Re: def'n of "computer network"

["James A. Donald" <jamesd@echeque.com>]

At 04:25 PM 7/15/95 +0100, Andrew Spring wrote:
> If your attacker has a more sophisticated statistical model of noise
> distributions than you do, then he can deduce the existence of message.

Since each hardware scanning device, and each image source, has 
idiosyncratic forms of noise, it is much harder to detect 
unusual forms of noise, than it is to emulate a usual form
of noise.

The attacker will get a huge number of false positives.  He
will not know if there is a whole lot of stego going on, or
he needs to adjust his noise models for a whole lot of
cranky and/or funky scanners.


--
  ------------------------------------------------------------------
We have the right to defend ourselves	|  http://www.jim.com/jamesd/
and our property, because of the kind	|
of animals that we are. True law	|  James A. Donald
derives from this right, not from the	|
arbitrary power of the omnipotent state.|  jamesd@echeque.com