1. Re: integer_median, Lucius, me and that.
Tor wrote:
>I finally found a faster median function (all on my own 
, although
>it will only work with integers as it is heavily inspired by good old
>bucket sort: <snip>
Interesting concept. I played with a similar idea, also based on the
bucket sort. Then I realized, it's fundamentally flawed: the execution
time does not depend just on the population size, it also depends on
the vagaries of the data, the total spread of the values. In other
words, a relatively small sample with a huge difference in extreme
values can take longer than a large sample with minimal differences...
jiri (sorry to be a spoilsport)
