The special problems of experimentally evaluating knowledge acquisition and knowledge engineering tools, techniques and methods are outlined, and illustrated in detail with reference to two series of studies. The "rst is a series of experiments undertaken at Nottingham University under the aegis of the UK Alvey initiative and the ESPRIT project ACKnowledge. The second is the series of Sisyphus benchmark studies. A suggested programme of experimental evaluation is outlined which is informed by the problems with using Sisyphus for evaluation.

No abstract

No abstract

No abstract