Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 (Tek) 9/28/84 based on 9/17/84; site hercules.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxt!houxm!vax135!cornell!uw-beaver!tektronix!teklds!hercules!franka
From: franka@hercules.UUCP (Frank Adrian)
Newsgroups: net.crypt
Subject: Re: Why no hardware random numbers?
Message-ID: <427@hercules.UUCP>
Date: Tue, 26-Mar-85 14:11:07 EST
Article-I.D.: hercules.427
Posted: Tue Mar 26 14:11:07 1985
Date-Received: Thu, 28-Mar-85 23:55:13 EST
References: <868@utcsri.UUCP> <2139@wateng.UUCP> <321@mtxinu.UUCP>
Reply-To: franka@hercules.UUCP (Frank Adrian)
Distribution: net
Organization: Tektronix, Beaverton OR
Lines: 20
Summary: 

In article <321@mtxinu.UUCP> ed@mtxinu.UUCP (Ed Gould) writes:
>>                      The problem with this type of random number generator
>> is that the sequence of values you get is irreproducible, which can be
>> a real drag if you are doing simulations and want to check your results,
>> or you want to try two different algorithms on the same sequence of
>> numbers.
>> -- 
>>    - peter bain
>
>It seems, then, that we really need both.  A good software PRN
>generator for testing, reproducing, and debugging, and a hardware RN
>generator (I don't say pseudo-random, because I'm willing to define
>certain natural phenomena as truely random) for better randomness.
>

Not really.  If you want reproducability on a machine with ONLY a hardware
RN generator, write a program which samples the generator, puting the results
into a file.  Then read the file sequentially for a reproducable set of random
numbers.  This gives both repeatability and "true" randomness.
				Frank Adrian