Re: Cohen's paper on byte order
From: Brian Gladman (email@example.com)
From: "Brian Gladman" <firstname.lastname@example.org> Date: Tue, 8 Apr 2003 19:52:41 +0100
"Roger Schlafly" <email@example.com> wrote in message
> "Brian Gladman" <firstname.lastname@example.org> wrote
> > Now a question for everyone with an interest: does the following, added
> > the end of section 3.1, encapsulate what is needed?
> > "Where these sequences are represented externally as enumerated arrays
> > 8-bit unsigned integers (commonly referred to as octets, bytes or
> > characters), the integer with an array index n will be formed from bits
> > to 8*n+7 of the sequence in such a way that higher bit indexes are
> > associated with lower numeric significance."
> I think it belongs in sec. 3.2, because that is where a "byte" is
> defined. Sec. 3.2 says:
> It is also convenient to denote byte values using hexadecimal notation
> with each of two groups of four bits being denoted by a single character
> in Fig. 1.
> I would replace that with
> It is also convenient to identify byte values with integers from 0 to
> 255, and use hexadecimal notation for those integers.
There are quite a few problems with changes to section 3.2.
Firstly section 3.2 is a part of the internal semantics of the algorithm
that is not supposed to be visible externally. To make the definitions in
section 3.2 a part of the external interface would require a change to
section 3.1 to achieve this 'promotion' and this would confuse the current
well defined separation of the internal and external semantics.
Secondly section 3.2 might seem to specify bytes but it does not really do
this. What section 3.2 specifies is the internal representation of finite
field elements and it would be quite inappropriate to suggest that this
definition should be visible externally.
And, of course, it would be potentially confusing to associate integer
semantics with the finite field objects that section 3.2 defines.
This is all far too messy for my liking, which is why I suggested an
addition to section 3.1.
If I am asked to choose between leaving the FIPS as it is now or making
changes in section 3.2, I will vote to leave it alone.