[cryptography] non-decryptable encryption
natanael.l at gmail.com
Mon Jun 18 08:27:04 EDT 2012
I'm not a crypto expert, but I have read a bit about it, and have some
comments. I am pretty sure that my comments below are accurate and
relevant. Feel free to correct me if I am wrong.
One: On the second paper, you assume a prime number as long as the message
is secure, and give an example of a message of 500 characters. Assuming
ASCII coding and compression, that will be just a few hundred bits. RSA
(using primes too) of 1024 bits is now being considered insecure by more
and more people. I'm afraid that simple bruteforce could break your scheme
quite fast. Also, why not use simple XOR in that case?
Two: You don't mind keys with repeating numbers. Why?
On the first paper: One: The description reminds me of steganography.
Two: Also, why swap out the symbol set? You won't gain anything by creating
new symbols, converting bits to them and back. It's just the same as adding
a substitution cipher step. It does not prevent statistical analysis.
Three: Essentially, you don't really seem to take predictability of keys
into consideration. If you want key based encryption and decryption takes
<1 second for a CPU core to decrypt, you need to have as many as somewhere
about 2^128 possible keys with equal probability to prevent bruteforce. DES
use 64 bits (2^64 possible keys) and can be bruteforced today. If you don't
allow enough possible keys and don't consider how key generation is done, I
won't trust it.
Four: One time key encryption can NOT be broken by guessing if the key is
random enough. It's a fundamental property of XOR. ANYTHING can be the
message, because there are no invalid decryptions. Guess for another
message and you get another key. Guess another key and you get another
message. While AES encryption generally ONLY give you jibberish until you
guess the right key, with XOR you can have thousands and millions of keys
that all give valid messages from the same ciphertext.
Six: You claim encryption is usually not done on the binary representation,
but on the letters. But AES does certainly not care about the meaning of
the bits, it just encrypts bits it's given.
Seven: It looks like you are misinterpreting Benford's law. It is not
intended to be applied to cryptographically random numbers, but to
"descriptive" numbers (in lack of better words). Number of cars, amount of
money, height of flowers, etc. Lower is more common, with a certain
Eight: How do you sort out garbage data on decryption?
What I would do instead of using your method is to compress the data, add a
whole lot of random padding and use normal AES encryption.
Just so you know, NSA and their likes don't mind guessing and.number
crunching until they find patterns if they want to decrypt your messages.
You seem to forget that cryptoanalysis looks for ALL kinds of patterns, not
just pure math and such.
And the example ciphertext you provided: That's not how cryptoanalysists
usually work, they usually work with harddrives full of data, listen to
radio transmissions, etc. Your message is probably too short for analysis
while a paragraph of that could be breakable.
- Sent from my tablet
Den 18 jun 2012 12:56 skrev "Givonne Cirkin" <givonne at 37.com>:
> My name is Givon Zirkind. I am a computer scientist. I developed a
> method of encryption that is not decryptable by method.
> You can read my paper at: http://bit.ly/Kov1DE
> My colleagues agree with me. But, I have not been able to get pass peer
> review and publish this paper. In my opinion, the refutations are
> ridiculous and just attacks -- clear misunderstandings of the methods.
> They do not explain my methods and say why they do not work.
> I have a 2nd paper: http://bit.ly/LjrM61
> This paper also couldn't get published. This too I was told doesn't
> follow the norm and is not non-decryptable. Which I find odd, because it
> is merely the tweaking of an already known method of using prime numbers.
> I am asking the hacking community for help. Help me test my methods. The
> following message is encrypted using one of my new methods. Logically, it
> should not be decryptable by "method". If you can decrypt it, please let
> me know you did & how.
> I have a website about this: www.givonzirkind.weebly.com
> For information about the Transcendental Encryption Codec click on the
> "more" tab.
> Also, on Facebook,
> <https://www.facebook.com/TranscendentalEncryptionCodecTec>Givon Zirkind
> You @ 37.com - The world's easiest free Email address !
> cryptography mailing list
> cryptography at randombit.net
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the cryptography