Posted May 11, 2006 22:18 UTC (Thu) by nix (subscriber, #2304)
[Link]
Wikipedia says this is due to Robert Coveyou.
Myself I prefer Knuth's line (from the introduction to TAOCP chapter 3) that (in italics) `random numbers should not be generated with a method chosen at random'. He then goes on to provide several examples of awful PRNGs, such as the original rand(), and shows why they're bad...
(... and a few pages later, whose name pops up, but Robert Coveyou!)