On the shortcomings of the axiomatic approach as presently used in computer science

In current practice, many issues in computer science are formulated and treated as formal logic theories, even though this style of axiomatization is inappropriate for the topic at hand. Comparison with the scientific approach in more mature disciplines, such as mathematics, physics and applied science, reveals the causes and the unwanted effects of this phenomenon and shows, more generally, that the methodological basis for computer science is still far from being fully developed. The impact on a few important research areas is briefly discussed. It is argued that this state of affairs can be improved by recognizing that the traditional characterization of the discipline of computer science, which may have been appropriate during the initial years, is now due for a major revision, and by broadening the 'pure' computer science curricula through the inclusion of judiciously chosen additional topics in mathematics, physics, and systems theory.<>