The question in, er, question relates to an issue I had been having myself with the behaviour of lexical scoping in Python, but had been able to work around sufficiently easily to not devote time to finding a proper solution. This in itself is an unfortunate truth of work; there isn’t enough time to investigate every interesting problem thoroughly. StackOverflow might just turn out to help with that niche: I presume that over time it’s going to evolve into a Not-So-Frequently-Asked-But-Still-Interesting-Questions repository. I’ve put some effort myself into answering some questions about basic computer science and distributed systems. The reputation farming I can do without, the badges are quite a neat feature – they encourage participation passively. The problem is that there really is a blind-leading-the-blind feel to some question answers, especially in the areas of data structures and the ever popular and yet heavily abused ‘big-O’ notation. If people speak authoritatively enough then they will be taken as authoritative, garner reputation points which serve as a feedback loop, amplifying their authority on future occasions. Unfortunately, there seem to be more people willing to upvote correct sounding answers than those who know whether an answer is actually correct. Time will tell if that is a general truth or just an early adoption issue. I suspect, alas, that it might be the latter.