A Modular System for Generative CAI in Machine-Language Programming

A computer-assisted instruction (CAI) system, which has been used to teach machine-language programming, is described. This system is generative in the sense that it can generate both meaningful programming problems and their solution programs. The problems are produced by concatenating a randomly selected set of problem primitives and parameters. These primitives are represented internally as a sequence of basic machine-language programming concepts. A list of logical sub-tasks or ``flow chart'' of the solution program is then derived and presented as a guide to the student. The degree of freedom allowed each student in the coding of his program is a function of his previous performance. Similarly, as a student advances, more of the easier sub-tasks will be coded for him. Each step of the student's solution program is monitored, and he is given remedial feedback when he is incorrect. In cases where there are many alternative correct program segments, the system will simulate execution of the student's program and determine the validity of his program segment. Results of classroom use of this CAI system are also discussed.