BY NATHAN ENSMENGER / WILLIAM ASPRAY
In recent years labor historians have devoted considerable attention to issues of race and gender in the history of labor-management relations and the dynamics of the workplace environment. The conventional wisdom argues that corporate managers often use women and minorities as low-wage, low-skill replacements for skilled white male laborers. Occupations tend to become sex-typed as being either male or female, depending on their relative position in the wage and status hierarchy. An influx of women and/or minorities into an occupation is usually considered to indicate that routinization, degradation, and deskilling has occurred. Women have rarely held high positions within the scientific or engineering community in significant numbers, at least until fairly recently.
There is evidence that the story of gender and software labor is a little less clear-cut. As a number of scholars have suggested, women have played an important role in the history of software development. The first ENIAC programmers were women, and Jennifer Light has argued that these women significantly influenced early computing and programming practice.100 The Association for Computing Machinery's first "Man of the Year" was a woman.101 Women have not only held a greater percentage of jobs in software than might otherwise have been expected, they were also able to advance farther and faster than there peers in other high-tech industries. Clearly there is something interesting going on in the history of the software professions that deserves further scholarly examination.
What do we know about women and software? Women were the very first programmers, or Â»codersÂ« as they were called in the earliest years of computing. The intended role of these women was clearly articulated in the three volumes on "Planning and Coding of Problems for an Electronic Computing Instrument," written by Herman Goldstine and John von Neumann in the years between 1947 and 1949.102 These three volumes served as the principal textbooks on the programming process at least until the early 1950s. The Goldstine/von Neumann method assumed that the computer would be used for complex scientific computation, and the division of labor in the programming task seems to have been based on the practices used in programming the ENIAC. Goldstine and von Neumann spelled out a sixstep programming process: (1) conceptualize the problem mathematically and physically, (2) select a numerical algorithm, (3) do a numerical analysis to determine precision requirements and evaluate potential problems with approximation errors, (4) determine scale factors so that the mathematical expressions stay within the fixed range of the computer throughout the computation, (5) do the dynamic analysis to understand how the machine will execute jumps and substitutions during the course of a computation, and (6) do the static coding. The first five of these tasks were to be done by the "planner" who was typically the scientific user and overwhelmingly often was male; the sixth task was to be carried out by "coders"--almost always female (on the ENIAC project). Coding was regarded as a "static" process by Goldstine and von Neumann, one that involved writing out steps of a computation in a form that could be read by the machine, such as punching cards, or in the case of ENIAC in plugging cables and setting switches. Thus there was a division of labor envisioned that gave the most skilled work to the high-status male scientists and the lowest skilled work to the low-status female coders. It turns out that the coders on the ENIAC project ended up doing many more tasks than envisioned. Programming was a very imperfectly understood activity in these early days, and much more of the work devolved on the coders than anticipated. To complete their coding, the coders would often have to revisit the dynamic analysis; and with their growing skills, some scientific users left many or all six of the programming stages to the coders. In order to debug their programs and to distinguish hardware glitches from software errors, they developed an intimate knowledge of the ENIAC machinery. "Since we knew both the application and the machine," claimed ENIAC programmer Betty Jean Jennings, "we learned to diagnose troubles as well as, if not better than, the engineers."(103) Thus what was supposed to have been a low-skill, "static" activity prepared these women coders well for careers as programmers--and indeed, those who did pursue professional careers in computing often became programmers and did well at it. A few women, Grace Hopper and Betty Holberton of UNIVAC and Ida Rhodes and Gertrude Blanche of the National Bureau of Standards in particular, continued to serve as leaders in the programming profession. (104)
However, during the 1950s, business applications began to surpass scientific applications; a computer manufacturing industry grew up to service the rapidly expanding need for computers for business applications; and a tremendous demand grew up for programmers. The number of new programmers, most of whom were male at first, swamped the number of female coders who had become programmers. Programming quickly became primarily a man's job. If the Braverman/ Kraft thesis about the deskilling of programming labor were correct, we would expect to see the employment of women in software increase as the occupation became less skilled and more routine. In a 1964 survey, 76 percent of the respondents expected to see the ratio of women in programming increase: "The only limitation is the number of qualified applicants," stated one manufacturer. (105) There are indications that certain types of female