W3C home > Mailing lists > Public > public-rdf-in-xhtml-tf@w3.org > August 2004

Quick glance through xhtml2 meatainformation module

From: Dominique HazaŽl-Massieux <dom@w3.org>
Date: Fri, 06 Aug 2004 11:33:20 +0200
To: public-rdf-in-xhtml-tf@w3.org
Message-Id: <1091784800.1416.2454.camel@stratustier>
I glanced quickly through the XHTML 2 meta-information module [1], and
even more quickly through the XHTML 2 meatainformation attributes module
[2]; I'm posting my comments here, since I haven't had time to
investigate them enough to post formal comments to the HTML WG [I'm
leaving in vacation in a few minutes :) ] but I hope this may help other
people doing so.

Overall, it looks really good and well-thought; the issues I spotted
while glancing through it:
- there is no way to distinguish between string literals and XML
literals within the <meta /> element, since they appear to be both (!)
the default content model of <meta />; I guess this could be solved by
requiring that the datatype attribute be set for one or the other (the
metainformation attributes modules says that datatype defaults to
"xs:string"; I think this makes the case that it can't be an xml literal
in that case)
- the introduction to datatype mentions the need of setting the unit of
a height property, but doesn't explain how to do so afterward; IIRC,
this strongly relates to another TF topic of the SWBPDWG 
- the absence of reference to RDF could be damaging; e.g., as of today
the use of a particular property is not constrained to refer to a given
schema language (and indeed, not to RDF, OWL, ...), which may makes the
RDF-ization of an XHTML document impossible if people were to start
using random other properties list [ I guess this relates to their
highlighted issue "We need to codify the mechanism for mapping prefixes
to namespaces for meta data profiles." ]
- the QName model for the "rel" (and "rev") property doesn't make clear
how a prefixed name should be interpreted; I guess one way to alleviate
this would be to reserve non-prefixed names to those defined further
below (ŗ la "alternate", ...)
- of interest, the "rev" attribute (inverse to "rel") introduces the
possibility of expressing syntactically an inverse property, the same
way you can do in N3 with the "is <property> of" construction; note that
RDF/XML doesn't have such a construction AFAIK
- since both <meta> and <link> support the common attributes modules
(Core, Events, I18N, Bi-directional, Edit, Embedding, Map,
Metainformation, and Hypertext described in [3]), it would be worth
reviewing them very carefully to see how/whether these attributes would
affect their interpretation; xml:lang and its history with RDF/XML is
probably worth a very close look, as may be hypertext/embedding

I think it would really be worth spending some time writing a few of the
RDF/XML test cases in XHTML to detect inconsistencies/incompatibilities;
I unfortunately don't have the time to do it for now.

Hope this still helps a bit,


2. http://www.w3.org/TR/2004/WD-xhtml2-20040722/mod-metaAttributes.html
Dominique HazaŽl-Massieux - http://www.w3.org/People/Dom/

Received on Friday, 6 August 2004 05:33:34 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:50:18 UTC