Folks, Dana sent me a private note in response to my posting that I've
gotten his permission to excerpt/quote from....
ah, but it would seem to me that it does so by labeling the
aural form of the language, not its writen form. I would
describe our present conversation as English content
expressed in Roman script.
Dana,
I fear I've gotten too old to be able to distinguish this from
hair-splitting, or to enjoy either activity. In the presence of a
standard that seems to work well for what appears to be (superficially
at least) a community with similar needs, I find myself a lot more able
to deal with "they really aren't similar because..." and "won't work
because" statements than with fine classification issues. (However, I
don't want to discourage anyone else from that line of discussion if
they find it helpful.)
Now, that said...
My copy of 636 isn't here, but I'd imagine a careful study of it would
show language codes tied to a "what is a language" abstraction that is
pragmatic rather than precise (or even consistent). I agree with your
description of the "present conversation". However, let's assume we
both read Greek script well enough (and in the same traditions) to
pronounce the phonemes that would typically be associated with a string
of Greek characters. Given that, we could write this "English content"
in "Greek script". Most people would classify our doing so as a "code",
not as "Greek" or "English".
For whatever it is worth, all uses of ISO 636 to describe computer-based
documents that I'm aware of ultimately use both a language code and a
character set code. That might be pretty close to the content/script
distinction; maybe not.
Since I don't have a charsets WG chair I can unload the job on, let me
put on my AD hat for a minute. This group might reach consensus that
language tagging at the body part level in MIME is of no earthly use to
anyone. If that turned out to be the case, this proposal would
presumably disappear.
Assuming that it isn't, a list of languages and language codes is going
to be needed. While I appreciate, and agree with, Harald's suggestion
to let IANA register names for special circumstances in which ISO TC46
can't usefully move rapidly enough for us, making language categories
and lists is really not an IETF area of expertise: we should use
something else if we can find it. If there are alternatives to ISO
636, it would be useful to start a discussion of their relative
strengths and weaknesses as a base.
--john