Re: The strength of "post-encryption encryption"?

daw@xxxxxxxxxxxxxxxxxxxxxxxx (David Wagner) writes:

Unruh wrote:
Also the only provably secure encryption is an "xor"

"Douglas A. Gwyn" <DAGwyn@xxxxxxxx> writes:
Well, that's not so.
To the extent that a perfect one-time pad system qualifies
as secure, which is presumably what you were referring to,
it doesn't have to use XOR operations.

Unruh wrote:
No, but in our modern binary world it does.

No. You're just confused. Doug Gwyn has got it right.

P.S. Not every operation works, but there are a large class of
operations that do work. For instance, any group operation would
suffice. XOR is the simplest example of a group operation, but it
is not the only example.

Weird. You extracted one line from my post, the rest of which made exactly
the point you are making. I was saying that the USUAL operation used for
one time pads ( and stream cyphers) tends to be xor. Sure, as I stated, you
can use the one time pad stream as a set of keys to AES and encrypt the blocks of the
text each with a separate key. Or any other invertable operation.
It would be silly, but possible. However in
general you want the fastest possible operation. That is almost certainly
xor. And since xor is as safe as any other operation in the infinite list
of possibles, why in the world not use the fastest.