Provides design guidelines for producing successful expert critiquing systems and presents COPE, a critic research testbed. It also unifies the two traditions in the study of human expertise which have been directed towards reducing expert error and judgement bias.