From: Bill Stewart <stewarts@ix.netcom.com>
Date: Sat, 23 Sep 95 17:30:41 PDT
To: wb8foz@nrk.com
Subject: Re: "random" number seeds vs. Netscape
Message-ID: <199509240029.RAA22318@ix8.ix.netcom.com>
MIME-Version: 1.0
Content-Type: text/plain


At 06:29 PM 9/23/95 -0400, you wrote:
>Can you take the time between two effectively async events, and use the
>low order bits of same?
>For example, time between a netverk pack arriving and the next tick of
>the TOD clock? And/or the system's disk IO or such? 

You might get some entropy there, but you can get more (though perhaps not
measurably more) by using the values of both instead.  For instance
        rand = MD5(rand, lowbits(T(clock) - T(packet)))
vs      rand = MD5(rand, T(clock), T(packet))

As long as you're not overoptimistic about how many real bits you've gotten,
might as well actually use the version that keeps more low-quality bits
than try to improve them.  Also, do watch out for things that aren't
really asynchronous, e.g. the next clock tick will probably happen at
some even multiple of 1000000/Hz microseconds.
#---
# Bill Stewart, Freelance Information Architect, stewarts@ix.netcom.com
# Phone +1-510-247-0664 Pager/Voicemail 1-408-787-1281
#---