Some of these tests might benefit from choosing values from a highly skewed distribution. Gamma(0.01, 0.01), for instance will exercise small and large values much better than an exponential distribution.
On Sat, Aug 28, 2010 at 4:49 PM, Luc Maisonobe <luc.maison...@free.fr>wrote: > We could start by a random set of values just like the current tests do. > The current test structure is basically one big accuracy test for each > function, consisting in 10000 calls to random values obtained thanks to > Math.random and normalized to the appropriate domain of the function > (-inf; +inf) or (-1; +1) or (0; +inf) ... >
