I almost always listen to CBC Radio One when I’m in the car. So it’s no surprise that I was doing so this Wednesday on my way to a doctor’s appointment. It was rather a long drive, which gave me a fair bit of time to listen in on an episode of Spark featuring an interview with UCLA’s Jane Margolis, author of Unlocking the Clubhouse: Women in Computing. Ms. Margolis studies gender and race inequality in computer science, a background-interest of my own.
Nora Young’s lead-in to this interview of course mentioned the dramatic imbalance of males to females who participate in computer science, but what really caught my ear was that her interest in talking to Ms. Margolis was to find out whether or not this imbalance matters, and why. This is a question that’s always seemed really relevant to me: the imbalance is striking, but should we care? I hoped it would take me long enough to drive to the doctor’s office to find out.
Typically, I hear only a few justifications for concern over the dearth of women in computer science, the most convincing and concerning (to me) being that there may be plenty of young ladies who’d love to become involved, but feel as though the field is a boys’ club, that it is too intimidating, or that it is prohibitive in some other way. If it’s the case that there are budding young nerds out there who feel as though their dreams are being denied, clearly this needs to change (whether said nerds are male or female). More often, though, the reason I hear is that diversity in the field is important – that women bring different perspectives and skill sets to the table than men. While this may be true, it hardly seems so earth-shatteringly important that we rouse women from their preferred paths to draw them into a field in which they may have little or no interest.
Margolis, however, takes this latter justification a step further. She contends that a lack of gender diversity in the field may actually be detrimental to women. The analogy drawn is to an all-male design team who created early models of air bags that were life-saving for males, but fatal for females. To bear out the analogy, technologies created with men in mind may not be the right fit for women. This is a fair concern, but I’m still not convinced it’s the golden goose.
Consider this, would the involvement of women in the above scenario necessarily have solved the air bag problem? The only conclusion I can reasonably reach is ‘maybe’. The reason for my lack of conviction is that female involvement would only have prevented these problems if the women involved performed better requirements gathering than the males. In other words, the mere presence of a woman may not have cued the team to the fact that something modelled only with the average man in mind might actually be harmful to a woman. Only if the women (and men) considered the specific needs of the end user would the outcome have been changed.
Customer requirements gathering is a critical part of computer science. Programmers like to create. And we like to write code. Sometimes, we really like what we’ve made. This doesn’t necessarily translate to customer satisfaction (in fact, the gap here is present so often that many methodologies call for nearly constant customer involvement to avoid repeatedly producing the ‘wrong’ product). As such, it’s really important for designers and developers to consult with end users about what software must do for them. A general comparison can be drawn between customer requirements gathering and considering someone for whom you are purchasing a gift: you should always buy the gift the recipient wants, not the one you want (there is a tendency to get this backwards despite the painful simplicity of this rule).
The analogous activity in the case of the air bags would have been gathering requirements for all possible end users of said air bags. Is it true that women are the only ones who can effectively perform customer requirements gathering with respect to female end users? Certainly not. Requirements gathering is about considering all potential end users, and if possible, interacting with them to discover their expectations and needs. To be quite blunt, any dummy could have realized that men wouldn’t be the only ones getting into car accidents that might deploy an airbag. Similarly, there’s no inherent reason that a group of male software developers can’t understand that women are amongst the end users of their products.
Admittedly, customer requirements gathering is tricky business – there are times when customers are too busy, don’t know or can’t communicate specifically what they want, or are unavailable to the software team altogether. Granted, in situations where communication with these end users is impossible, it might be helpful to employ a diverse group of software developers who can substitute their own needs for those of the absent end user. But, when the breadth of software products is considered, it seems unlikely that any given software-development team is ever going to be diverse enough to cover all possible cases. For example, while a team might have a perfect balance of men and women on the team, their software might be intended to help farmers sell grain at the best possible price in the area where they reside. Are any of the developers farmers? Nope, but dammit, some of them are ladies!
To sum up, ladies are not the only software developers who can develop software for lady software users. So, although I enjoyed the interview with Jane Margolis (she had plenty of excellent observations on how male and female computer scientists tend to differ, as well as great arguments for multi-disciplinary teams), and although I’m stoked to read her book, I’m still not convinced that the perceived underrepresentation of females in computer science necessarily matters.
If you’d like to listen to Nora’s interview with Jane, here’s the podcast. 33:27 marks the spot.