Arguing that the main promise of computers lies in their potential for moment-by-moment adaptation of instructional content and form to the changing cognitive needs of individual learners, this paper clarifies the implications of this potential for tutoring research, and points out overlooked relations between different strands of research relevant to the construction of intelligent tutoring systems. The discussion focuses on four major topics: (1) cognitive diagnosis; (2) subject matter analysis; (3) teaching tactics; and (4) strategies for teaching. The implications of the goal of providing dynamically adaptive instruction for each of these factors are analyzed, and the results of the analyses are stated as principles about intelligent tutoring. The major conclusion is that, in order to provide adaptive instruction, a computer tutor must have a strategy which translates its tutorial goals into teaching actions, and that, as a consequence, research on teaching strategies is central to the construction of intelligent tutoring systems. A four-page list of references concludes the document. (DJR)