[Python-ideas] Should our default random number generator be secure?

[Sturla Molden]

On 14/09/15 22:45, Robert Kern wrote:

> Funny story: MT is provably *not* strictly equidistributed; it
> produces a exactly 624 fewer 0s than it does any other uint32 if you run
> it over its entire period. Not that it really matters, practically
> speaking.

I probably would not live long enough to see it ;)


Sturla