Remembering Michael S. Mahoney

Michael S. Mahoney, professor of the history of science at Princeton University, died in 2008. Born in 1939, Mahoney was already a seasoned historian of mathematics when he became one of the first senior historians to take an interest in the history of computing. He was by no means the first: for example, individuals such as I. B. Cohen at Harvard University and Derek de Solla Price at Yale University had been interested since the 1960s. Moreover, several institutions were already actively engaged: the Smithsonian Institution, the Charles Babbage Institute (founded 1979), the Computer Museum in Boston, and similar organizations in Europe. A journal, the Annals of the History of Computing, was established in 1979. However, Mahoney’s unique mission was not so much to do the history of computing, as to see it done well.

The collection of papers on the history of computing, admirably edited by Thomas Haigh, stands as a memorial to Mahoney. ¹ It is arranged in three categories which effectively define Mahoney’s contributions to the discipline: the shaping of the history of computing; constructing a history of software; and the structures of computation. Haigh has provided an excellent, and by no means uncritical, commentary on the collection. In this essay, I will follow Haigh’s organization.

1. The Shaping of the History of Computing

Mahoney burst on to the scene with a paper “The History of Computing in the History of Technology” published in the Annals of the History of Computing in 1988. In this somewhat polemical paper he urged both historians and computing practitioners to take up the challenge of doing his history well. The paper is highly characteristic of Mahoney’s rhetorical style, and to read it is to hear his voice. Mahoney loved teaching, especially undergraduate and lay audiences, and he had a voice that projected far. Unknowingly he tended to use the same rhetorical boom regardless of the size of his audience or auditorium. I always tried to sit at the back where the volume was lowest. This paper was delivered at a time when the literature of the history of computing was quite thin. Today, twenty five years later and with the publication of hundreds of scholarly articles and scores of monographs, the paper is a snapshot of an emergent discipline. Mahoney played an important part in establishing the validity of this discipline.

Mahoney’s comments were often addressed to practitioners rather than historians. Indeed he had an admiration and affection for computer professionals. Typical was his advice on doing history well given to speakers at the History of Programming Languages Conference in 1996. One of his recurrent themes was expressed here quite brilliantly:

When scientists study history, they often use their modern tools to determine what past work was “really” about; for example, the Babylonian mathematicians were “really” writing algorithms. But that’s precisely what was not “really” happening. What was really happening was what was possible, indeed imaginable, in the intellectual environment of the time; what was really happening was what the linguistic and conceptual framework then would allow. The framework of Babylonian mathematics had no place for a metamathematical notion such as an algorithm. (p. 39).

As it happened, in 1972 the world’s most famous theoretical computer scientist Donald Knuth had written a widely admired paper “Ancient Babylonian Algorithms” which took exactly the line that Mahoney discouraged. Not many computer scientists would have criticized Knuth, albeit obliquely; this was a powerful message. Mahoney probably never realized it, but he was quite an intimidating authority figure and his strictures had the potential to discourage amateurs from engaging with the history of their subject. This is not to say that Mahoney was personally intimidating—far from it, he was kind and encouraging in person—but he carried everywhere the prestige of a professorship in one of the world’s great universities.

2. Constructing a History for Software

Mahoney was primarily interested in theoretical computer science and so-called systems software (that is, infrastructure software such as operating systems). This interest tended to constrain his view of the broad spectrum of software, which properly...