Newsgroups: comp.ai.genetic
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!gatech!EU.net!Austria.EU.net!cosy.sbg.ac.at!edvz.sbg.ac.at!news
From: helmut@cosy.sbg.ac.at (Helmut Mayer)
Subject: Re: Random Number Generator Effects
Message-ID: <D9CoM3.I99@wst.edvz.sbg.ac.at>
Sender: news@wst.edvz.sbg.ac.at (USENET News System)
Reply-To: helmut@cosy.sbg.ac.at
Organization: University of Salzburg / Austria
References: <3q7tf6$ku6@ixnews3.ix.netcom.com>
Date: Mon, 29 May 1995 17:30:03 GMT
Lines: 66

Jonah Thomas writes
# In <D972ov.E4u@wst.edvz.sbg.ac.at> helmut@cosy.sbg.ac.at (Helmut Mayer) 
# writes: 

# >Yes. In classical terms the theory of statistical mechanics can be
# >transponed to electrons. That's why we speak of "electron gas" in a
# >metal.(Maxwell-Boltzmann-Distribution).
# 
# OK, but that isn't a uniform distribution is it? 

It's not, though it's very flat. (like exp(-E/kT), where k~10^-23 (Boltzmann  
constant, E is energy and T temperature)

# Aren't you
# measuring something like how long it takes for electrons to reach a
# target?

More the average time between two collissions of the electron with the cristall  
lattice they move in. This time is again very random for a single electron.

# Or how many reach the target in a given time?  Doesn't it
# change with temperature?

Yes. This again goes against your repeatability demand which I'm not quite sure  
of, if it is needed and even more, if it is useful for a PRNG. Just occured to  
me: The term Pseudo RNG implies that it is not a real RNG. What's the  
difference between a PRNG and a RNG?

# You might do something like put your
# resistors in pairs, and record which crosses some threshold first.

That's the step towards a noise generator. Usually, you can select white noise  
with such a device which as pointed out earlier has theoretically uniform  
distribution in the power spectrum. Not technically, because there is no  
perfect electronic device. But again, even the distribution will drift a  
little, not too bad for a real random process, I guess.

# For people to use a nonrepeatable rng they must have a lot of
# faith in it.

Why? They could do some experiments (and I'm sure many people have done) with  
it and then decide weather they can use it or not.

# A good rationale explaining why your design really
# _ought_ to give uniformly random results would be very useful for
# creating that faith.

If you are really interested, I'll take a closer look to noise generators.

# Everybody who decides 
# your stuff is good enough is a potential extra customer and you need 
# them for your product to be a success.  Best of both worlds.

Unfortunately, I don't run a business..;-)

Best regards,

Helmut
---
Helmut A. Mayer, University of Salzburg, Department of Computer Science
*** Member of siGis ***
Jakob-Haringer-Strasse 2, A-5020 Salzburg, AUSTRIA
Tel.: +43-662-8044-6315  Fax: +43-662-8044-611
"...All I want is PEACE and LOVE on this planet, aint how that GOD planned  
it?..." - Chuck D.

