From: cme@ellisun.sw.stratus.com (Carl Ellison)
Date: Thu, 16 Sep 93 08:49:20 PDT
To: cypherpunks@toad.com
Subject: Re:  Random Numbers
Message-ID: <9309161545.AA09335@ellisun.sw.stratus.com>
MIME-Version: 1.0
Content-Type: text/plain


>Date: Thu, 16 Sep 1993 02:38:26 -0400 (EDT)
>From: Mike Ingle <MIKEINGLE@delphi.com>
>Subject: Random Numbers
>Message-Id: <01H300QM3KUI8ZG7IN@delphi.com>
>

>That sounds doable. Tune a radio to a spot where there is static, plug it
>into the audio in port of a SoundBlaster card, sample a few Kbytes, and
>MD5 it to get your key. Nobody is going to reconstruct that.

A poster a while back on sci.crypt pointed out that if you plug nothing
in to the audio input (eg., /dev/audio on a Sun), what you get is
circuit noise.  If you then compress it, you get pretty good random input.
It's not something someone else can monitor or interfere with.
If you then pull off some bits for encryption keys and encrypt the
rest using those keys, ....

I have a small C program to do that -- do any left-over compression
and build a shell script to drive IDEA for encryption.  It's avbl by
mail if anyone wants it.  If enough want it, I'll post it on sci.crypt.

>                                     < MikeIngle@delphi.com >
[p.s.]
> I got the impression Mykotronx isn't too happy about the status
>of the Clipper/Spooktap project.  :-)

Yeah?  Do tell.  What did they say?