I think these sorts of articles are asking the wrong question when they ask "why aren't there more women in CS?" or "how can we get more women in CS?"
Instead the question should be (IMHO) "are there any barriers to entry for women in CS?"
I suspect the opinion people have on this issue largely mirrors their view on affirmative action. Some believe in a meritocracy that is blind to race, age and gender. They believe that any discrimination for one group is discrimination against another (which, incidentally, is axiomatically true). Many in this camp believe that lowering standards for one group encourages a perception that any member of that group did less to get where they are (correct or not).
The other side believe that an artificial environment needs to be created to correct an imbalance. Rooted in this principle there is often a belief that this situation only exists because of artificially created and perpetuated gender roles.
Personally I believe that gender imbalance in this (and other) professions is only indirectly related to gender.
In this post as one data point, the author notes that she felt socially isolated from other girls. The question then is: is this because she didn't share the same interests as her peers or did she develop separate interests because of this social isolation? I suspect the answer is a little of both.
The stereotype of an engineer is that of a judgmental introvert, two traits that tend to negatively correlate with having many social connections. Of course this stereotype isn't universal but we're talking about patterns in a large group, not specific individuals. If such personality traits are causative (rather than simply being correlated) in choosing CS as a career, you then need to ask if there is a gender basis for these traits? Males and females form very different social structures when left to their own devices.
For me, I'm not sure how productive all this gender hand-wringing really is.