Browse Books

Reviewer: Eric A. Weiss

The “father of expert systems” (as he is called in the jacket blurb) and his two associates bring the message of what expert systems are and what they are doing right now for visionary companies. The bulk of the book recites the experiences of 21 organizations, largely US and Japanese, in the creation and use of expert systems— stories that are interspersed with a few definitional and tutorial chapters. In an appendix, Paul Harmon lists about 140 expert systems that are currently in use, each with a single-paragraph description. (Interestingly, none of them are for the task of creating expert systems.) A short glossary includes these definitions: Artificial intelligence (AI) is a subfield of computer science concerned with the concepts and methods of inference by a computer and the symbolic representation of the knowledge used in making inferences. An expert system is an AI program that achieves competence in performing a specialized task by reasoning with a body of knowledge about the task and the task domain. Knowledge consists of facts (statements whose validity is accepted), assumptions, and heuristics. “Competence” and “reasoning” are undefined. On the basis of clear differences in style, I suspect that the few excellent definitional and tutorial parts of the book were written by Feigenbaum, while the application yarns were written by McCorduck in the exciting form of popular prose associated with Tom Peters, who contributed an enthusiastic foreword. Although said to be addressed to senior management, this book will be most useful as a necessary source of argument and persuasive tales for those inside-the-company champions of expert systems who are essential to the introduction of such systems, and whose careers hang on their acceptance and their success or failure. The inexcusable lack of an index almost disqualifies the book from serious consideration.