A proper law
Brian Beesley
BJ.Beesley at ulster.ac.uk
Thu, 6 Mar 2003 17:10:46 +0000
On Wednesday 05 March 2003 21:29, Brian Morrison wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On Wed, 05 Mar 2003 21:16:37 -0000 in 3E666935.1048.B050DC@localhost
>
> "David Hansen" <davidh@spidacom.co.uk> wrote:
> > On 5 Mar 2003 at 19:23, Brian Morrison wrote:
> > > Current methods involve using white noise generated by a
> > > semiconductor device such as a noise diode being turned back into
> > > bits and then processed to ensure randomness
> >
> > One is relyng on the latter process being able to spot non-
> > randomness.
> >
> > I wouldn't trust too much to such a process.
>
> There are special device drivers for just this process, I think they can
> be made provably random, or at least demonstrably so.
>
> Perhaps someone with some real knowledge of this area could comment? As
> usual preventing physical access to the hardware is the critical step
> and not many of us can achieve that.
What's "provably random"?
AFAIK the one trillion known decimal digits of pi pass every known
mathematical test of randomness. Yet pi is very far from a random number...
Brian Beesley