This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
In a comment of August 2004 [1], Paul Biron suggests a correction to the description of the language type, to make it accept the empty string, in the same way that the xml:lang attribute can have the empty string as a value (with the meaning: any language information given for parents does not necessarily apply here). He also suggests ensuring that xsd:language is aligned with RFC 3066 bis. [1] http://lists.w3.org/Archives/Public/www-xml-schema-comments/2004JulSep/0086.html In February 2005 ([2], member-only link), the WG appears to have been persuaded that a correction should be issued. [2] http://lists.w3.org/Archives/Member/w3c-xml-schema-ig/2005Feb/0096.html I believe that since that time, a contrary view has gained ground, which holds that xml:lang should simply be defined as a union of xsd:language and a string-based type with a single enumerated value, namely the empty string. I find that approach satisfactory, myself. But the comment from Paul Biron, and the analogous comment from the SVG working group, suggest that even if we don't change xsd:language, an explanatory note would be useful. On the empty string issue, the WG needs to decide whether (a) to issue a correction to 1.0, including the empty string in the lexical and value spaces of xsd:language, (b) to do nothing, or (c) add a note observing that xml:lang and any construct intended to behave like it should be defined as a union of xsd:language and the empty string. On the 3066bis issue, the WG needs to decide whether to align xsd:language in XSD 1.0 with 3066bis, or allow our language type and the language codes specified by the IETF to go out of alignment with each other.
The Working Group discussed this issue today (14 December 2007) and agreed that we do not want to change the xsd:language type to allow the empty string as a value, and that we do want to add a note pointing out that xml:lang does allow empty strings and thus needs a type like the one given in the schema document at http://www.w3.org/2001/xml.xsd (a union of xsd:language and an enumerated type with one value: the empty string). I'm marking the issue as 'decided', since the WG did not wish to review a wording proposal, and 'needsDrafting'.
The note agreed upon on 14 December has been integrated into the status quo text. So I'm marking this resolved.
As the originator of record, I note my agreement with the disposition of this issue and CLOSE it accordingly. Since in some sense it reflects a concern raised by Paul Biron in 2004, I also note that Paul was present on the call that agreed on the disposition and that I therefore presume him to be content with it.