Introduction

The Semantic Web [18, 128] aims to make the vast amount of information on the Web accessible to machines through the annotation of Web content using machine-understandable formats such as RDF, and enable comprehension and integration of this information through the use of ontologies [55], which may be specified using the Web Ontology Language OWL [103]. However, these annotations refer only to static knowledge, and ontologies are – generally speaking – static descriptions of background knowledge in a particular domain. Web services [4] are concerned with providing functionality over the Web, and are thus more than chunks of static information; an example of such functionality is the sale of books over the Web, e.g., Amazon.2 Mainstream Web service technologies such as SOAP3 and WSDL4 provide means for the structured XML-based annotation of, and interaction with, Web services. However, the description of the functionality of services using these technologies is limited to natural language text and a description of the structure of input and output messages. These limitations make it hard – especially for a machine – to understand the functionality of a service, let alone automatically discover, combine, and execute Web services. Consequently, the location, selection, combination, and usage of Web services requires considerable human effort [57, Section 4.5].

There is a conjecture that the combination of Semantic Web and Web service technologies, called Semantic Web services, has the potential to overcome these limitations [104]. To facilitate combining these technologies, several approaches to Semantic Web service description have arisen. They range from bottom-up approaches that extend existing technologies, such as WSDL-S [2] and SAWSDL [54], to top-down approaches that introduce new languages for the semantic description of Web services and subsequently “ground” such descriptions in existing technologies. The two most prominent top-down approaches are OWL-S (OWL-Services) [102, 8] and the Web Service Modeling Ontology (WSMO) [121, 57]. The former is tied in with the DL (Description Logic) sub-language (species) of the Web Ontology Language OWL [46, 77], and requires the use of OWL DL for the description of services.5 WSMO provides a language-independent conceptual model for the description of services; it does not require using specific language, but requires languages that implement WSMO to follow the structure of the conceptual model.