From: Eugen Leitl (eugen@leitl.org)
Date: Thu Sep 12 2002 - 13:24:21 MDT
On Thu, 12 Sep 2002, gts wrote:
> I'm not convinced that a noisy microphone is a true RNG. By "truly
I'm not truly convinced that anything is a "true RNG". I personally find
relying on a known wrong theory (we know for a fact QM is not a TOE) in a
matters of life and death unwarranted. I wonder what makes people
contemplate using QM key distribution for satellites instead of using a
good old PKI infrastructure, or, better, a one time pad to distribute the
keys. It's basically a faith based decision. "In laws of physics we
trust". Even while we know for certain we haven't got the whole picture.
Hello? Which bus am I on?
Assuming digital physics is true, I must try to make guessing the state as
hard as possible. A large nonlinear system is a) amplifying intrinsic
initial state knowledge uncertainty b) amplifying noise it picks up from
anything in within lightcone it resides. Choosing a very small, strange
system like a single photon hitting a beam splitter makes me cringe with
unease.
Assuming digital physics is not true, we still get full QM goodness, since
now the noise floor driving the nonlinear process amplifying it is true
QM.
Clearly win/win in my book.
> random" or "genuinely random," I mean 1) unpredictable, and 2)
> undetermined. Only quantum processes satisfy the second requirement.
You'll notice that quantum noise permeating from the system floor
completely swamps all analog systems. Only digital systems manage to pull
themselves up by the hair from the quantum noise swamp by means of a dirty
trick.
> Noise from a microphone is not necessarily due entirely to quantum
> processes. Some of that noise will be determined by events that occur
> at the macroscopic non-quantum level.
You're putting arbitrary walls up where none exist. Picture a purely
classical gas box of ideally elastic perfectly spherical balls. They have
some initial position, some initial velocity distribution, can collide
with themselves and the walls. Now if you let this system run (this is
being used as an illustration of semideterministic (because hard to
quantify external factors make reproducibility a nightmare) noise in
modelling), you'll see that infinitesimal displacement of even a single
ball gets amplified exponentially over time, very quickly completely
changing the state of the system. The snapshot shows a completely
different distribution of balls in space.
Now chuck Newton out the window, enter Herr Heisenberg. You're conducting
a measurement on a system (looking at Brownian noise of a membrane in that
gas box). There's your source of QM noise (whether genuine divinely
approved 24ct noise, or the shabby counterfeit article) amplified to
macroscale, where you detect it.
> We've barely tried to communicate at all, Eugene. You entered the
> quantum tunneling discussion with a "bah humbug" post, followed shortly
> thereafter by a personal insult. Aside from those unfriendly messages
By insulting you I was trying to make you understand that you're not
listening. I don't just refer to myself, you exhibited a certain recurring
pattern when talking to other people. On online fora you don't see people
rolling up their eyes, and make an excuse by claiming their cocktail ran
out, and never coming back to resume the conversation. This place needs
different means of signalling.
> you've made little or no attempt to "connect" (until now, thank you.)
> But let's let bygones be bygones... I note that you seem to be a very
> intelligent and knowledgeable fellow and this is not supposed to be a
> over-polite tea-party in the first place. :-)
I'm glad you didn't take offense. Let's continue this conversation.
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 09:16:59 MST