> Message: 7
> Date: Thu, 28 Mar 2002 22:41:31 -0700
> From: "John (EBo) David" <ebo@leml.la.asu.edu>
> Subject: Re: Random numbers
> To: plug-discuss@lists.plug.phoenix.az.us
> Reply-To: plug-discuss@lists.plug.phoenix.az.us
>
> KevinO wrote:
> >
> > Craig S. wrote:
> > a completely random (ok this is impossible in practice ...
> >
> > You CAN get truly random numbers but it requires a little specialized hardware
> > to do it. Make your noise using an analog technique and then sample and
> > digitize it. With a little care and not much money it is very possible.
>
> there was a scientific amirican artical about 10 years ago on building a
> chaos white noise random generator usinging a diode operated in it's
> unstable region and sampled...  They got the thing to sit there and
> wonder all over the place.  Appearently the dude that made it was able
> to get the built for just a couple of dollars at radio shack.  You would
> likely be able to build it with an old junk radio or TV ;-)
>
>   EBo --
>
> --__--__--
>
>

ok, here is what I understand the definition of randomness to be. It is a arbitrary
selection of value between two magnitudes. Now I visualize that varying value as
entropy in a bigger picture (after all even though we have a upper and lower bound
for the selection of value infinity is much much larger than either bound.) As n in
the magnitude tends towards positive and negative infinity this entropy starts to
become a neglible factor and we can interpolate values within the bound with a high
degree of precision. Now in the physical world we don't have a infinite number of
circuits to crunch numbers. So the trick becomes bounding these limits with a large
enough value to make the randomness neglible or to sample the value enough times
within the bounds to make a determination of what the value is not. Once we know
the irrelevant values a determination can be made as to what the value is.Either
way determining a new bound or sampling values takes time. So what we want in
encryption or selection of random numbers is to make being able to determine the
original values so time consuming that the information is either useless by the
time the answer is found or the cost of determining the values exceeds what those
values are worth. That is the core of a good encryption scheme. Of course as
technologies advance so must encryption standards so that the cost of decryption
stays higher than the value of the data.

I havew tried to cover about 4 pretty deep subjects in two paragraphs. Coupled with
the knowledge that I am not a encryption expert (just a knave in the trade) there
is much that can be expanded on. But it is at least an insight as to why I made the
statment I did about random number selection.

Craig S.