Abstract
The need for systems of classifying tasks which would lead to improved generalisations and predictions of human performance is stressed. A learning and performance theory which ascribes task dimensions a central role is considered critical. Some issues and alternatives in developing and evaluating taxonomies of human task performance are discussed. Recent work with the ‘ability requirements’ and ‘task characteristics’ approaches are reviewed and empirical work with them are evaluated. Particular attention is given to their utility in describing human task performance, predicting learning, standardised task batteries, estimating job task requirements, and organising findings in the performance literature. A research paradigm and a number of recent applications are described for linking the abilities and task characteristic approaches. A taxonomic system which combines these approaches appears to have promise for enhancing predictions and generalisations about human performance.