Hacker News Reader

Comment by rnovak | Hacker News Reader

I would really love to submit this encryption method to the PCI auditors.... "uses encryption extensively".

Edit: seriously? The first example is a shift cipher and a one time pad of all 'M'.

Perhaps you'd be happier with the term "obfuscation." But the code does meet the textbook definition of "encryption." Maybe not strong encryption...

Where did I say it wasn't encryption? I said I'd like to submit it to my PCI auditor as _my_ method of encryption, much easier than proper applied crypto.

Yes: the term the VX scene uses is "crypter".

Code size and unusualness/uniqueness are goals here, as they're for obfuscation, not strong encryption - the code needs to be small, hard for an AV to recognise with a heuristic (so tables and baked constants are bad) and ideally amenable to being passed through a polymorphic/metamorphic mutation engine.

And, the key 'MZ' used in DecryptStrings() seems to be a "tribute" to DOS executable header :)

It's more accurate to call it a Caesar cipher in GF2(8) instead of a one-time pad.

Caeser cipher _is_ a shift cipher, which is exactly what I said it was...

I was referring to "and a one time pad of all 'M'".

Yeah, sorry, I'm not good enough with the math behind crypto enough to have recognized that :(

learn something new every day.

Perhaps my notation was a bit too terse. GF2(8) is a Galois field of order Pow(2, 8). The xor operation is addition in a Galois field, so 8-bit xor with 'M' is the same thing as a constant additive shift by 'M' in a Galois field of order Pow(2, 8). So, it's a Caesar shift cipher using the GF2(8) definition of byte addition rather than the standard modulo 256 definition.

Reply on news.ycombinator.com