On 07/04/11 04:00, Piotr Bański wrote:
> Obviously, TEI XML assumes the XML Spec as augmented by, among others,
> the XInclude spec.
I don't agree that it is obvious.
The "language" property defined in the XInclude spec is not part of the
XML Infoset (which would make it normative for TEI); it's a transient
property defined by the XInclude spec as a convenience for defining an
algorithm for "fixing up" xml:lang attributes in transcluded documents.
The document which results from an XInclusion does not contain this
"language" property; the result of applying the "language fixup"
algorithm is only to create new xml:lang attributes in the output
document as necessary to identify the language of the transcluded
document (where it differed from the language which was in scope at the
point in the document into which it was transcluded).
So in short I don't think the "language" property defined in XInclude
has any relevance to TEI as such.
The xml:lang attribute defined in the XML spec does have relevance, but
it already has rather narrowly-defined semantics which the TEI community
is not licensed to modify or reinterpret.
Personally, I think the current discussion has some value in that it
points to a perceived need for such a facility, but I think that the
focus on xml:lang is not really helpful.
Syd is entirely correct to say:
>> don't believe there is any chance that the W3C XML Specification
>> (which is where xml:lang= is defined) includes remotely referenced
>> content when it says [...]
And though I understand what you mean by "virtual content", and I do
have some sympathy for people doing stand-off markup in TEI, and who
want the facility to indicate the language of the target of their
stand-off markup, I still think it's a mistake to try to press-gang
xml:lang into this role. A better approach, in my opinion, would be to
introduce a new attribute in TEI, e.g. refLang, with the required
semantics, whatever they might be.
eResearch Business Analyst
Victorian eResearch Strategic Initiative