**Subject:** Re: replace kernel random number function

**To:** Robert Elz *<kre@munnari.OZ.AU>*

**From:** Mike Pelley *<mike@pelley.com>*

**List:** tech-kern

**Date:** 10/24/2000 11:06:12
> > sys/arch/sparc/sparc/clock.c assumes that random(9) returns uniform
> > random number (at least the comment says so).
>
> Unless I am horribly confused, isn't "uniform" a necessary property
> of anything that is truly a random number? That is, if the number
> selection is biased, it isn't truly random is it? Is there something
> about the kernel's random number generation (the good one) which makes
> it not generate uniform numbers ?
The original should probably have said a uniform pseudo-random number. A
series of truly random numbers (if such a thing exists ;o) is not bound by
any constraints, including uniformity. This series can represent any
possible sequence, including those that look "biased". Even over a long
series there is often noticeable clumping of numbers close together, which
can be inconvenient if a uniform distribution was the goal.
Mike.