Workshop 11 - Reference Ontologies vs. Applications Ontologies

Two schools of thought are gradually beginning to crystallize in the domain of information systems ontology. On the one hand is the school which focuses primarily on the representational adequacy of an underlying ontological theory, leaving for others the task of transforming this theory into working applications. On the other hand is the (much larger) school which focuses primarily on the construction of ontologies as working applications at the expense of representational adequacy, and which is associated with current developments under the heading of the Semantic Web. The goal of the workshop is to bring together representatives of these two schools of thought for a structured debate, which is designed as a contribution to the clarification of the foundations of information systems ontology in the future.

The terms "reference ontology"‚ and "applications ontology"‚ were introduced by Nicola Guarino, who draws the basic distinction as follows:

... what kinds of ontologies do we need? This is still an open issue. Some people believe that very general ontologies involving rich axiomatic characterizations are important, others think they are a waste of time, and prefer to concentrate on lightweight ontologies, focusing on the minimal terminological structure (often just a taxonomy) which fits the needs of a specific community.

Guarino uses "reference ontology"‚ (which he now calls "foundational ontology"‚) to refer to ontological theories whose focus is to clarify the intended meanings of terms used in specific domains. Smith has proposed a more radical definition within the framework of philosophical realism: a reference ontology is a theory of some independently existing domain of entities which seeks to maximize descriptive or representational adequacy to the maximal degree compatible with the constraints of formal rigor and computational usefulness.

One goal of this workshop would be to prepare the ground for more coherent debate by fixing on one correct usage of terms such as "reference ontology"‚ "foundational ontology", "application ontology"‚ "lightweight ontology"‚ and indeed of "ontology"‚ itself.

Additional goals will be:

1. To clarify the range of current approaches to ontology and to demonstrate their relative merits and demerits.
2. To set forth the basics of the reference ontology approach and of its underlying rationale. To survey the institutions involved in developing reference ontologies. To address the role of realism, and of intended meanings, and intended models, in information systems ontology.
3. To set forth the basics of the application ontology approach and of its relation to the use of the specific machinery of Description Logics.
4. To present reports on case studies of the use of ontology in applications with special reference to domains such as medical informatics and spatial reasoning.

The benefits of the reference ontology approach are that it can bring a rich framework for resolving terminological incompatibilities and classificatory underdetermination. The benefits of the applications ontology approach are that it can yield ontologies which run in real time.

are available from the Workshop Organizer Barry Smith and from http://ontology.buffalo.edu/ki2003.