Dynamic Hill cipher
in [Cryptography]
|
From: Bryan on 1 May 2010 15:32 Maaartin wrote: > I suppose he wants to keep the coefficients of the PRNG secret as > well, which leads to non-linear equations for them (or am I talking > non-sense and should go to bet now?). This can get complicated for me, > but there are methods for computing the coefficients from the sequence > which may get adapted for the case you know only a linear combination > of them. I wouldn't take the challenge seriously. Shen doesn't understand the subject well enough to define a reasonable test. He'll put forth a claim that you can demolish, but demand you do something different to prove your case. Even a system that is perfectly linear could meet Shen's requirements, as long as the PRNG has reasonable statistical properties. That's all you need to consider to refute his argument for security, so if the challenge were fair you ought to able to assume it. -- --Bryan
From: Mok-Kong Shen on 1 May 2010 17:14 Maaartin wrote: > Bryan wrote: >> Even a system that is perfectly linear could meet Shen's requirements, >> as long as the PRNG has reasonable statistical properties. That's all >> you need to consider to refute his argument for security, so if the >> challenge were fair you ought to able to assume it. > > I'm refusing his argument (as I ever did), I only state that secret > coefficients makes it non-linear, i.e., much harder. I'm looking for a > paper about analysis of LCM. A question of ignorance: What do you denote by LCM? (I only know the meaning "least common multiple".) BTW, in a previous post I have lowered the threshold of the difficulty of the challenge task. On the other hand, for obvious reasons, like in all contests, I am taking the liberty to set the closing date of acceptance of the challenge offer on 1st June, 2010. Regards, M. K. Shen
From: Maaartin on 1 May 2010 18:15 On May 2, 12:00 am, Bryan <bryanjugglercryptograp...(a)yahoo.com> wrote: > You should also require the *decryption* program, that takes the > ciphertext and key and returns the plaintext. You should require that > for the same key, the decryption program generally inverts the > encryption program. Yes, I realize you knew that, but you forgot to > specify it, and the challenge before you is not fair. Well, I forget quite a lot of things. But before taking the challenge, I'll try much harder to recall all necessary things. > You do have a sensible idea: negotiate the challenge to be objectively > testable with self-evident results. Alas, it's trickier than it looks: > For decades, noted skeptic and MacArthur Fellow James Randi offered a [snip] Funny story.
From: Mok-Kong Shen on 1 May 2010 18:15 Mok-Kong Shen wrote: > BTW, in a previous post I have lowered the threshold of the difficulty > of the challenge task. On the other hand, for obvious reasons, like in > all contests, I am taking the liberty to set the closing date of > acceptance of the challenge offer on 1st June, 2010. Mr. Maaartin, I suppose that this give you ample time to specify all details that you need to guarantee that the contest is fair. You could certainly tentatively post your specification for negotiation with me, so that experts of the group may examine and see whether there were loopholes that need to be closed up before you accept the offer. Regards, M. K. Shen
From: Maaartin on 1 May 2010 18:20
On May 1, 11:14 pm, Mok-Kong Shen <mok-kong.s...(a)t-online.de> wrote: > A question of ignorance: What do you denote by LCM? (I only know the > meaning "least common multiple".) My fault, I meant LCG (Linear congruential generator). > BTW, in a previous post I have lowered the threshold of the difficulty > of the challenge task. On the other hand, for obvious reasons, like in > all contests, I am taking the liberty to set the closing date of > acceptance of the challenge offer on 1st June, 2010. Ok, I need to think about it and to specify it in a bullet-proof way. It may happen, I found out, that it's too easy; in this case I'll tell you. |