Downloads & Free Reading Options - Results

Using the Web Ontology Language (OWL)[1] to model information can have great benefits. It is semantically richer than many commonly used models, such as entity-relationship (ER) diagrams for relational database systems or class diagrams for information models. The W3C [2], which is responsible for OWL, does not endorse a standard graphical representation, but the Object Management Group s Ontology Definition Metamodel (ODM)[3] defines how to use the Unified Modeling Language[4] (UML) s wellknown class diagram notation to depict OWL ontologies. As OWL grows in popularity, it is reasonable to expect that many organizations will want to use OWL for their information models. This raises the question of the most efficient way an organization can produce its OWL-based model, especially if (as is common) it already has a significant investment in another model. A second and equally important question is what an organization does with its OWL model: An organization with a significant investment in using a relational database system for data persistence may be loath to abandon that model. In other words, it may want to use the ontology as a specification but continue using RDBMS technology as the implementation until it has the resources to update all its systems. This paper describes our experience converting a legacy ER-based information model to OWL, and using that OWL model as the basis for generating DBMS and other implementations. The methodology is broadly applicable. It relies on well-known standards and technologies, in particular OWL, UML, and the Model Driven Architecture.[5]