This article by Stubbornella on women in technology was a great summary of what I've seen in my programming career. I've seldom witnessed overt discrimination, there's no girlie calendars in the break-room, but we're terrible at assessing and promoting coders according to how effective and efficient they are. Instead there's a tendency to reward the qualities that Nicole lists under 'Cowboy-coders', even when these people have a negative impact on our ability to ship product.
Why does this matter? It's insanely hard to build great software, it's really tough to find strong engineers, and a system that discourages a large proportion of entrants with the potential to grow into those engineers is broken.
The point that Nicole mentions but doesn't go into depth on, is that this isn't just about women. Some of the most technically brilliant male coders I've worked with are still quietly working away with zero recognition from either their own companies or the outside world. I've always loved talking about my work, and that's driven me to overcome my innate shyness and learn to perform in public, and generally be pretty vocal with my opinions in discussions. This is vital in a profession infested with arrogant jerks, and sometimes I wonder if I've had to become one too. I try hard to "act like I'm right, but listen like I'm wrong", but that's a very fine line.
We tend to take this as a law of nature, that's how people get ahead in the programming world, but it's a sign of both poor management and a very ineffective culture. By contrast my partner Liz is a trained actuary, her job involves a decade of taking exams and some very deep math, and almost all of her old department was women. You don't have to be a pushy jerk to get noticed as an actuary, there's actually a system for actively finding and rewarding talent, and as if by magic, there's a lot more women in the profession.
Our problem is that we're fixated on a romantic image of programming, one full of individual rock-stars producing astonishing products thanks to all-night coding sessions. The problem is this doesn't work, at least not consistently or reliably. The reality of building great software is that it's a long process that takes a lot of teamwork. Our immature reverence for aggressive self-promoters leads to poor outcomes, and even if you don't give a fig about discrimination, that's still a massive problem.
So how do we fix it? I'm actually still a bit ambivalent about Google's scholarship specifically for women, though I can see why it makes sense in a lot of ways. I'd much rather see companies and individuals in leadership positions actively seeking out strong engineers to speak and attend conferences, rather than waiting for people to come to their attention, since that rewards loud-mouth folks like me. We need to find more people who are quietly doing great work, without waiting for them to stick up their hand.