[tei-council] [Fwd: xml:lang attribute and P5 RNG schemas]

[Syd Bauman]

I don't think the WG was non-committal on this at all. The WG was
completely firm (with 1 dissent) on using xml:lang=. The WG was a
little less firm (3 explicit votes in favor, no dissension) on how
the value of xml:lang= would be (optionally) associated with a
<language> in the header. The 3 explicit votes all favored the plan
I brought up in Ghent, namely that <language> should bear a
(required?) ident= attribute, and the value of xml:lang= would be
compared to it.

The WG made no suggestion as to the type of attribute xml:lang= would
be, but I have always presumed we would declare it as tei.langtag or
some such, which in turn would be declared to be as close as we could
get to an RFC 3066 language tag. I.e., in DTD language it would be
CDATA, if we could use RelaxNG directly it would be xsd:language, a
datatype defined by W3C Schema just for this[1]. The only minor catch
is that it is defined as RFC 3066 language tags, and by the time P5
comes out there may be a successor to 3066. I think we should presume
W3C will follow suit relatively quickly (they did move from 1766 to
3066 when that happened, I think).

Note
----
[1] See http://www.w3.org/TR/xmlschema-2/#language.