--On Thursday, February 23, 2017 13:44 -0800 Jim Schaad
<ietf(_at_)augustcellars(_dot_)com> wrote:
I have used and have seen used the term of US-ASCII to refer
to 7-bit ASCII as opposed to the full 8-bit ASCII. I think
that this generally makes sense and therefore am unsure that
the term should be removed.
Jim,
I don't have time to go through this again, but...
(1) Depending on what you count, there is either no such thing
as "8-bit ASCII" ("full" or otherwise) or it can refer to any of
the following:
(i) 7-bit ASCII in octets with leading zero bits, as
described in RFC 20 and other places.
(ii) ISO 8859-1 (aka "Latin-1")
(iii) The entire ISO 8850 family, or at least the subset
of it that have all or most of ASCII (or at least ISO
646 IRV) in the low-order seven bits
So it either doesn't mean what you think it means, or it is
hopelessly ambiguous.
"US-ASCII" is a different matter. It is a term of art in the
IETF, referring to a "charset" that was specified as part of
MIME and is now listed in the IANA charset registry. It is
essentially equivalent to the first definition above.
Unfortunately, "US ASCII" is used in some other contexts to
refer to other things, in particular to distinguish between the
repertoire of [7 bit) ASCII (i.e., ANSI X3.4) and that of ISO
646 IRV and BV. I personally think the IETF would be better
off if we retired the term but YMMD (indeed, I'm probably in the
rough), and I'm therefore ok with its use in this context as
long as there is a normative reference pointing to a clear
definition.
My problem is not "remove the term" but being absolutely sure
that whatever terminology you use is completely unambiguous.
best,
john