Cryptography Reference
In-Depth Information
response of the computer and the computer can turn data into the
random choices that guide the grammars.
Here's an extreme example. You want to set up a conversation
The sun's a thief, and
with his great attraction
Robs the vast sea: the
moon's an arrant thief,
And her pale fire she
snatches from the sun:
The sea's a thief, whose
liquid surge resolves The
moon into salt tears. . .
—William Shakespeare
in
Timons of Athens
with a friend across the country. Ordinarily, you might use the basic
talk
protocol to set up a text-based link. Or youmight use one of the
Internet phone programs to exchange sound. In either case, the bits
you're exchanging can be monitored.
What if your talk program didn't contact the other person directly
but logget into aMUD somewhere on theNet as a persona? The other
person's talk program could do the same thing and head for the same
room. For the sake of atmosphere, let's make it a smoke-filled room
with leather chairs so overstuffed that our textual personae get lost in
them. There are overstuffed mastodons on the wall to complement
the chairs.
Instead of handing your word bits over to the other person's per-
sona directly, your talk program encodes them into something in-
nocuous like a discussion about last night's baseball game. It might
be smart enough to access the online database to get an actual score-
card to ensure that the discussion was accurate. When the other
person responded, his talk program would encode the data with a
similar grammar. The real conversation might be about very private
matters, but it might come out sounding like baseball to anyone who
happened to be eavesdropping on the wires.
Both sides of the conversation can use the same grammar. This
convention would make it possible for both sides to hold a coherent
conversation. After one persona commented about the hitting of Joe
Swatsem, the other could say something about Swatsembecause the
same grammar would control what came afterward.
The entire systemis just an automated version of the old gangster-
movie conceit about talking in code. One gangster says, “Hey, has the
shipment of tomatoes arrived yet?” The other responds, “Yeah. It'll
cost you 10,000 bananas.” The potentials are amazing.
The Disguise
Grammar-based mimicry can be quite realistic. The
only limitation is the amount of time that someone puts into
creating the grammar.
How Secure Is It?
At its best, the grammar-based systemhere can be
as hard to break as RSA. This assessment, though, doesn't mean
that you can achieve this security with the same ease as you
can with RSA. There is no strong model for what is a good key.
Nor has there been any extensive work done on breaking the
system.
