Published: 2000 April
Herausgeber: U. Frank
Buchtitel: Proceedings of the German Workshop "Modellierung" 2000, Koblenz, Germany, April, 5-7, 2000
In this paper we present a new approach for language-neutral modeling of large-scale ontologies. The gist of our approach lies in the way we treat the majority of axioms. Instead of capturing axiom semantics in some specific representation language, we categorize axioms into different types and specify them as complex objects that refer to concepts and relations. A separate layer that is language-specific, in fact it may even vary for different inference engines working on the same language, describes how these objects are translated into a goal representation. In addition to its far reaching independence with regard to specific representation languages, this approach benefits engineering since the semantics of important types of axioms may be much more elucidated in our ontology engineering tool, OntoEdit, than in comparable tools. Furthermore, our approach is principled in a way that allows for comparably easy adaptation of our tool to requirements for modeling axioms in specific domains.