[Cryptography] Photon beam splitters for "true" random number generation ?

[dj at deadhat.com]

>
> Their intended market is gaming machines, about the only situation where
> the
> cost is justified, so there's little point to backdooring them.  They're
> reasonably nice devices, and easy to interface to, I've had support for
> them
> in my code for awhile.
>

The last time I paid attention to the Nevada Gaming Commission random
number generation rules, the one thing they did not require was random
numbers. They required statistical distributions that were uniform over
long time spans. I.E. It was within the rules to skew the results one way,
then the other, providing the odds remained the same in the long run.