DARWIN

In recent years, ontology-driven applications gained more and more momentum by smartening functionalities ranging from inferencing to querying and data integration. Nowadays, ontology-driven applications are finding their way also into traditional application domains. Development environments for constructing the core, i.e., the semantic infrastructure, of those ontology-driven applications suffer from three crucial deficiencies. First, inadequate abstraction mechanisms force software engineers to deal with intricacies of underlying semantic technologies. Second, change-agnostic editors neglect the importance and meaning of changes by dealing with semantic infrastructures as editable semantic artifacts only. Third, limited co-evolution support of highly interdependent semantic artifacts hamper maintenance of infrastructures evolving together with the applications. This leads to increased development times, decreased software quality and consequently, impedes efficient adoption of semantic technologies for smartening applications. DARWIN proposes a novel extensible framework to develop and maintain semantic infrastructures based on model engineering techniques. The innovations are manifested in three key research goals. First, DARWIN aims at tailoring existing conceptual modeling languages allowing to model semantic infrastructures on a technology-independent abstraction layer and to automatically generate technology-specific implementations. Second, DARWIN focuses on precisely tracking evolution at modeling level, being the pre-requisite for incremental change propagation to the implementation level. Third, DARWIN targets semi-automatic co-evolution of interdependent semantic artifacts to enhance maintainability while ensuring global consistency. Solutions are realized by means of a prototype and will be integrated for demonstration purposes within the commercial UML tool Enterprise Architect.