In other words, instead of ending discrimination, we should just pile more layers of it on top -- rather than paint your white car pink, you'll load your white car into a pink 18-wheeler and just drive that to work, proudly proclaiming that you have diversity in your color choices (and then one day, someone will point out that you don't have a rainbow-colored ride, so you'll load your pink truck on a rainbow-colored train and demand that someone lay tracks from your house to your office).
I don't think ending discrimination is a simple thing and that in this case increasing the visibility of female experts in an area can have nontrivial antidiscriminatory effects (e.g. encouraging young women to go into an area, making women feel more comfortable at a conference, etc.) even if it involves positive discrimination.
Anyway, I'd still be interested to hear from you how (following your metaphor) you would suggest to dismantle the white car directly.
"in this case increasing the visibility of female experts in an area can have nontrivial antidiscriminatory effects (e.g. encouraging young women to go into an area, making women feel more comfortable at a conference, etc.) even if it involves positive discrimination."
It may make women feel more comfortable at conferences, but there is a much deeper problem here. Discrimination is discrimination, regardless of whether or not it affects a disadvantaged group in a positive way. There is also the risk that such discrimination will only deepen the problems that caused the disproportionate representation in the first place: people might conclude that a woman's success in a technical field was the result of the advantages she was given (e.g. by being given an unfair boost in being invited to speak at conferences), and in the worst case this will reinforce notions that women are not as good as men. I have heard the last concern from women in computer science and technical fields -- women who worked hard to get where they are, who now worry that other people will doubt their abilities because of this sort of thing.
"Anyway, I'd still be interested to hear from you how (following your metaphor) you would suggest to dismantle the white car directly."
The car is not the problem; the paint on the car is the problem. While it is more expensive in the short-term to repaint the car, in the long run you realize savings (less fuel than the truck, no need to lay down rails, etc.). The point here is that the problem can be attacked head-on: the problem with female representation at conferences is a reflection of the problem with female representation in the field, which may be harder to solve but whose solution will ultimately save effort and pain in the long run. Why waste resources on solutions that make the situation even more complicated, when we could spend resources on solutions that simplify the situation?
Yes, it is hard, but there are people working on it. The woman who organized this conference should have worked with those people. I see efforts on the part of the professors in my department to bring more women into the field by focusing on the 101 course. It is not clear which approach works best; some professors have tried making the introductory course more graphics-oriented, theorizing that young women would be drawn to something more visual; others have tried to make the intro course more relevant to fields where women dominate, hoping to make the material more comfortable or familiar to potential dual-majors. I know one professor whose idea was to make the course more like a math course, with intellectually challenging problems and with the idea that high-school boys are more likely to have spent their time hacking, and thus will have an advantage in any course that is "hacker oriented" (and so making a theoretically rigorous course would help level the field).
There are probably other approaches that could potentially work better, but the point is this: the root of the problem is in the lack of women choosing to apply to or enroll in engineering and CS programs. I saw this problem first-hand as an undergrad (there were no women in my EE class; the department simply had no applicants at all), and I see it now in a CS grad program. This is not a problem that can be solved by layering more discrimination on top of it; instead, we need to remove the discrimination at the deeper levels. If women are being turned off the CS by 101 courses, or if they are not even bothering to take 101 courses or even apply to schools with CS programs, that is the problem that must be solved. I suspect that the ultimate solution will be found in middle school, because that is the age where girls with talent for math and science seem to lose interest in technical fields.
Thanks for taking the time to write such a detailed answer. We might actually disagree less than I thought.
However, with your main point I still disagree. If I understand you correctly, you say that the problem can only be fixed bottom up: that we can't fix equal representation in tech by having some conferences with equal representation. I don't think it is that clear cut: for example, having an event with a disproportionately large amount of (1/2) female expert speakers might allow more women at an earlier stage of their career (depending on the event, maybe even high school) to picture themselves on such a career path.
So, within your metaphor, the white car might come out a little pinker after taking a ride in the pink truck...
Apart from that, I completely agree with you and women feeling restricted to other career paths much, much earlier is an important problem to solve. Even your example, the CS 101 class, is probably a later stage in that progress.