An ontology for software models and its practical implications for semantic web reasoning

Abstract

Ontology-Driven Software Development (ODSD) advocates using ontologies for capturing knowledge about a software system at development time. So far, ODSD approaches have mainly focused on the unambiguous representation of domain models during the system analysis phase. However, the design and implementation phases can equally benefit from the logical foundations and reasoning facilities provided by the Ontology technological space. This applies in particular to Model-Driven Software Development (MDSD) which employs models as first class entities throughout the entire software development process. We are currently developing a tool suite called HybridMDSD that leverages Semantic Web technologies to integrate different domain-specific modeling languages based on their ontological foundations. To this end, we have defined a new upper ontology for software models that complements existing work in conceptual and business modeling. This paper describes the structure and axiomatization of our ontology and its underlying conceptualization. Further, we report on the experiences gained with validating the integrity and consistency of software models using a Semantic Web reasoning architecture. We illustrate practical solutions to the implementation challenges arising from the open-world assumption in OWL and lack of nonmonotonic queries in SWRL. © 2008 Springer-Verlag Berlin Heidelberg.