Great Principles of Computing by Peter J. Denning and Craig H. Martell (review)

One typical feature of the (short) history of computing as a discipline is the debate on its different disciplinary traditions: is it a kind of mathematics, is it engineering, or rather is it a science?¹ One of the principal advocates of the science view on computing is Peter Denning, past president of the ACM and an influential figure in US-based computer science policy and education. Great Principles of Computing, coauthored with Craig H. Martell, develops Denning’s most recent views on the science of computing and the so-called “great principles framework” (see also http://denninginstitute.com/pjd/GP/GP-site/welcome.html).

Peter Denning is well known not just for his work in computing—most notably his work on operating systems and virtual memory—but perhaps even more for his work on computing. Since the early 1970s, Denning has been a prominent figure in the disciplinary debates on computing often in connection to computer science education and funding. In 1971 he chaired the task force COSINE to develop an undergraduate course on operating systems. Already at that time, Denning was advocating an approach based on first principles in order to provide a sound conceptual basis for computing. This concern was motivated by among others, the need to apply these principles in the industry in order to tackle problems of hardware and software complexity in a more systematic manner.² In the early 1980s Denning, then president of the ACM, was actively seeking more recognition, both financially and intellectually, for experimental computer science.³ His concern is rooted in the idea that computer science funding will decrease if the field doesn’t provide decent experimental work, one of the classic benchmarks used in the traditional sciences.

This emphasis on experimentation recurs in Denning’s later works on the science of computing where it is now integrated within the more general view of computing as the science of information processes, natural and artificial. Despite Denning’s outspoken view on computer science as a science, this view does not exclude the two other research traditions of mathematics and engineering. In the 1980s, a task force chaired by Denning to reflect on the core of computer science presented a highly influential report, “Computing as a Discipline,” which presents computing as an interdisciplinary endeavor involving mathematical, engineering, and scientific practices instantiated in computing as theory, abstraction (modeling), and design. The report concluded that “the three processes are so intricately intertwined that it is irrational to say that one is more fundamental than the other.”⁴

In Great Principles of Computing, Denning certainly acknowledges this interdisciplinary character of computing. Perhaps one of the clearest indications of this is the emphasis throughout the book on the significance of both the material and physical aspects of computing (such as architecture) and its more abstract developments (including algorithms). However, from Denning’s viewpoint, the emphasis is on the science of computing and, more particularly, a science that can be described from its (first) principles. A principle of computing, in Denning’s vocabulary, is concerned with “manipulating matter and energy to produce intended computations” (p. 11). Defined in that way, this principles approach “emphasizes the fundamental laws that empower and constrain ... [computing] technologies” (p. 11). The principles fall into six categories: communication, computation, coordination, design, evaluation, and recollection.⁵ These categories are nonexclusive and are understood as possible perspectives on computing. For instance, security has principles coming from all six categories and can thus be viewed from the perspectives of communication and computation.

As was the case when he was chairing the COSINE task force, Denning is convinced that a principles-based framework allows for a more coherent and systematic approach on computing, as opposed to a presentation of the field’s content according to its core technologies like algorithms or data structures. According to Denning, such a core technologies approach gives an overly technological interpretation of the field. This reflection on the best way to present the field of computing is...