Edinburgh Napier University

  Complex applications such as those to support the emerging field of biodiversity informatics, require database systems able to represent and manipulate their data. Object-oriented databases (OODBs), designed to support complex domains such as Computer Aided Engineering, appear to be a solution. A criticism of early OODBs was the lack of an ad hoc query language such as SQL, which led to the development of OQL [CBB+97], a generic ad hoc query language. However OQL does not take full advantage of the semantics captured in the OODB and thereby provides little more power than SQL. This paper presents a database model and query language developed while building a database application for a complex domain, that of botanical taxonomy. Our extensions centre on the provision of relationships as first class constructs in the data model. We describe a prototype database system (Prometheus) in which we have augmented an existing OODB (ODMG compliant) with relationships, and extended OQL with generic operators for manipulating the corresponding structures. By providing richer semantic concepts in the database, the power of the query language is enhanced, thereby providing more powerful ad hoc querying at the application level. The system which has been fully implemented and used to build a taxonomic database, would also benefit other complex domains where semantic-rich relationships are important for capturing the semantics of the domain.