Re: Basic tools of encryption: Transposition and substitution?

Doug wrote:

> Is the difference between "compression" and "encryption" just one
> of those random things where a line is drawn in the sand arbitrarily,
> or is compression actually not much use for encryption?

The latter, sadly. To be useful in a security sense (rather than in a
traffic-limiting sense) the compression has to approach perfect - perfect
compression being theoretically impossible - and it's too hard to do in
practice.

But suppose it were easier - you approach perfection, but if your adversary
is a little closer then he wins - you have to be better than him to win, and
more, you have to be sure he doesn't get lucky and find something which you
haven't thought of which will allow him an advantage.

The player with the best models wins (defining "best" as the one that wins,
of course :).

In security we do not want to be in a competition - we want it to be
impossible for our adversary to win.

Besides, the (almost-) perfectly compressed message still has to be
encrypted if you don't want a casual observer knowing what it is. The
encryption can be simple in some circumstances, as the adversary can't know
when he has correctly decrypted _one single_ message - but if eg there are a
lot of known pt/ct pairs using eg a substitution cipher with one key then
the adversary can easily decrypt the next message - if his assumption that
the key hasn't changed is correct, then he has recovered the plaintext.

It isn't much use either.

-- Peter Fairbrother

Quantum mechanics - the dreams stuff is made of.
Received on Tue Jan 3 03:41:54 2006