Component models for semantic web languages

Abstract

Intelligent applications and agents on the Semantic Web typically need to be specified with, or interact with specifications written in, many different kinds of formal languages. Such languages include ontology languages, data and metadata query languages, as well as transformation languages. As learnt from years of experience in development of complex software systems, languages need to support some form of component-based development. Components enable higher software quality, better understanding and reusability of already developed artifacts. Any component approach contains an underlying component model, a description detailing what valid components are and how components can interact. With the multitude of languages developed for the Semantic Web, what are their underlying component models? Do we need to develop one for each language, or is a more general and reusable approach achievable? We present a language-driven component model specification approach. This means that a component model can be (automatically) generated from a given base language (actually, its specification, e.g. its grammar). As a consequence, we can provide components for different languages and simplify the development of software artifacts used on the Semantic Web. © 2009 Springer Berlin Heidelberg.