employees were seen, at least in the 1960s, as being more stable and reliable than their male counterparts, based upon some typical sexual stereotyping: "Women are less aggressive and more content in one position...Women... are more prone to stay on the job if they are content, regardless of a lack of advancement. They also... are less willing to travel or change job locations, particularly if they are married or engaged. For these reasons there is a considerably lower turnover rate in women programmers and as a result, the initial investment in training pays a greater dividend for their employees." (106) Employers were warned away, however, from hiring "the most undesirable category of programmer," the female "about 21 years old and unmarried," who was likely to marry, become pregnant, or waste precious energy worrying about her social commitments for the weekend. (107)
There is no doubt that some male programmers were threatened by a perceived incursion of females into their profession. For many of these men, women were associated with low-skill clerical labor, even though many of the ENIAC `girls' had actually possessed college degrees in mathematics. The new generation of female programmers was being recruited from the ranks of keypunch operators or `coders.' In an era when programmers were anxious to distinguish programming as a creative intellectual activity from coding as manual and narrowly technical labor, these women represented the lowest rungs of the occupational hierarchy ("There's nothing lower than a coder") (108). An influx of low-skill, lowwage labor threatened both the professional selfidentify of the programmers and their superior bargaining position in the labor market for software workers. It is hard to imagine, therefore, that they would have been pleased or flattered by Helen Gurley Brown's exhortation to the readers of Cosmopolitan that they go out and get jobs as programmers making $15,000 after five years. (109) Many of the advertisements for "automatic programming" languages and systems used women as a proxy for less expensive, more tractable labor. If you could teach your secretary to program in COBOL, there was no need to pay for expensive programming talent. There are other historical questions to be asked about gender and software labor. Recent statistics on computer science enrollments and software industry employment indicate that the number of women in computing has been dropping since the early 1980s. Why? It has been argued that many women perceive computer careers as being overly competitive, incompatible with a well-rounded family oriented lifestyle, and solitary rather than social. (110) Writers such as Sherry Turkle and Tracy Kidder have described the various ways in which the programmer subculture emphasizes culturally masculine traits such as competitiveness, practical joke playing, and aggressive hacking and cracking. (111)
How and why did this masculine subculture develop? How does it relate to the perpetual software labor crisis? Anecdotal evidence suggests that women are attracted to programs in information systems, rather than computer science or computer engineering, because "information systems is perceived as more people-oriented and more attuned to the uses of information technology." (112) What does this tell us about the historical and social construction of computer knowledge and specialties? In what ways has the absence of women from the programming profession been used to emphasize its rational, "scientific" qualities? Labor historians have developed an extensive literature on work and gender; historians of software should make use of their expertise and experience.
99 Willis Ware, "As I See It: A Guest Editorial," Datamation 11, 5 (1965), 27 100 Jennifer Light, "When Computers Were Women," Technology & Culture 40, 3 (July, 1999). 101 Admiral Grace Hopper received her "Man of the Year" award in 1962. Needless to say, it was extremely unusual for an association of technical professionals to grant its highest honor to a woman, especially in the early 1960s! 102 These technical reports are most easily found today in reprint form in William Aspray and Arthur Burks, eds., The Papers of John von Neumann on Computing and Computer Theory (Cambridge, MA: MIT Press, and Los Angeles: Tomash Publishers, 1987). 103 Fritz, p. 20. 104 Frances Elizabeth ("Betty") Snyder Holberton was awarded the Association for Women in Computing's Ada Lovelace Award in 1997. Grace Hopper described her as being "the best programmer that she had known during her long career." (W. Barkley Fritz, "The Women of Eniac," Annals of the History of Computing 18, 3 (1996)). 105 Report, "Advanced Programmers, Women Employment Seen Rising," Datamation 10, 2 (1964) 106 Valerie Rockmael, "The Woman Programmer," Datamation 9, 1 (1963), 41, 41 107 William Paschell, Automation and employment opportunities for office workers; a report on the effect of electronic computers on employment of clerical workers (Washington, D.C: Bureau of Labor Statistics, 1958); also Rockmael, p. 41. 108 "Checklist for Oblivion," Datamation 10, 9 (1964) 109 The quote from Helen Gurley Brown appears in an advertisement for the Computer Sciences Corporation, "In case you missed our first test," Datamation 13, 9 (1967). 110 Peter Freeman and William Aspray, The Supply of Information Technology Workers in the United States (Washington, D.C.: Computing Research Association, 1999), p. 113. 111 Sherry Turkle, The Second Self: Computers and the Human Spirit (New York: Simon & Schuster, 1984); Tracy Kidder, The Soul of a New Machine (New York: Avon Books, 1984) 112 Freeman and Aspray, p. 111 http://www.sas.upenn.edu/~nathanen/software. pdf