QUOTE(Mja @ 2008-03-17 11:17)

I didnt say he should use the PSK, I said that the initial post claimed that PSK was used. PSK as such is used for the term when pre shared encryptionkeys are used, or well, at least thats what I thought.
Yes, that's what you thought.
QUOTE(Mja @ 2008-03-17 11:17)

Going OT a bit:
...
Cryptography, to be safe, should use certificates, or at least keypairs, assymetric.
These keys should be used for key-exchange, as well as signing.
Keys used for the encryption should have a max time to live, either based on the amount of data encrypted by it, or time passed since last change.
As long as a decent length of keys for the assymetric encryption is used, and symmetric keys for the encryption of transfered data is changed on a reasonable basis, the keylength of the symmetric key has less impact on the security, given the fact that even if you manage to decrypt it, you will only have access to a limited amount of data.
Just because the standard containers use all this doesn't mean that it should be used in every single design.
There are ways to exchange keys without the use of asymmetric ciphers, but since you've been taught the basic stuff - that's what you think is the case in all implementations. And since you've already exchanged keys in a PSK-situation, why do it again?
You keep talking about a COMPLETELY DIFFERENT SUBJECT! by dragging the asymmetric ciphers along - and FYI it's an english invention and NOT an american. Thank you very much!
QUOTE(Mja @ 2008-03-17 11:17)

I guess this is where your "HOW random?" question arrise, since if you only generate pseudo-random keys, for example with a static seed hardcoded, the same sequence of keys will be used every time, nothing random in that.
If you dont "re-seed" the random-algorithm, a predictable pattern of random values (read: It looks random, but can be predicted, the curse of randomization in computers) will be the result. In this case, even if you swap keys during the encryption, someone that crack ONE key will be able to deduct the previous and following ones, as long as they know the algorithm.
No that's also wrong!
It's not the "re-seeding" that's the problem here but the genericly cryptographicly insecure random number generators implemented in the standard libraries. And you're still talking about a situation with asymmetric key-exchange - IT'S NOT!
QUOTE(Mja @ 2008-03-17 11:17)

Anyhow, I think I´ll drop the subject, seeing there are others far more interested and well informed in the areas of cryptograhpy, I´ll only make a fool of myself discussing it more

Listen to Caecigenus, nod, and then do what you like

Apparently - yes.
But you still had to make that looong post, spreading false claims.
@Mja: Du är inte lite kaxig du... känner man inte lite troll-vibbar?