News for hackers. Stuff that matters.
Re:Time Spent Waiting for an Answer in StackOverflow (by lan... by tomerv on Wednesday October 31, 2012 @01:12AM (#4721331)
By filtering out answers that arrive after more than 5 hours, you're really skewing the data, and in a very arbitrary way. Why not 4 hours, or 6? Note that filtering also caused languages to change relative positions (e.g. PHP climbed from 5th place to 1st).
Re:Time Spent Waiting for an Answer in StackOverflow (by lan... by zzleeper on Wednesday October 31, 2012 @01:26AM (#4721365)
Well he could always compute the median. Less robust to outliers. 100 questions answered in 1 minute, with one question answered in 1 day will give you an avg of 15 minutes instead of the more representative 1 minute
Re:Time Spent Waiting for an Answer in StackOverflow (by lan... by guilespi on Wednesday October 31, 2012 @01:24AM (#4721362)
It does, the 'clustering by hand' in categories according to hardness is completely unscientific.

I guess if done properly more categories would arise (e.g. bad formed questions) and possibly trends would change.

It's fun none the less ;)