Abstract :
This paper describes a CAI system being used in an introductory computer science course. The system is generative in the sense that it generates its own problems, derives its own solutions, monitors student solutions, and generates remedial statements. Due to its generative nature, the system provides a great deal of individualization of instruction. It selects the path taken by each student through the course concepts on an individual basis as a function of its prior experience with the student. It also tailors the difficulty of the problem generated, the degree of monitoring, and the amount of help provided to suit each student´s past performance. This paper describes the special properties of generative CAI which make it particularly useful for problem-oriented courses. It discusses how the teaching algorithms are designed and how problems are generated. It also describes how the system algorithms can be used to analyze student responses, determine the nature of the student error, and generate pertinent remedial feedback.
