Re: Key entropy, stream entropy, block entropy, block population entropy AKA uniique stream length

On 6 Feb, 19:23, g...@xxxxxxxxxxxxxxxx (Greg Rose) wrote:
In article <1170751546.573225.152...@xxxxxxxxxxxxxxxxxxxxxxxxxxx>,

<j...@xxxxxxxx> wrote:
It is ok they just try to protect their positions and i don't care but
do not try to call me a fool when i am right.

a) I didn't call you a fool. I merely said that
you are mistaken. It's an easy and common mistake
to make.

I merely tell you again, you are wrong and i am sure you are aware of
it, and that bother me much more than your refusal to accept facts.

b) Kristian has already explained why it's a
mistake. But just to reiterate: if there are only
2^256 keys, there are only 2^256 output streams,

Yeah and so have i told you for many hours now the number of streams
created from a block can not be be greater than the key entropy.

However every stream *IS NOT JUST ONE BLOCK*
The stream can have large a large internal state that make it possible
to render each possible *OUTPUT BLOCK* in more ways then the key have
entropy. Why do you play you do not understand this.

I know you know that a key can be expanded, i know that you know that
if the expansion result in a *PERMUTATION* that permutation alone can
have more states then the KEYENTROPY.

If a 256 bit key is expanded to 2048 bit AKA a 256 byte shuffle the
shuffle in itself will have 256! states. If you reverse the shuffle
and create a mirror to XOR it with and store the result in a PRNG
BUFFER..
You create a PRNG STREAM of 2048 bits.

You shuffle both streams again XOR them together and update the PRNG
BUFFER by a XOR with the result.
If you can not see that the there is more ways than 2^256 ways to
create one *SPECIFIC BLOCK* it must be pure refusal.
The permutations alone have 256! faculty, and given that the PRNG
BUFFER do not have the same value within when it reach its end that it
hand from the start.......
I would say the that it is even possible that the PRNG BUFFER could
have 2^2048 outcomes for every specific permutation.

And if you take a look at those surely *overestimated possibilities*
(because i do not think the PRNG reach every state for every SHUFFLE.
One can easily derive that any block of PRNG OUTPUT can be constructed
from more then 2^256 different internal states.

Example: As you can see a string of just 2048 *1* or zeroes is even
possible to create in as many ways as any other specific state. You
can actually have any state in the *PERMUTATION* and create them given
right state in the *PRNG BUFFER*

no matter whether you expand blocks and XOR them,
or whatever... if the decisions how to combine
blocks are made depending on the key. If the
decisions how to combine blocks are made
independent of the key, new entropy has been
added.

Well ok now i had it *PLEASE RUN MY CIPHER* with a one byte key or two
byte password/key and you will notice one thing about the stream *IT
HAS SAME ENTROPY THAT A 256 BYTE KEY*

Because it is a keyexpansion permutation shuffle.

Please you people here who doubt me run the algorithm with keys of
different lengths and check the output in DIEHARD.
It will encrypt a 200 MB files (if i remember correct) with just
zeroes.
The stream have same properties for 2 or 3 byte key as it has for a
256 byte key due to the keyexpansion so please fuckin run the program
with a short key.

So would you be so kind to run it

Jonas Thörnvall



This is no lie


c) It isn't *my* position, or Kristian's, it
belongs to Claude E. Shannon. And, since it is the
foundation of information theory, I don't feel a
need to protect it. I, and if I may be so bold as
to speak for him, Kristian, are just trying to
increase the amount of clue in the world.

Greg.
--
Greg Rose
232B EC8F 44C6 C853 D68F E107 E6BF CD2F 1081 A37C
Qualcomm Australia:http://www.qualcomm.com.au


.