We are on the brink of a new generation of World Wide Web (WWW) which, in his recent book ``Weaving the Web'', Tim Berners-Lee calls the Semantic Web. Unlike the existing WWW, where data content is primarily intended for human consumption, the Semantic Web will provide data whose content is also machine processable. This will enable a wide range of intelligent services such as information brokers, search agents, information filters etc., a process that Berners-Lee describes as ``Bringing the Web to its full potential''.
The development of ontologies will be central to this effort. Ontologies are metadata schemas, providing a controlled vocabulary of terms, each with an explicitly defined and machine processable semantics. By defining shared and common domain theories, ontologies help both people and machines to communicate more effectively. They will therefore have a crucial rôle in enabling content-based access, interoperability and communication across the Web, providing it with a qualitatively new level of service: the Semantic Web.
In order for ontologies to fulfil their rôle in the semantic integration of the Web, there will need to be some standardisation of Web ontology languages. The W3C is already moving in this direction with languages such as RDF and RDFS. However, in order to achieve the widest possible acceptability, these languages have deliberately been kept very simple and have relatively weak semantics. Much richer ontology specification languages will be needed in order support the design, sharing and integration of the complex ontologies that will be required in order to exploit the full potential of the Semantic Web.
This requirement has already led to the development of languages such as OIL and DAML+OIL. It is anticipated that a W3C standardisation working group will soon be established in order to develop a W3C ontology language recommendation based on DAML+OIL. However, although OIL/DAML+OIL may be sufficient for many applications, language extensions will certainly be required by some applications: the requirement for an extension of the ontology layer to include some form of rules language has already been identified, and even this may not be adequate for all applications and/or application domains.