On Thu, 28 Jan 1999 Valdis(_dot_)Kletnieks(_at_)vt(_dot_)edu wrote:
On Wed, 27 Jan 1999 15:05:15 PST, Chris Newman said:
It's RFC 2482 or RFC 2231. Take your pick. Of course, language tags are
a MUST accept, MAY generate case.
Umm.. I admit I was only paying half attention here - is this a
"MUST accept and process correctly" or a "MUST accept, syntax check,
and then discard if you're not up to handling it"?
In other words, are they required to handled, or merely not flagged
as an error? And if "accepted and ignored", does that have any implications
for funky output on the screen? Or does it just mean we're going to
discard the information on (say) whether it's the Korean or Japanese
variant of a given glyph, and for the iso8859-1 limited people it will
look the same either way?
I'd say a basic implementation MUST be capable of handling received
language tags in such a way that they don't intrude into the user's
experience. A more advanced implementation might provide a way to see the
language tag or act on it in terms of glyph rendering or text-to-speech
algorithm (or both).