Specialized knowledge acquisition tool support compared to manual development-a case study

A case study into the effect of applying a domain-oriented knowledge acquisition (KA) tool to an expert system project is presented. A KA tool that acquires knowledge for experiment planning has been implemented. The effects of applying this tool have been evaluated and compared to manual development by conventional means. A manually acquired knowledge base has been used as a reference. To give an appropriate background, brief descriptions of the domain, the expert system and the KA tool are provided. Lessons learned considering productivity, flexibility, training needed, etc. are the main topics discussed. A salient result from this study is that the effort required to implement the KA tool, compared to the effort for the corresponding expert system, is lower than expected. Specialized KA tools may even prove to be economical for single applications