From: Hal <hfinney@shell.portal.com>
Date: Sat, 26 Mar 94 09:53:47 PST
To: cypherpunks@toad.com
Subject: Re: Digital Cash
Message-ID: <199403261837.KAA28005@jobe.shell.portal.com>
MIME-Version: 1.0
Content-Type: text/plain


I think there are two issues here.  One is the intractability of defeating
encryption protocols such as RSA, digital signatures, blinded signatures,
etc.  These form the basis for digital cash and they appear to be quite
secure.

The other issue, which I know less about, is the possibility of cryptograph-
ically strong obfuscated code.  Mike Duvos first mentioned this.  You could
have an algorithm running on your own computer and have it be impossible to
determine what it is doing, or (presumably) to effectively alter the internals
of the algorithm.

This seems a lot more difficult to achieve, since all the information needed
to tell what the program is doing is in principle in your hands.  Yet the
ability to actually determine this is computationally out of reach.  It's
not just a matter of the kinds of complexity and obscurity we have been
discussing here (self-decrypting code and such tricks), but rather some
mathematically strong transformation has been done on the structure of the
code to hide it in a cryptographically strong way.

I vaguely recall hearing about such technologies, but I can't remember
where now.  Can anyone provide some references, or (better) a summary of
how this works and what can actually be accomplished along these lines?

Thanks -
Hal