A blackboard-based software maintenance expert system

Intelligent Systems architectures have been proposed to tackle the problem of integration in the next generation of Software Engineering Environments. We use a model drawn from Al to address this issue, and apply the model to the process of software maintenance. In this context, we cm identify two different types of integration, respectively activity integration and views integration. Activity integration is needed to model the interaction between the maintenance programmer with abstraction recovery tools and the application domain. Views integration is crucial to represent the mental model of the system under maintanance and to perform the impact analysis when the software system evolves. To evaluate the approach we have developed a Software Maintenance Expert System (SMES) based on the blackboard architecture and used computational linguistics techniques to represent control knowledge.