thegeezer <[email protected]> schrieb: > On 06/26/2014 11:07 PM, Kai Krakow wrote: >> >> It is worth noting that my approach has the tendency of generating random >> characters in sequence. > > sorry but had to share this http://dilbert.com/strips/comic/2001-10-25/
:-) I'm no mathematician, but well, I think the swapping approach fixes it. What this makes clear, however, is that randomness on its own does not completely ensure unpredictable items if combined with other functions. One has to carefully think about it. I, for myself, would always stay away from using modulo to clip the random numbers. It will always create bias. My first idea introduced predictable followers (you always knew that the next char had a specific probability related to the tail length of the list). You can actually learn from Dilbert comics. ;-) -- Replies to list only preferred.
