From: yanek@novavax.nova.edu (Yanek Martinson)
Date: Tue, 24 Nov 92 23:33:02 PST
To: cypherpunks@toad.com
Subject: possibility of a reply to an anonymous message
In-Reply-To: <9211250508.AA01456@netcom.netcom.com>
Message-ID: <9211250732.AA03979@novavax.nova.edu>
MIME-Version: 1.0
Content-Type: text/plain


> (The following message arrive in my mailbox, with a request to repost
> it anonymously. Sometimes these "manual" methods work pretty well!

If the person is not on the mailing list, please forward this to him/her/it?

> lots of other ways to get this function.  But the problem is getting
> an answer back with out unmasking.  

This is not difficult at all, it just costs bandwidth.  The basic idea
is to anonymously post or mail a message, and at the end attach a
public key, and the name of a public forum such as a usenet newsgroup.
Then whoever wants to reply, will encrypt the message with the public
key, and posts to the appropriate newsgroup.  No-one but the poster of
the original message can decrypt the reply.  since everyone receives
all the articles in the newsgroup, it is not possible to trace back who
decrypted the message.

> conjure up a better scheme, but I see the backward security to be
> simply a matter of tearing the connection (return path down) faster
> than it can be backtraced.  It would be nicer if some kind of

I would NOT rely on this under any conditions.  This could only provide 
enough security that someone who is NOT interested in you can not trace you.
Hey, for that kind of security you could use something like rot13...
Anyone that is in the least interested, will immediately investigate the
recent traffic at nodes the message went through.

> definitely work. With a large number of the forwaring nodes, and
> multi-hop paths, it would become quickly very impractical to shake
> down the originator,  So the return path could be left "up" for 

Unless someone really wants to.

> for a reasonable reply.  Better yet of cource would be a completely
> uncrackable return path that you could leave up long enough to use
> institutional advertising.  Eg: run a business from.

What I described above can work indefinitely (at least until your private
key is compromised)

> Another way to leave up long lasting connections (return paths)
> is to select one really HARD point node.  This is the one they

Again, I would not use this approach.  Such a node would become the target of
concentrated penetration efforts and would break down sooner or later, and
probably would remain operational, while logging all data.


--
Yanek Martinson    mthvax.cs.miami.edu!safe0!yanek     uunet!medexam!yanek
this address preferred -->> yanek@novavax.nova.edu <<-- this address preferred
Phone (305) 765-6300 daytime   FAX: (305) 765-6708  1321 N 65 Way/Hollywood
      (305) 963-1931 evenings       (305) 981-9812  Florida, 33024-5819