From: gts (gts@optexinc.com)
Date: Wed Sep 18 2002 - 19:20:20 MDT
Mike Lorrey wrote:
> > In principle, a determined but apparently unpredictable sequence can
> > be
> > cracked, even if the method necessary to crack the sequence would
> > require an unrealistic expenditure of effort and resources, and/or
> > require knowledge of initial conditions that are at present
> > unavailable.
> > A sequence that is both unpredictable and undetermined would on the
> > other hand be uncrackable even in principle.
>
> The problem is that such nondeterminism makes it entirely useless for
> encryption. If you cant decrypt something (i.e. via a deterministic
> algorithm), then you are wasting your time. Adding plaintext to a
> unpredictable and undetermined keystream via some deterministic
> algorithm renders a ciphertext sequence that is non-random,
> deterministic, and is therefore as attackable as one encoded via an
> unpredictable determined keystream.
Consider, as a simple example, the generation of a password (e.g., the
password you would need for use with conventional encryption with the
IDEA algorithm [my favorite] such as that found in PGP). To generate a
password that is truly uncrackable in principle (aside from sheer brute
force attacks) would require that you have access to numbers that are at
once unpredictable and undetermined.
-gts
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 09:17:08 MST