From: mpd@netcom.com (Mike Duvos)
Date: Wed, 31 Jan 1996 07:42:07 +0800
To: cypherpunks@toad.com
Subject: Re: Alleged RC2
In-Reply-To: <9601301829.AA11121@alpha>
Message-ID: <199601301943.LAA05677@netcom11.netcom.com>
MIME-Version: 1.0
Content-Type: text/plain


Mike McNally writes:

> Any ideas on whether the comment in the source about the "effective
> key length" trick being an export control deal is true?

It sounds plausable.

> If there were a known version of this floating around known to have a
> 40-bit restriction, is it likely that the restriction would be done by
> always supplying "40" as the "bits" parameter, or would be it by
> simply limiting the user key length?

The "bits" parameter guarantees that there are exactly 2^bits
distinct possibilities for the key schedule.  It does this by
re-calculating the key schedule as a function only of its
rightmost "bits" bits, after expansion of the user key to 
128 bytes.

One would not wish to directly limit the length of the user
key, since it would most likely be a passphrase of some sort. 

The "bits" parameter allows the effective key length to be
set in a manner which is translucent to the application and
its user interface. 

--
     Mike Duvos         $    PGP 2.6 Public Key available     $
     mpd@netcom.com     $    via Finger.                      $