When a large amount of data is stored in OWL files, it is not efficient to maintain and query those data. The OWL syntax is based on XML, which is a meta-markup language. Thus, it is suitable for data description and data exchange, rather than for data storage and data management. Furthermore, enabling multiple users to work with the same ontology in parallel and make modifications mandates the use of databases (e.g. object-relational) for storing ontologies. Therefore, this paper proposes to extract data from OWL files and store those data in object-relational databases. Object-relational databases are good for data storage and data management. But OWL is semantically richer than SQL3. As a result, some of the semantics captured in ontologies can be lost when ontologies are stored in object-relational databases. In addition, some of the structures of ontologies can be lost. E.g. a class in an ontology can be stored as it were an atomic undifferentiated piece of data in a single CLOB column in an object-relational database. When the ontology is stored in such a way, the query and scalability features of the object-relational database cannot be fully exploited to access the data.
展开▼
