A couple of years ago, Robin Hauser Reynolds, a filmmaker and photographer in the Bay Area, learned that her daughter, who had been taking computer-science classes, had decided that she wasn’t cut out to pursue computer science as a career. In one particular class, in which there were only a few female students, she felt that she didn’t fit in. She also perceived herself to be doing poorly, despite getting decent grades. “She called home a couple of times and said, ‘Hey, Mom, I’m so bad at this, this is horrible, I hate it.’ And meanwhile I’d seen a bunch of newspaper articles that said, ‘Hey, if you want a job out of college, you should study computer science,’” Reynolds recalled. She began seriously contemplating a question that has occupied Silicon Valley executives for the past couple of years: Why aren’t there more female programmers in the U.S., and what can be done about it?
The result of Reynolds’s inquiries was screened at the Tribeca Film Festival on Sunday, with the première of “Code: Debugging the Gender Gap,” a documentary that aims to make sense of the dearth of women in computer science. “Code” has already received disproportionate amount of attention for a documentary by a relatively unknown filmmaker; Reynolds and her film, which was financed partly through a crowdfunding campaign, had been profiled in a number of major publications well before the première, reflecting the broad interest in the tech industry’s diversity problem. Last year, several Silicon Valley companies acknowledged, for the first time, just how few women they employed in tech positions (fewer than twenty per cent, in most cases). In January, Intel pledged to spend three hundred million dollars, over five years, to make its workforce more diverse, and in February, a discrimination lawsuit brought against the venture-capital firm Kleiner Perkins Caufield & Byers by a former employee, Ellen Pao, went to trial, revealing sexist and gendered attitudes on the part of some of the firm’s most prominent executives.
To those who have been following the discussion, some of the ground that the film treads will seem well worn. “Code” describes how women engineers, including Ada Lovelace and Grace Hopper, were influential in the early years of computer science, but have become footnotes in an official history that privileges the work of male engineers. It documents the rise in the proportion of computer-science graduates who were women through the mid-eighties, and attempts to explain the precipitous decline that has taken place since then, from nearly forty per cent to less than twenty per cent. (Some of the film’s interviewees propose that the prototypical image of the antisocial, uncool male nerd that emerged in pop culture during the eighties might have discouraged women from pursuing computer-science degrees.) And it highlights some of the ongoing attempts to change the proportions: a conference named after Hopper; a curriculum change at Harvey Mudd College; a number of nonprofits recruiting female students, from diverse backgrounds, to learn to program.
“Code” also addresses a question that has been discussed less often. When Reynolds described the film’s theme to her mother, her mother asked, “Well, Robin, why does it matter who’s coding as long as we have the products?” It’s a valid question: If women don’t want to program, what’s the harm? Reynolds told me that it led her to seek out, in her interviews, cases in which less diverse engineering teams created worse products than they otherwise might have. “I said, ‘Can you give me an example of where not having a diverse coding team has affected the product?’” she recalled.
The results of those inquiries are fascinating. Roz Ho, the senior vice-president of engineering at Ericsson and a former Microsoft executive, recalled a period in the late nineties when Microsoft was adding animated assistant characters—remember Clippy?—to its Office products. Ho recalls,
In the end, Ho said, Microsoft ended up shipping Office with mostly male-looking characters; as is well documented, Office Assistant flopped (though it’s unclear what role the characters’ perceived gender played), and Microsoft eventually discontinued it. Ho’s recollection, on which Microsoft declined to comment, is one of several such examples in the film. It also describes the early development of airbags, which had been designed to protect average-sized male bodies and ended up being dangerous for small women, and of voice-recognition software that couldn’t “hear” the voices of women.
Silicon Valley’s lack of diversity can also influence entrepreneurs’ perspectives on the kinds of products that are needed in the first place. George Packer, writing in the magazine, in 2013, about Silicon Valley’s unusual brand of idealism, recounted a conversation with a young entrepreneur named Dave Morin. “San Francisco is a place where we can go downstairs and get in an Uber and go to dinner at a place that I got a restaurant reservation for halfway there,” Morin told Packer. “And, if not, we could go to my place, and on the way there I could order takeout food from my favorite restaurant on Postmates”—a delivery service—“and a bike messenger will go and pick it up for me.” In the article, Packer observed, “It suddenly occurred to me that the hottest tech start-ups are solving all the problems of being twenty years old, with cash on hand, because that’s who thinks them up.”
People who are working to make Silicon Valley more diverse, especially those who themselves come from underrepresented communities, have been keenly aware for some time of how programmers’ own experiences influence their perspectives on the kind of software that is needed in the world. Code for Progress, a nonprofit in Washington, D.C., that is highlighted in Reynolds’s film, recruits women and people of color as fellows, trains them to code, and helps to place them in engineering positions at companies and nonprofits. The organization encourages the fellows to build software programs that address the needs of underrepresented communities, including their own. One Code for Progress project, a Web site called Buscando, helps people and organizations keep track of social services and other resources for Central American refugee children; fellows have also reported working on tools to help public-housing tenants report problems and to encourage, educate, and connect low-wage workers. Programmers, whatever their gender, race, or socioeconomic background, “need to be able to think about a broader perspective and a broader user base,” Reynolds told me. “But the truth is, we’re all human, and most of us are comfortable with what we know.”
