What Should Be Automated?
Matti Tedre

Tumaini University | matti.tedre@tumaini.ac.tz

 

One of the most influential figures in the early development of computer science, George Forsythe, argued in 1968 that “the question ‘What can be automated?’ is one of the most inspiring philosophical and practical questions of contemporary civilization [ 1].” Almost 20 years later, Peter Denning wrote that computer science is “the body of knowledge dealing with the design, analysis, implementation, efficiency, and application of processes that transform information” and suggested that “What can be automated?” is the fundamental question underlying all of computing [ 2]. That question emphasizes the very foundations of computing as a discipline—it asks, in a very general way, what in principle can be automated with any kind of machinery. Later Denning et al. [ 3] refined the question to “What can be (efficiently) automated?” Although the discipline of computing has diversified greatly since 1968, this “fundamental question” has rarely been challenged.

In addition to that one fundamental question, Denning listed 11 topic areas of computing and outlined a number of fundamental questions asked in each topic area [ 2]. In the end Denning had 50 questions altogether. However, in 19 of the 50 fundamental questions that Denning mentioned, the question deals with how instead of what. Denning’s 50 fundamental

questions include questions such as: “How can large databases be protected from inconsistencies generated by simultaneous access…?” “How can the fact that a system is made of components be hidden from users who do not wish to see that level of detail?” and “What basic models of intelligence are there, and how do we build machines that simulate them [ 2]?”

The theoretical question “What can be automated?” is the central question in computability theory, which studies what in principle can be computed with any kind of machinery. The theoretical question “What can be efficiently automated?” is one of the central questions in computational complexity theory, which studies the amount of resources, such as time and storage, that it takes to solve different kinds of computational problems. Researchers in those fields often are indifferent about the specific technologies that might be used to automate processes. Many computing researchers do not, however, stop at theories about automation, but their work includes implementing systems that automate various things. “How can one automate things efficiently and reliably?” is the crucial issue for practically oriented computing researchers [ 1, 3, 4]. Although scientifically and practically oriented researchers have different aims from theoretically oriented researchers, the theoretical questions about

limits of computing are fundamental to empirical research on computation and computers, and crucial to the design and implementation of computing systems.

Answers to questions that ask what are different from answers to questions that ask how. Generally speaking, theoretical and empirical scientists tend to ask what and why, and engineers tend to ask how. The theoretician’s questions—what can be automated? and what can be efficiently automated?—are concerned with finding out which processes can be automated and which cannot, and finding out which processes can be automated efficiently (according to some criterion of efficiency). For the theoretician’s questions there can be straightforward yes/no answers. For the practitioner’s question, “How can process p be automated?” there usually are many competing answers. For instance, implementations a, b, and c can all automate process p and be equally efficient in their use of space, time, and other resources, or they can all be non-optimal in different ways.

The theoretician’s and practitioner’s questions above imply that computing researchers deal with issues of computability, tractability, efficiency, operation-ality, usability, maintainability, and reliability. Those issues are concerned with what can be efficiently automated, why some things can be efficiently auto-

[ 1] Forsythe, George. “Computer Science and Education.” In Proceedings of IFIP Congress 1968, 92–106. Edinburgh, UK: August 5–10 1968, Volume 2.

[ 2] Denning, Peter J. “The Science of Computing: What is Computer Science?” American Scientist 73, no. 1 (1985): 16–19.

[ 3] Denning, Peter J., Comer, Douglas E., Gries, David, Mulder, Michael C., Tucker, Allen, Turner, A. Joe, and Paul R. Young. “Computing as a Discipline.” Communications of the ACM 32, no. 1 (1989): 9–23.

[ 4] Raatikainen, Kimmo. “Issues in Essence of Computer Science.” An English translation of essays for Tietojenkäsittelytiede 2/3 (1991-1992). Available at https://www.cs.helsinki. fi/u/kraatika/Papers/ IssuesInEssenceOf-ComputerScience.pdf

September + October 2008