NetNews Usenet Archive 1992 #27/NN_1992_27.iso/spool/sci/crypt/5063

Text File | 1992-11-20 | 1.7 KB | 45 lines

Newsgroups: sci.crypt
Path: sparky!uunet!pipex!pavo.csi.cam.ac.uk!camcus!nmm1
From: [e-mail] (Nick Maclaren)
Subject: Re: unpredictable random generators (terminology)
Message-ID: <[e-mail]>
Sender: [e-mail] (USENET news)
Nntp-Posting-Host: bootes.cus.cam.ac.uk
Organization: U of Cambridge, England
References: <[e-mail]>
Date: Fri, 20 Nov 1992 13:23:10 GMT
Lines: 32

In article <[e-mail]>, [e-mail] (Prof. Dr. Klaus Pommerening) writes:
|> In the literature there is a certain confusion as to how call this sort
|> of (unpredictable) pseudo-random generators. Examples are:
|>  - secure
|>  - unpredictable
|>  - cryptographically strong
|>  - perfect
|> I prefer the latter term, because it is so short, and, more
|> important, the cryptographic strength implies that the generator
|> passes all efficient statistical tests -- truly perfect, isn't it?

I am afraid that this isn't true.  Consider the following generators:

    1) A multiplicative congruential generator with known modulus of the
order of 10^1000 and a known multiplier that is near-optimal in the spectral
test, sense.  This would be excellent for simulation (if fast enough!) but
useless for serious cryptographic work.

    2) An effectively independent random number generator that delivers
numbers with an unknown relative error in their uniformity of O(10^-3).  This
would be excellent for most cryptographic use, and is useless for most
simulation.



Nick Maclaren
University of Cambridge Computer Laboratory,
New Museums Site, Pembroke Street,
Cambridge CB2 3QG, England.
Email:  [e-mail]
Tel.:   [phone]
Fax:    [phone]