TARGET's role in knowledge acquisition, engineering, validation, and documentation

We investigate the use of the TARGET task analysis tool for use in the development of rule-based expert systems. We found TARGET to be very helpful in the knowledge acquisition process. It enabled us to perform knowledge acquisition with one knowledge engineer rather than two. In addition, it improved communication between the domain expert and knowledge engineer. We also found it to be useful for both the rule development and refinement phases of the knowledge engineering process. Using the network in these phases required us to develop guidelines that enabled us to easily translate the network into production rules. A significant requirement for TARGET remaining useful throughout the knowledge engineering process was the need to carefully maintain consistency between the network and the rule representations. Maintaining consistency not only benefited the knowledge engineering process, but also has significant payoffs in the areas of validation of the expert system and documentation of the knowledge in the system.