3. Masculinity and the Machine Man

At a Houston insurance company, computerized data entry (background) replaced conventional keypunching (foreground), resulting in a reported 20% cost saving- but little change in women's work.

While academic computer science has suffered a decline in women's participation since the mid-1980s, corporate computing has been full of women since the beginning. A typical punch-card operation employed roughly equal proportions of men and women, but in different roles. Women were mainly to be found coding data from a keyboard onto punched cards, while men were found feeding the cards into specialized, non-programmable machines like collators, sorters, tabulators, and multipliers. There was little room for advancement as a keypunch worker. Rarely did they advance to machine operators, and never to the level of supervisors. Starting in the 1950s, computers began to see widespread use in administrative settings, for tasks such as payroll processing, billing, and accounting. The most successful computer models in the 1950s and early 1960s were extensions of and compliments to existing punched card machines, and the staff that used them were mostly the existing punched card machine operators and midlevel staff of the department being computerized. Occasionally larger companies needed to plan for the new administrative procedures and would hire systems analysts, a booming field after World War II, who worked on the design of data processing applications. Another new job title at the time was that of programmer. Different from the routine mathematical labor of early scientific computing, it was instead a hybrid of aspects of operators and systems analysts and inherited the gendered division of both those titles.

This division was often pegged as natural preferences of men and women for certain types of work, but labor historians now understand this to be a form of "occupational sex typing." Sex typing is presenting gender segregation based on natural aptitude yet using shifting definitions and characteristics of aptitude to emphasize Charteris tics associated with one gender over another. It is also worth noting that at this time, gender discrimination was legal, and many occupations specified gender in their descriptions. The sex typing of data processing jobs has been put down to the "preferences" of machine supervisors, who were 99% men (a 1960 survey of almost five hundred companies found only one who reported a woman supervisor, and only two had managers that were women). There was an incentive to keep a firm gender separation between data processing work, leftover from the clerical labor market. In the 1870s, clerical work was mainly performed by men and was seen as a good first step for an aspiring business man, but by the 1920s was predominantly women's work, with the low-pay dead-end jobs seen as work that a woman could do for a few years before marriage, and this division was maintained through pay differentials and policies such as firing a woman upon marriage.

In 1971, Honeywell's Keyplex Data Entry System seemed to promise tidy masculine management over female clerical workers.

By the late 1950s, "Data processing supervisors were not content merely to defend the status quo. The arrival of the computer and the ever-increasing importance of data processing promised ambitious men the chance to elevate their positions within the corporate hierarchy." The manager position became the new definition of masculinity, but most firms promoted from within, and thus needed to convince young men to take on entry-level positions, which they were reluctant to do if the same position could be occupied by a woman. This status anxiety caused companies to physically separate men and women, and to retitle positions based on gender.

One unidentified woman is in the back row at far left, 1951.

The National Machine Accountants Association, later renamed the Data Processing Management Association, is a good case study for the treatment of women in these early days of computing. In 1953, about 10% of the members of the NMAA were women, and the first few issues of The Hopper, the association newsletter, included photos of minor Hollywood starlets scattered throughout as pin-ups. The main activity entrusted to women was organizing the "Ladies Program" for its annual meeting, plans for which included fashion shows, brunches, talks on interior decorating and gourmet dining, and a session entitled "Women and Automation." Keypunch workers were not welcome as members of the NMAA at all, and barely let in keypunch supervisors, many of whom were women that had advanced from the role of keypunch worker. In 1962 an association spokesperson said that they needed to "upgrade the Association... and get a better caliber of person interested. I think we could well lose some key punch supervisors and pick up systems analysts." Trade literature from the time also reflects this gender division, with advertisements for office machines featuring predominantly young, skinny, white women who were sometimes explicitly sexualized, and those for scientific computers showing white men in dark suits. Some advertisement took the denigration of women even farther, claiming that their machines were so easy to use that women could be hired now for their looks alone, rather than their intelligence, or joking that the system could understand Dumb Blonde.

Grace Hopper received the DPMA's inaugural Computer Sciences Man of the Year award at its 1970 annual meeting in Seattle. She is hugged by Cal Elliot, its executive director.

It was not until 1970 that any explicit claims of sexism appear in data processing literature, and that same year the DPMA awarded its first "Computer Sciences Man of the Year" award to Grace Hopper, but this progress was met with hostile backlash. In the early 1970s, the majority of women in data processing were in data-entry jobs, mainly still keypunch operators. By 1982, the relatively new category of computer and peripheral equipment operator, was composed of nearly two-thirds women, owing perhaps to the increasing use of minicomputers and new word processing technology that allowed typists and clerical workers to reclassify as operators. Through 1992, the total number of women in computer-related occupations continued to exceed the total number of men, but they remained clustered at lowest levels of work. From 1982 to 1992, the percentage of women in programmer positions remained steady at 37%, but the field itself grew by two-thirds. The number of women in computing since 1992 has remained around 1.5 million, but the positions that women occupy have shifted from the low-level data-entry work rapidly towards systems analysis and management. While the amount of women in programmer positions has continued to fall from 2002 to 2006, and it is important to note that this position has fallen in importance over time, the percentage of women in high-status computer positions has actually increased by 9%.

There is a tendency in examining women within the history of computing to start with the women programming ENIAC and then wonder why there were no women left in the "computing profession" a decade later. This is problematic for a few reasons. Firstly, there is no one computing profession, especially in today's world where the majority of people interact with computers at their job on a daily basis. Only some of this work is arbitrarily deemed "computing." Secondly, of these "computing" jobs, none of them have ever actually professionalized.