From: Noah Horton (nhorton@ectropic.org)
Date: Thu Sep 12 2002 - 10:25:56 MDT
I did not come in at the start of this thread, so excuse me if this has
been mentioned before, but I have to disagree with a point in this email
regarding the non-existence of a natural process with perfect entropy.
Most quantum activity is perfectly random, and that can be exploited.
An excellent example is photonic polarization. If a photon is transmitted
with 0 degree polarization, then when it hits a polarized filter with 45
degrees of polarization, it has a perfect 50% chance of transmission
through the filter. Thus if one uses a highly tuned laser to emit single
photons, and you pass these photons through the appropriate polarized
filter, and then place a photon detector on the other side of the filter,
then you can count received photons as a 1, deflected photons as a 0, and
thereby generate a string of arbitrary length of perfectly random binary
bits. Unfortunately that setup is rather costly, but it can be done.
For the record, this is essentially how Quantum Key Distribution for
cryptography is done.
-Noah Horton
On Thu, 12 Sep 2002, Eugen Leitl wrote:
> On Thu, 12 Sep 2002, gts wrote:
>
> > "Almost never"? You'll have to prove that to me.
>
> Run an FFT on a noisy microphone. Doesn't look very white, does it?
> Here's a true RNG which doesn't pass tests for randomness. It has
> (distinctly) less than 100% entropy. I can't think of a single natural
> source with 100% entropy. Simple device artifacts will introduce bias.
>
> Run a noisy microphone through a SHA-1, then through FFT. Looks completely
> white now, doesn't it?. Run it through a suite of randomness tests. It
> will pass all of them. Now you've got a true RNG which passes tests for
> randomness.
>
> Use a PRNG based on a cryptographycally strong block cypher. It still
> passes all tests for randomness with flying colors. But it's a PRNG, not
> an RNG this time.
>
> > > I don't expect you to get this time either, but this merely for the
> > > benefit of all those you're trying to confuse.
> >
> > We might actually enjoy our communications if you were not so rude.
>
> Okay, I'm sorry if I'm giving you a hard time. It's just I'm getting
> desperate trying to communicate with you. I consistently fail to connect.
> I have no idea why this happens.
>
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 09:16:59 MST