Virtual machines and practical languages: the development of programming notations in the 1950s

Im Forschungskolloquium werden zum einen laufende Examensarbeiten und Promotionsprojekte vorgestellt und diskutiert, zum anderen kommen in eingeladenen Vorträgen aktuelle Forschungsthemen der Wissenschaftsgeschichte zur Sprache. Das Kolloquium steht allen Interessierten offen und richtet sich besonders auch an Masterstudierende. Masterstudierende, die einen Vortrag halten möchten, mögen sich bitte spätestens im Ende März beim Leiter des Forschungskolloquiums melden.

_____

Mark Priestley: Virtual machines and practical languages. The development of programming notations in the 1950s

In the 1950s, knowledge about computers was hard to come by. Programming was widely viewed as an esoteric practice demanding a level of technical sophistication beyond what could be expected of a practitioner in a different field. Nevertheless, in part precisely because of the lack of widespread computing expertise, it was a widely accepted goal that scientists and others should be able to use computers directly in their work. One approach to making this feasible was to design notations, later called “programming languages”, to mediate between users and the machine. In many cases, these reflected the practices of a particular field in a machine-processible way, thus making programming more accessible to non-specialist users. Familiar examples include Fortran, which allowed scientists to program using familiar mathematical equations, and Cobol, intended to allow business users to program using the language of commerce. At the same time as they facilitated the encoding of domain-specific knowledge and technique, however, these programming languages interpreted the computer for non-specialized users. Learning a programming language gave users access to a “virtual” computational device that was often simpler or better tailored to a specific application area than the physical machine on which the language ran. I will describe how programming languages in the 1950s thus formed “epistemological bridges” between apparently incompatible areas of knowledge, allowing users to grasp some of the ideas underlying the new computers and simultaneously express at least some of their domain-specific knowledge and problems in a computational form. I will contrast this thesis with the view that the development of programming languages was largely a process of abstraction away from the machine and argue that it challenges the hardware/software dichotomy which continues to shape much thinking about the computer itself.

Mark Priestley is a historian of computing whose primary interest is in the development of the theory and practice of computer programming. After an undergraduate degree in Mathematics and Philosophy, he worked as a programmer and university lecturer before undertaking a doctorate in the history and philosophy of computing. His thesis examined the influence of logic on the early development of programming languages and was published as A Science of Operations (Springer, 2011). From 2012 to 2016 he was engaged in the ENIAC in Action project (MIT Press, 2016), and later projects have included studies of the operation and use of the Colossus machines used at Bletchley Park in World War 2, and of John von Neumann’s programming work in the 1940s. The results of the latter are described in his latest book, Routines of Substitution (Springer, 2018). He is currently Senior Research Fellow at the UK’s National Museum of Computing.