Published in Proceedings of InterSymp-2002: The 14th International Conference on Systems Research, Informatics and Cybernetics: Baden-Baden, Germany, July 30, 2002. 11 pages. Publisher website: http://www.iias.edu.
As the volume of data and human-centered information available to decision-makers continues to increase at an ever-accelerating rate, the need to represent information in software-processable formats becomes more apparent. At the same time, the availability of information from diverse sources through the World Wide Web provides the opportunity to widen the scope of input to decision-support systems, if this information can be made accessible through automated means. Past approaches to information-centric interoperability have been based on the use of a shared static object model, but this becomes impractical when we consider the loosely-coupled decentralized nature of the Web.
This paper discusses the motivations driving a change from static to dynamic information models. It defines a representative use case, and describes a service-based architecture that allows for extending existing information sources to allow programmatic access. The proposed architecture uses existing and emerging Web Service specifications, enhanced by an ontology definition language, to create an environment that does not require information service providers to use static shared models, while allowing information consumers to learn ontologies from the services themselves. Clients such as decision-support systems can thus build their own information context at run-time based on models received from multiple sources.