From: "Joseph Reagle Jr." <reagle@gl.umbc.edu>
Date: Sat, 29 Jan 94 07:23:30 PST
To: Yaroslav Faybishenko <fby@netcom.com>
Subject: Re: Quantum cryptography
In-Reply-To: <199401290255.SAA27845@mail.netcom.com>
Message-ID: <Pine.3.89.9401291011.A22328-0100000@umbc8.umbc.edu>
MIME-Version: 1.0
Content-Type: text/plain


On Fri, 28 Jan 1994, Yaroslav Faybishenko wrote:
>   i assume the message would be translated in blocks (instead of in a 
> stream), and then those blocks would be compared to verify that no one 
> is eavesdropping. So, if there is actually someone eavesdropping, then 
> the very first block that is transmitted would be corrupt, so when Alice 
> and Bob compare their versions of the block, they would discover that 
> the channel is insecure, so they would stop the conversation.  

	I believe here you are kind of describing the private channel 
(the quantum channel) though I've never heard of it described in quite 
this way.  When I refer to the Public Channel, I am referring to the 
channel that Alice and Bob talk on so that they may compare the blocks or 
bits that they have both received.  Hence, a powerful enough Eve could 
insert herself in the private channel (this is very much discussed) as 
well as the public channel (which I haven't seen addressed.)

	Many of the papers discuss Eve inserting herself in the 
private channel and impersonating Alice (to Bob), and Bob (to Alice), 
but this results in a non-correlation of the measurements when compared 
over the public channel.
	But if Eve may impersonate both people on _both_ channels it may 
be possible to dupe Alice and Bob.  I expect what is important is the 
definition of what a public channel is?  (Still haven't got access to my 
original thoughts on this, will say more later.)

Regards,
Joseph M. Reagle Jr.   | 
reagle@umbc.edu        | It is a good day to die! --Worf.
jreagl1@umbc8.umbc.edu |