer. This is the point at which assertions from around the Web can be used to derive new knowledge. The problem here is that deduction systems are not terribly interoperable. Rather than design one overarching reasoning system, Berners-Lee instead suggests a universal language for representing proofs. Systems can then digitally sign and export these proofs for other systems to use and possibly incorporate into the Semantic Web.

Berners-Lee ended his presentation examining what could be done practically today. He observed that the higher layers of his architecture are likely to take around ten years yet to come to fruition -- most of the new work today is happening on ontologies. Practical solutions include the use of XSLT to derive RDF from XML sources, the work on topic maps and RDF convergence, the emergence of general-purpose RDF databases and engines, and general and specific GUIs for RDF data. Berners-Lee noted that a rearrangement of the metadata activity within the W3C would also have a bearing on Semantic Web work.

Though many attendees enjoyed the talk and his vision, there is still a significant measure of skepticism over the practicality of the Semantic Web. More than one delegate invoked the failed, lofty ambitions of Artificial Intelligence in the '60s and '70s. Others noted that his vision has little in common with the current future of "web services" being promoted by most XML vendors. This was, however, one of the most complete expositions of his vision that Berners-Lee has yet delivered to the XML community.

• Slides from Berners-Lee's talk