Book reviews

Within the expert-systems community, 'knowledge acquisition' refers to the necessary task of eliciting and representing expert knowledge for coding as a computational system. This practically motivated task cannot avoid an issue of central import to a theory of machine intelligence: the relationship between the human knowledge to be acquired and the represented knowledge to be stored in the computer. Through intended to provide guidance for the knowledge elicitor, Knowledge Acquisition is therefore most interesting to readers of Minds and Machines for its insistence on a distinction between human knowledge and represented knowledge, and its focus on knowledge acquisition rather than knowledge representation as the key to system flexibility and usability. Brul6 and Blount believe that the computer metaphor for human knowledge has dominated many approaches to knowledge acquisition, resulting in inevitable difficulties: