Identify Blackboard Pattern Based on Ontology

Abstract

Blackboard pattern identification is significant for the programmer to maintain the software system. Whether and how the system uses the blackboard pattern could help the programmers unfamiliar with the target system. This paper proposes a blackboard-instance identification approach based on ontology, which not only judges whether the target system uses the blackboard pattern but also provides the blackboard pattern implementation of the target system. The target system is described by ontology and input into the ABox of the knowledge base, the blackboard pattern is described by ontology and input into the TBox of the knowledge base. And the inference engine will reason out the raw pattern instance. Finally, the final pattern instance will be outputted by iterative refinement. To study the accuracy of our approach, sixty-eight projects have been tested and two of them have been analyzed the components' identification accuracy.