Toward Multi-viewpoint Reasoning with OWL Ontologies

Despite of their advertisement as task independent representations, the reuse of ontologies in different contexts is difficult. An explanation for this is that when developing an ontology, a choice is made with respect to what aspects of the world are relevant. In this paper we deal with the problem of reusing ontologies in a context where only parts of the originally encoded aspects are relevant. We propose the notion of a viewpoint on an ontology in terms of a subset of the complete representation vocabulary that is relevant in a certain context. We present an approach of implementing different viewpoints in terms of an approximate subsumption operator that only cares about a subset of the vocabulary. We discuss the formal properties of subsumption with respect to a subset of the vocabulary and show how these properties can be used to efficiently compute different viewpoints on the basis of maximal sub-vocabularies that support subsumption between concept pairs.