Towards Ontological Context Mediation for Semantic Web Database Integration: Translating COIN Ontologies Into OWL

The COntext INterchange (COIN) approach to information integration [Goh et al, 1999] [Bressan et al, 2000] [Firat et al, 2002] uses ontological mappings and enables powerful context-sensitive query mediation for semantic integration of knowledge across multiple heterogeneous database sources. Its existing applications include financial analysis for the financial services industry, as well as airfare and car-rental aggregators for the travel industry [Firat, 2003]. COIN’s original development preceded the Semantic Web. In this paper, we address how to combine the best of COIN’s capabilities with those of standardized Semantic Web ontologies and rules, i.e., OWL [W3C, 2004] and RuleML [RuleML, 2003] [Grosof, 2004]. As a substantial first step, we give (in summary form) a translation of COIN’s ontology representation into OWL description logic, thereby enabling COIN to utilize OWL ontologies. We further identify, at a high level, how to use RuleML logic program rules, together with the ontologies, to perform the kinds of reasoning that the overall COIN approach does, including mapping ontologies and mediating queries. Challenges for future work include treatment of equational ontologies, constraint solving, and abduction.