Other Directory Sites: SeekWonder | Directory Owners Forum
The following archive was created by hippie-mail 7.98617-22 on Fri Aug 21 1998 - 17:16:52 ADT
Peter Wayner (pcw@access.digex.net)
Mon, 6 Apr 1998 13:06:55 -0400
Over the weekend, I saw the new Bruce Willis film, "Mercury
Rising", which revolved around an autistic kid who could read
the government's new secret code with the name "Mercury". The
whole process works by the kid staring intently at a block of
text that, for the sake of argument, is as impenetrable as a hex
dump. (There are some greek letters in the movie's text, but I
say ignore them.)
Here's my question: For what crypto systems is this impossible
to do without a copy of the key? We know that a one-time pad can
yield any message given the right key. So the kid couldn't pull
the right message out of thin air. But how likely is it that two
plausible messages will be emerge from the same DES encrypted
message. That is, given ciphertext C produced by k1, how likely
will it be that another key, k2, exists such that both
Decrypt(C, k1) and Decrypt(C,k2) might make sense. The paper on
deniable encryption (Dwork et al) describes how to do this in
one space. How likely is it that it occurs in DES, IDEA or RSA.
It's an interesting question to imagine breaking RSA this way.
Autistics are often known to have the strange ability to simply
"see" the right answer to a complicated math problem. They just
look at a long multiplication and "see" the right result. Has
anyone done a study on whether they can be taught to "see" the
answer to intractable problems in math like the inverse
logarhythm, number factoring or some NP-complete problems like
3SAT? If they could do this, then someone could break RSA with
access to the public key.
-Peter
The following archive was created by hippie-mail 7.98617-22 on Fri Aug 21 1998 - 17:16:52 ADT