In the other stats thread[1], users self-reported themselves as Linux 34% and Windows 22%, which I was highly doubting. There are probably a lot of closet Windows users here, or it's more l337 to claim Linux (even on an anonymous poll?). I also understand there's a selection bias related to the article's content.

I won't say polls are useless, but I keep being surprised how unreliable they are.

[1] http://news.ycombinator.com/item?id=4421971