View Single Post
Old 01-14-2013, 06:46 PM
palmipede palmipede is offline
Join Date: Jan 2013
Posts: 13
Default Re: Computing Pr(f(x) != g(x))

The distinction between exact and stochastic methods is not air tight because exact methods often have to take care of round-off errors and these may accumulate to noticeable quantities.

Complexity is the main issue in my opinion, the more complex (whether the code or the data) the more likely an implementation is buggy and the longer it takes to make it correct. I got shot for that reason just today.
Reply With Quote