I am not sure if that's a bug or a feature or something else, but I thought I'd report it here. There is this question http://mathoverflow.net/questions/90807/computer-invariant-closed which I originally voted to close since it did not seem to have any mathematical content to it. Then, the OP replied to a Gerry Myerson's comment to the question, and the question suddenly started looking more meaningful, though unacceptable in that form. I however started typing an answer to it, and when I hit the button to post the answer, the most delightful result occured - the question now is marked as closed 17 minutes ago and has an answer posted 8 minutes ago. I don't think it's bad that system allows to post an answer which was being typed before the last vote to close was cast, but it's a little bit unexpected!

Actually I post comments to closed questions after the closure now and then (though so far I haven't seen a closed question which would tempt me to post a full answer). The resulting time signature is, indeed, funny but I find this feature (or bug) that allows users with sufficiently high ratings to override the "official dismissal" desirable. I'm too lazy to experiment now but I won't be surprised if above certain reputation level, you can post a full answer to an officially closed question.

Here's the question that Kevin was referring to: http://mathoverflow.net/questions/5954/is-it-best-to-run-or-walk-in-the-rain-closed and here is the related meta thread here: http://tea.mathoverflow.net/discussion/644/answer-to-a-closed-question which links to some posts on meta-SO about the matter.