> It'd seem data passed through an encrypter would be less random
that would mean a serious weakness of the crypto-algo;
"The most common precise definition of the randomness of a string of digits or of a sequence of black and white cells on a tape is that it is random if there is no way of describing it with a string of shorter length."
-- http://forum.wolframscience.com/printthread.php?s= 92e4c3a7f4b586209fe7057ac05dd7b3&threadid=1413
plaintext is usually not very random, that is, it can be compressed significantly;
in this case: randomness(plaintext + crypto.key + crypto.algo) is low,
but cryptographic algorithms are designed with the key feature in mind,
that crypto-text should reflect the least possible structure of the plaintext;
therefore encrypted data is undistinguishable from random,
until the encription is considered broken
You don't need to encrypt a file containing (pseudo-)random noise. It already looks like encrypted data. You can't get away with claiming it's just random noise. Or else we could just claim the same when caught with an encrypted bomb-design in our hands.
Okay, in the case of a PRNG, you can reveal the hoax by regenerating the file from the seed value used by the PRNG-algo, but this method is also applicable to the encrypted pseudorandom data stream.
Slashdot Mirror
> It'd seem data passed through an encrypter would be less random that would mean a serious weakness of the crypto-algo; "The most common precise definition of the randomness of a string of digits or of a sequence of black and white cells on a tape is that it is random if there is no way of describing it with a string of shorter length." -- http://forum.wolframscience.com/printthread.php?s= 92e4c3a7f4b586209fe7057ac05dd7b3&threadid=1413
plaintext is usually not very random, that is, it can be compressed significantly;
in this case: randomness(plaintext + crypto.key + crypto.algo) is low,
but cryptographic algorithms are designed with the key feature in mind,
that crypto-text should reflect the least possible structure of the plaintext;
therefore encrypted data is undistinguishable from random,
until the encription is considered broken
You don't need to encrypt a file containing (pseudo-)random noise.
It already looks like encrypted data.
You can't get away with claiming it's just random noise.
Or else we could just claim the same when caught with an encrypted bomb-design in our hands.
Okay, in the case of a PRNG, you can reveal the hoax by regenerating the file from the seed value used by the PRNG-algo, but this method is also applicable to the encrypted pseudorandom data stream.