Science and computer science
Nathaniel Stewart · ACM Computing Surveys · 1995
During the few decades that computer science has been identi ed as a separate discipline, the question of the intellectual nature of the subject has frequently been posed. In particular, the question was discussed recently in the 1993 Turing Award lecture of Juris Hartmanis, \\On computational complexity and the nature of computer science " [1]. It is clear that there are components of computer science that could be viewed as sub elds of mathematics or engineering. Indeed, competent work in theoretical computer science meets rigorous mathematical standards, and competent work in applied computer science meets the standards of good-quality engineering work (prototypes are built, proof-of-concept projects are conducted, and results are evaluated on the basis of their usefulness in practice). The open question is the extent to which the remaining parts of computer science can be viewed as science. To study this question is not necessarily to engage in sterile debate: it may be, for example, that the eld is not yet a science, but that it could become one, and that certain policy changes could accelerate this process. Hartmanis argues [1, 2] that computer science is di erent enough from the other sciences to permit di erent standards in experimental work, and that computer-science \\demos " can be viewed as a replacement for the experimentation found in other elds. I do not agree. Computer science is, or has the potential to be, a science similar in character to physics and the other natural sciences. However, its traditions, in the areas of experimentation and formulation of theories, may delay its acceptance and inhibit its development (as a science).