As the means to investigate domain-dependent use of Prolog Programming Techniques, we have built a system which can automaticly synthesise models (Prolog programs) from user specifications. At this stage, we would like to assess the effectiveness of both, the concepts applied and the system produced.
At this informal talk I will explain the design I've been doing for the system's evaluation. I would like to relate that with earlier KBS talks on Evaluation and Methodology in AI and/or HCI usability testing.