ABSTRACT
This chapter examines the way costs of knowledge-based systems that are calculated and segregated, and then introduces a low cost, high yield approach to the process of developing knowledge-based systems that can result in dramatic reductions in costs. There are relatively few knowledge engineers currently available compared to the demand for their services. The most compelling statistic is the life-cycle cost of shells compared to high end systems. Shells are obviously easier to use, since they include the inference engine, dialog system and knowledge base software already programmed and the user/programmer need only insert the logic and dialog messages as necessary. An alternative, offered only for the numerous expert systems (ES) environments that are smaller in scope, is the use of shell-based tools, which have unit costs in the range of ten per cent of the high end systems. The unit cost and the total cost of high end ES are, in the million dollar range for large, complex systems.
