Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watmath!clyde!rutgers!seismo!nbires!vianet!devine
From: devine@vianet.UUCP
Newsgroups: sci.crypt
Subject: Re: generating RSA keys
Message-ID: <194@vianet.UUCP>
Date: Mon, 8-Jun-87 21:09:27 EDT
Article-I.D.: vianet.194
Posted: Mon Jun  8 21:09:27 1987
Date-Received: Sat, 13-Jun-87 02:07:43 EDT
References: <537@ucdavis.UUCP>
Organization: Western Digital, Boulder Tech Ctr
Lines: 24
Keywords: Keys, RSA

In article <537@ucdavis.UUCP>, uucp@ucdavis.UUCP (uucp) writes:
> I have been trying to come up with a scheme to implement an algorithm
> that would allow the user to pick any key (presumably text), and map it 
> to a _good_ key (as defined above).

  As I understand your proposal, the entire strength lies in the selection
of a key by the user and in keeping the algorithm secret.  It is unlikely
that the algorithm can be kept secret very long in a hostile environment.
So, the proposal seems to boil down to a sophisticated one-way password
encryption.  A brute force attack might guess some of the passwords.

> My feeling is that if the key text that is gotten from the user is long
> enough (say 10 characters, 6 bits per char -- around 60 unique bits
> or from 1 to 1E19), then the results should be fairly good.

  If you are using text, there will not be 6 random bits per char (on a
CDC machine?).  A reasonable guessing strategy would be to look at only
certain ranges of characters (lower-case letters for instance).  That
yields a much lower range.

  A good paper that describes how to select primes for RSA's algorithm
is by John Gordon of Cybermation Ltd in England entitled "Strong Primes
are Easy to Find".  He calculated a 76 decimal digit prime in about 20
minutes on a 1 MHz Apple II.