From: yanek@novavax.nova.edu (Yanek Martinson)
To: gnu@toad.com (John Gilmore)
Message Hash: f4ef0eb33c0140dd43a83e7fe043c317870b190aed48d946feb0535921ae2daa
Message ID: <9211280500.AA08808@novavax.nova.edu>
Reply To: <9211280415.AA12997@toad.com>
UTC Datetime: 1992-11-28 05:00:47 UTC
Raw Date: Fri, 27 Nov 92 21:00:47 PST
From: yanek@novavax.nova.edu (Yanek Martinson)
Date: Fri, 27 Nov 92 21:00:47 PST
To: gnu@toad.com (John Gilmore)
Subject: Alternative to physically meeting
In-Reply-To: <9211280415.AA12997@toad.com>
Message-ID: <9211280500.AA08808@novavax.nova.edu>
MIME-Version: 1.0
Content-Type: text/plain
> person identified in the name field". Don't sign someone's key unless
> you are sure you can make that statement (like, they're standing in the
> same room with you and they verify that they key ID matches their real key).
> Don't sign a key that you received by email or over a modem; it might
> be from someone impersonating your friend (when they left their keyboard
Here's an alternative method if you know the person (know them well
enough to recognize the voice on the phone):
You transfer the key over a non-trusted channel such as electronic
mail.. Then both of you run a secure hash function (for example MD5)
on the key. The result (128bits in the case of MD5) is then converted
to alphanumerics using something like base64. In the case of 128bit
hash, you end up with 22 character verification code.
Then you call each other up on the phone, and spell out the 22 letters
and verify they match what you independently computed. If they do,
that means the key transferred over e-mail is correct.
This is of course susceptible to the kind of attack where someone stands
with a gun pointed at you and makes you give the wrong key, but that
attack can also be done if meeting in person. I.e. someone tells you they
are going to kill you as soon as you step out of the room if you don't give
the compromised key.
But at least with this attack one of the persons knows they key is
no good, and you will avoid using it for sensitive material.
Can you think of any other attack that this method is susceptible to?
--
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
Return to November 1992
Return to “yanek@novavax.nova.edu (Yanek Martinson)”