Jim Foley
CRA Board Chair
foley@cc.gatech.edu
College of Computing
Georgia Tech

>> Response by Robert Glass

[An earlier version of this article appeared in Computing Research News 14(4), September 2002, page 6.]

The Georgia Tech College of Computing was founded in 1990, with the broad goal of integrating computer science with other disciplines. This article is my own answer to the question "What is computing and how does it differ, if at all, from computer science?" The answer of course depends on what we mean by computer science - and not just computer science research, but computer science as a discipline.

Abelson and Sussmann [1] write "Computation provides a framework for dealing precisely with notions of 'how to'." In a recent presentation, Sussmann [2] contrasts the "how to" of computer science with the "what is" of mathematics. I would go further. Science discovers the laws of nature -the "what is" of nature. Engineering uses the laws of nature to create physical artifacts. In contrast, computer science discovers and uses the laws of "how to" compute and "how to" organize information - to create computational and information artifacts. Computer science is also concerned with the organization - that is, the architecture - of the physical artifacts that perform computations and store and transmit information. (Computer science is both science and engineering - a good thing, in my opinion, but that's another story.)

What then of the discipline of computing? I've been at Georgia Tech's College of Computing for much of the time since I first moved South in 1991. Our strategic plan [3] gives a short explanation of computing: "The College was founded to focus on computing - the integration of computer science and other disciplines to address problems of wide interest. Our interest in end-results leads immediately to our concentration on the human element of computing in much of our research and to our aggressive interdisciplinary orientation."

My longer and more detailed way to say this is that computing is computer science plus the engagement of computer science with other disciplines that:

• Inform us about people in their use of computers for any of the myriad of uses to which we put computers, ranging from serious computations to communications to entertainment; or
• Present meaningful problems/applications that can benefit from computer science solutions; or
• Suggest unsolved computer science research questions; or
• Affect the architectural characteristics and properties of the physical artifacts that perform computations and that store and transmit information.

Again - when I speak of computing and computer science, I speak of the two disciplines, not just of the research within the disciplines. Thus some discipline might "Present meaningful problems that can benefit from computer science solutions" where those solutions are known and do not require research, but do require that practitioners understand the computer science tools and techniques and theories that constitute those solutions. Researchers, on the other hand, will seek new tools and techniques and theories to solve new problems or to better solve known problems.

The concern with how people use computers (and are affected by that use) opens a rich world of research that draws as much on psychology and sociology and learning sciences and organizational behavior as it does on computer science.

Computing includes computer science, so the discipline of computing spans from deep theoretical questions about the nature of computation and information to the domains and ways in which computers are used or could be used. Hence computing simultaneously looks inward to solve its own fundamental problems and looks outward to other disciplines to join in solving their own computation-related fundamental problems or to use their methods and insights in solving computing problems.

These dual facets of computing are what I have started calling the "innerface" and the "outerface" of computing: the theoretical and systems-oriented aspects are the "innerface" while the end-user-oriented aspects are the "outerface."

Fig. 1 suggests that:

• Computing contains computer science
• The foundations of computing fall partially within and partially outside computing (coming from disciplines such as math, probability, statistics, cryptography, philosophy, linguistics, semantics, cognitive and learning sciences, library science, sociology, and anthropology). From these foundations we draw theories, methodologies, and insights.
• Working on research challenges coming from applications and other disciplines is an important element of computing
• Addressing the research challenges from other domains can involve computing research in computing foundations, computer science, or computing.

Fig. 1. A Structure for the Discipline of Computing.

The more engaged one becomes with other disciplines and human issues in one's research, the more one is doing computing research in addition to or instead of computer science research. The path from computer science to computing is a continuum: there is no hard dividing line. Many computer science departments have individual faculty members who do computing research. Many of my College of Computing colleagues at Georgia Tech do computer science research. I have done computer science research and computing research.

An important aspect of computing research is "interdisciplinary research" - a term which means different things to different people. Fig. 2 captures some possible interpretations - or misinterpretations. It is only in the upper-right quadrant that I consider true interdisciplinary research to be occurring. Yes, sometimes a research team will start in one of the other quadrants as a "get acquainted with the other discipline and problem" strategy, but the target must also be the upper-right quadrant. This is where neither computing nor the other discipline can alone create new knowledge. Indeed, an interdisciplinary research team might recognize new problems that neither discipline alone would recognize, and in some cases the collaboration can lead to fundamentally new ways of thinking about problems.

Fig. 2. Interdisciplinary research occurs when new knowledge is being discovered in both disciplines.

How does computer engineering fit into the picture? Here again is a continuum. The distinction between the two is blurred and should be even further blurred than organizational boundaries often suggest or impose. There are differences of degree, not of kind. Computer engineering has a greater concern than does computing with implementation as opposed to architecture. Computer engineering more often uses computers as a sub-system of a larger system, in the sense - compute - actuate cycle as in automotive electronics, avionics, and control systems in general. Hence computer engineering often has more emphasis on controlling the physical world, while computing often focuses more on the information world. But there is considerable overlap: at the heart of things is a shared concern with "how to."

What of the relationship between computing and the two areas known as Information Technology and Computational Science? Fig. 3 shows two types of application domains - those that are compute-intense and those that are data-intense (the two overlap because there is a continuum here as well). Computational science is computing with a focus on compute-intense domains; information technology is computing with a focus on data-intense domains. Again, each of computational science and information technology can involve research in computing foundations, computer science, or computing.

Fig 3. How computational science and information technology relate to computing.

Computational science and information technology are not the sole domain of computing. Sometimes they housed in a computing college with strong links to the other discipline, sometimes in a separate department bearing the name of the discipline (such as medical informatics), sometimes in a larger department with strong links to computing or computer science (operations research is often found in industrial engineering) or in an interdisciplinary degree program managed by both disciplines.

My understanding and articulation of the structure of the discipline of computing is still evolving. I welcome your feedback, foley@cc.gatech.edu. My thanks to Elizabeth Mynatt and Janet Kolodner for helpful comments on an earlier version of this article.

References:

[1] Abelson and Sussmann, Structure and Interpretation of Computer Programs, MIT Press, Cambridge MA 1985.

[2] Sussmann, G., The Legacy of Computer Science, presented at NRC/CSTB Symposium on Fundamentals of CS, www7.nationalacademies.org/cstb/project_fcs_agenda3.html .

[3] http://www.cc.gatech.edu/staff/f/freeman/strategicplan.html