Re: Cohen's paper on byte order
From: Eugene Starokoltsev (firstname.lastname@example.org)
From: email@example.com (Eugene Starokoltsev) Date: 6 Apr 2003 05:54:09 -0700
"Douglas A. Gwyn" <DAGwyn@null.net> wrote in message news:<3E8F6055.firstname.lastname@example.org>...
> Mok-Kong Shen wrote:
> > But is that a problem 'specific' to AES??
> Yes! Haven't you been paying attention? Other
> protocol specs for information interchange are
> careful to spell out all the details necessary
> for correct communication of the formatted data.
> AES does not. It is missing an essential
> component. Putting it another way, it was an
> excess of abstraction to change the specification
> from the original one, that concerned bits in
> storage, into one about abstract mathematical
> objects without a definite mapping between the
> abstract objects and actual storage objects.
> The connection with reality was thereby lost.
I can't agree with you. Indeed AES is just a solid block to create
strong cryptographic systems. AES does not define a lot of essential
components for reliable secure unambiguous communication, starting
from modes of operation and ending by characters encoding. But it is
not a defect of AES as a standard - you just need to use another
standards together with AES or define yours.
AES is defined on a sequences of 128 bits only as data and a sequence
of 128, 192 or 156 bits as a key. This definition is very unambiguous
and consistent - exectly what I need from a standard. In addition,
Rijndael structure is so as all combinations of bit and byte
endiannesses in the communacation channel and endianness of the
hardware executing the algorithm lead to implementations of equal
complexity - it was a goal of Rijndael's authors.
Note that "external interface" of SHA-1 is defined on bit sequences
BTW thinking of "files" as just sequences of bytes is common for Unix
world but is not a universal constant. So the "reality" is not a
well-defined term :)