DRAFT review of XHTML Role Attribute Module

Review of:

 XHTML Role Attribute Module
 A module to support role classification of elements
 W3C Working Group Note 16 December 2010
 <http://www.w3.org/TR/2010/NOTE-xhtml-role-20101216>

Previous versions of this note included a mapping of the XHTML role
attribute to RDF, though this seems to have been removed in the current
version. The RDFa working group had been critical of the wording of this
mapping, as it appeared to place normative requirements on conformant
RDFa processors. However, I think we were broadly in favour of such a
mapping existing, as it provided an easy method of making simple
statements about XHTML elements themselves. (Whereas RDFa focuses on
making statements about XHTML documents, and about the things described
in them.) With that in mind, it seems a shame this feature has been
removed, though it's understandable if consensus could not be achieved
on this mapping.

The note includes a normative reference to CURIE Syntax 1.0. While CURIE
Syntax 1.0 is not the subject of this review, the RDFa working group
hopes that it is broadly compatible with RDFa's own CURIE syntax.

In RDFa 1.1, attributes which accept CURIEs also accept absolute URIs.
Is this feature supported in the role attribute? If so, then the note
could do with clarification. If not, then it would be useful to note
this difference.

The HTML working group's current drafts of HTML 5 include an attribute
named role, so HTML+RDFa implementations may need to be aware of the
role attribute. It would be useful if an informative section could be
included in the XHTML role attribute note detailing to what extent (if
any) it is compatible with HTML 5's role attribute.

The author of this review would also like to point out that markup like
the following:

 <mylang xmlns="http://www.example.com/dtd/mylang" 
  xmlns:xh="http://www.w3.org/1999/xhtml" xml:lang="en" xh:role="foo">

is inconsistent with the role attribute in XHTML, where it is not in any
namespace at all (not even in the XHTML namespace).


-- 
Toby A Inkster
<mailto:mail@tobyinkster.co.uk>
<http://tobyinkster.co.uk>

Received on Tuesday, 25 January 2011 23:12:11 UTC