Where are the Women and Minority Open Source Programmers?
Open source culture—in theory and largely in practice—is about as meritocratic as can be. Yet it's also nearly as dominated by white males as can be. Why is that? It's a question worth asking, especially in the wake of the Washington Post's observations a few days ago regarding Silicon Valley's "diversity problem."
Open source culture—in theory and largely in practice—is about as meritocratic as can be. Yet it’s also nearly as dominated by white males as can be. Why is that? It’s a question worth asking, especially in the wake of the Washington Post’s observations a few days ago regarding Silicon Valley’s “diversity problem.”
That women and minorities are underrepresented in the IT professions is not news, of course. But what makes this issue so interesting when viewed from the open source segment of the IT ecosystem is that many of the factors that would explain that underrepresentation in other contexts don’t apply for open source programmers.
That’s because the way open source works makes it much more difficult to discriminate against someone based on race or gender, since many open source programmers never see each other or their users in real life. When Linus Torvalds decides whether to accept someone else’s code contributions for the Linux kernel, for example, he usually has no idea what the person looks like. If he rejects the code, it’s almost certainly on technical grounds, not because he doesn’t like the gender or race of the person who wrote it.
(I’m using Linux and Torvalds here purely as an example, by the way, not because they’re any more or less relevant than other big-name open source projects and programmers to the question of diversity.)
After all, the principle that “hackers should be judged by their hacking, not bogus criteria” such as race, gender, age, professional status or nationality, has formed a core part of the “hacker ethic” since Steven Levy first described it in a comprehensive way in his 1984 book, “Hackers.”
This characteristic sets the open source space apart from other realms of the tech world, where it is easier to discriminate against people via traditional mechanisms. At major Silicon Valley companies, recruiters could favor certain groups when hiring, or managers could treat employees differently based on arbitrary characteristics.
To be sure, the same things can and probably do happen to a limited extent in the open source space, since it has its own companies and organizations with traditional hiring and management practices. Yet, in contrast to the world of closed-source software, a great deal of collaboration in the open source community takes place between individuals who deal purely in code, not personalities.
Despite all this, the few available measures suggest that female and minority programmers are even more underrepresented in the open source space than they are in the tech world in general.
Where are the Minority Hackers?
One unenlightened way to interpret this evidence would be to say that white males account for a disproportionate part of the open source community because they are better programmers. If open source is so radically meritocratic, after all, can’t we conclude that the people who end up in charge are the ones who deserve the positions most?
No. There’s another critical factor to consider, which is that open source, unlike the rest of the tech world, has traditionally rewarded participants in non-monetary ways. Probably the best articulation of this principle appeared in Eric S. Raymond’s essay “The Cathedral and the Bazaar,” in which he wrote that open source operates as a “gift culture” in which people write code in exchange for the reputation it gains them among their peers, rather than for money.
That’s less true today than it was when Raymond was writing in the late 1990s. Since then, the explosive commercial importance of open source software has made it much easier to earn a good salary writing it. Still, the fact remains that many open source coders today receive little financial reward for their work—especially if they are contributing code voluntarily over the Internet, which is precisely the way that obscures personal characteristics and ostensibly assures meritocracy.
The problem with the gift-culture principle is that it disadvantages people who can’t afford to spend their time giving away their work for free. And since women and minorities are, on average, less well-off than white men, the plausible conclusion is that white men are better positioned to participate in a gift-based community like the open source world than their counterparts.
In addition, poorer individuals tend to have less access to computers and the Internet, not to mention computer-science degrees. Those factors also make it harder for economically disadvantaged demographics to contribute open source code.
Of course, the fact that leading open source figures have occasionally exhibited less-than-welcoming attitudes toward “diverse” groups has not helped any of this. For example, in one of his more obscure writings, Raymond compared female computer scientists to “amazons,” “bimbos” and “impossible anonymous synthetic blondes in an upscale skin magazine,” imagery he apparently found “much less threatening” than embracing the reality that there could be highly competent, professional women programmers.
More recently, Torvalds, with characteristic brashness, stated publicly that “the most important part of open source is that people are allowed to do what they are good at” and “all that [diversity] stuff is just details and not really important.”
Incidents such as these are much less important than the consequences of gift culture in explaining why the open source community looks the way it does, but they’re certainly not helping to make it any more diverse.
Unix and Racism?
And if you want to get really philosophical, read this essay by Tara McPherson, an academic who argued that the modular design of Unix itself reflected “the way in which the emerging logics of … the covert racism of color blindness [were] ported into our computational systems.”
She meant that the core Unix design principle—that the operating system should be composed of discrete parts, each one designed to do a specific job and do it well—echoed the sort of silent segregation that emerged in American society once government-approved segregation became illegal in the 1950s and 1960s. Although Unix’s designers tended to be leftists who sympathized with the Civil Rights movement, she wrote, they unwittingly and subconsciously embedded segregationist principles into Unix, the operating system that remains conceptually central to the open source world today.
It’s a deep argument that only the academically inclined will fully appreciate, but it’s a fascinating one all the same.