Cognitive model and problem-solving processes of computer programmers

Abstract

Although researchers have started concentrating on the programming behavior of computer programmers, the material produced so far has not been merged into a testable theory; each study focuses on a particular problem without integrating the results into a workable overall model. In addition, most studies have concentrated on the measure of programming skills and aptitudes, rather than on the underlying cognitive processes differentiating programmers from non-programmers. While many studies focus on the human-system interaction, very little has been attempted to specify which strategies computer programmers use to solve problems, how they process and integrate information; whether these strategies are specific to a certain type of people, or whether these processes can be taught and improve with experience. Two partial models were examined, the syntactic/semantic model of programmer behavior, and the heuristic/algorithmic problem-solving model, in order to attempt to build a stronger base for the evaulation of programmer aptitudes. Analysis indicated that: on Master Mind, Experienced Programmers performed better than the General Population; Novice Programmers performed better than the Control Group; and Novice Programmers performance was related to their grade on the BASIC course. On Semantic Ability Test, Experienced Programmers performed better than General Population, althought not significantly so; Novice Programmers performance was related to tehir grade on the BASIC course, although not significantly so. (Abstract Shortened)