why compression doesn't perfectly even out entropy

[Perry E. Metzger]

Peter Monta writes:
> No, he's correct; cryptanalytic schemes like those you mention rely
> on statistical toeholds *in the context of a deterministic cipher
> algorithm*.  For one-time pads that are "cooked" or "screened" (and
> I agree that it's a silly thing to do), the toehold is much weaker,
> infinitesimal in fact.

Please learn what the context of the discussion was before
commenting. It was not about using cooked streams for one time pads.

Furthermore, I suggest you look up the Venona intercept work and tell
me again about how far an advesary will go with a tiny toehold.

.pm