Other Directory Sites: SeekWonder | Directory Owners Forum
The following archive was created by hippie-mail 7.98617-22 on Fri Aug 21 1998 - 17:20:12 ADT
Anonymous (nobody@replay.com)
Wed, 8 Jul 1998 00:28:02 +0200
On Tue, 7 Jul 1998, Perry E. Metzger wrote:
> My big question is this: are there tools for taking a set of random
> numbers dispersed according to a non-uniform distribution, like a
> poisson or normal distribution, and turning them into a set of random
> numbers over a uniform distribution? Given such tools, timing
> intervals between the geiger counter ticks is probably safe --
> otherwise, it may skew the results subtly.
The mathematical technique to do this is straightforward enough, although
I'm not sure of the cryptographic implications.
The inverse problem is more commonly encountered. Given a uniform
random number generator (often pseudo-random), generate numbers with
a normal or other distribution. The solution is to plot the integral
of the desired distribution curve from negative infinity to x, as a
function of x (the cumulative distribution). Call this function f(x).
Note that f(x) ranges from 0 to 1. Given a uniformly selected number
y in this range, output x such that f(x) = y. Then x will have the
desired statistics.
To invert this, then, produce the cumulative distribution curve f(x),
and then given a random value x with that statistical distribution,
output f(x). The resulting value will be uniformly distributed.
Of course you need an accurate model of the statistics of your distribution
in order to get a good uniform distribution, which is where the technique
may not work too well in practice for cryptographic purposes.
The following archive was created by hippie-mail 7.98617-22 on Fri Aug 21 1998 - 17:20:12 ADT