Detecting Dirty Queries during Iterative Development of OWL Based Applications

Incremental/iterative development is often considered to be the best approach to develop large scale information management applications. In an application using an ontology as a central component at design and/or runtime (ontology based system) that is built using this approach, the ontology itself might be constantly modified to satisfy new and changing requirements. Since many other artifacts, e.g., queries, inter-component message formats, code, in the application are dependent on the ontology definition, changes to it necessitate changes to other artifacts and thus might prove to be very expensive. To alleviate this, we address the specific problem of detecting the SPARQL queries that need to be modified due to changes to an OWL ontology (T-Box). Our approach is based on a novel evaluation function for SPARQL queries, which maps a query to the extensions of T-Box elements. This evaluation is used to match the query with the semantics of the changes made to the ontology to determine if the query is dirty - i.e., needs to be modified. We present an implementation of the technique, integrated with a popular ontology development environment and provide an evaluation of our technique on a real-life as well as benchmark applications.