This paper argues that the most important role the intelligent videodisc can fulfill is to provide a mechanism for effectively integrating testing and instruction. This integration will produce at least four important benefits: (1) increased learning by the student, (2) more interesting instructional materials, (3) gains in the efficiency of testing, and (4) less clerical work for the instructor. The success of videodiscs in instruction will depend largely on how well they maintain student motivation by providing attainable goals, frequent feedback, appropriate difficulty, and minimized failure, and whether they allow for such individual differences as attitudes and interests, time required for learning, student's knowledge base, and generalized ability. Difficulties in implementing a proposed diagnostic-prescriptive model and steps toward improving the chances for success in implementing the model include the need for large item pools and infinite numbers of different tests, scoring problems, burdensome recordkeeping, and complexity of the prescriptions. Some challenges to the advocacy of this model and its implementation are discussed. (CHC)