[tei-council] [Fwd: xml:lang attribute and P5 RNG schemas]
Lou Burnard
lou.burnard at computing-services.oxford.ac.uk
Sun Nov 28 16:48:42 EST 2004
The appended note to TEI-L reminds me that the current draft for P5 is a
bit noncommittal on the subject of xml:lang
I think we ought to try to decide
(a) whether P5 should say that xml:lang *replaces* lang or may be used
as an alternative
(b) whether the declared value of xml:lang should be IDREF or CDATA
At present the draft for CH recommends a particular way of constructing
values for xml:lang which is derived from relevant standards: it has the
status of a WG recommendation but has not yet been formally approved by
the Council.
For the record my vote is
(a) lang should be replaced by xml:lang
(b) it is CDATA
Lou
-------- Original Message --------
Subject: xml:lang attribute and P5 RNG schemas
Date: Mon, 22 Nov 2004 00:10:08 +0000
From: Peter Heslin <public at HESLIN.ECLIPSE.CO.UK>
Reply-To: public at HESLIN.ECLIPSE.CO.UK
To: TEI-L at LISTSERV.BROWN.EDU
I am using Emacs nxml-mode to edit TEI documents, and I am using the
pre-release version of the Relax NG schemas that I downloaded from the
TEI web-site. Somewhere on that site,[1] I read that P5 will recommend
using the standard xml:lang attribute rather than a TEI-defined lang
attribute. So I tried that, but nxml-mode complains that xml:lang is
not a valid attribute.
Sure enough, the TEI schemas do not declare xml:lang, whereas, for
example, the XHTML schema that comes with nxml-mode does declare it.
It seems odd to me that it should be necessary to declare an attribute
defined by the XML specification itself. So it may be that this is a
bug in nxml-mode rather than in the P5 schema. Three possibilities
occur to me:
1) I am mistaken in using xml:lang rather than lang in TEI P5, or
perhaps the switch to xml:lang has not become official yet.
2) This is a bug in nxml-mode, which should always allow the xml:lang
attribute in any element; it should be removed from the XHTML
schema.
3) This is a bug in the P5 schema; xml:lang should be declared by
schemas that want to use it.
Best wishes,
Peter
[1] I can't remember where I saw this, but google gives this abstract:
http://www.hum.gu.se/allcach2004/AP/html/prop142.html
More information about the tei-council
mailing list