Re: [XHTML 2] 25.2 role and QNames in attribute context (PR#7757)

On Thu, 26 Jan 2006 13:13:19 +0100, Steven Pemberton  
<steven.pemberton@cwi.nl> wrote:
>>> QNames are the way that the working group, and indeed the W3C, handle  
>>> having
>>> data that comes from differing sources. The working group is not  
>>> willing to
>>> change course at this time.
>>  The TAG finding I was able to find on this seems not really to promote  
>> this in any way:
>>  <http://www.w3.org/2001/tag/doc/qnameids.html>
>>  Especially given section 5 of that finding I don't think this is a  
>> good way to solve the problem. Having browsers to implement support for  
>> XML Schema in order to support supposedly simple features from XHTML  
>> 2.0 is not really satisfactory.
>
> The advantages of QNames in this context are many:
>
>  1. Classical usage (such as rel="index") can stay the same.
>  2. It is extensible, so that you don't have to reissue the language for  
> new values.

The Atom WG solved this problem by having registered values  
<http://www.iana.org/assignments/link-relations> and allowing new values  
to be added by authors as _IRIs_. This solves the extensibility problem  
and does not create the complicated algorithm UAs have to resort to for  
xml:base for example. (Which can also be modified dynamically, etc.)

I suggest that the HTML WG solves this problem in a similar matter as the  
Atom WG did namely allowing a value out of a defined specified set of  
values and allowing IRIs for new values. This seems to be in line with the  
"semantic web" stuff and doesn't create issues for XSLT etc. as Jim points  
out.


-- 
Anne van Kesteren
<http://annevankesteren.nl/>
<http://www.opera.com/>

Received on Thursday, 26 January 2006 13:56:26 UTC