Re: [CSS21] Please endorse xml:id

On Thursday, June 30, 2005, 5:53:53 PM, Anne wrote:

AvK> Chris Lilley wrote:
>>>> In what way would it prevent exiting CR?
>>>
>>> Because it would require support for xml:id
>>
>> 'should' is not the same as 'must'

AvK> Only when you have compelling reasons.


>>> and UAs might not have implemented that while having
>> implemented CSS 2.1.
>> 
>> Which would prevent what, exactly? Are you really saying that (for 
>> xml, remember, not for html or for xhtml 1.0/1.1) there are UAs that 
>> implement 100% of CSS 2.1 and yet do not, or could not, implement 
>> xml:id?

AvK> Most UAs that exist today — Mozilla, Opera, Safari — are XML UAs with
AvK> some knowledge of the XHTML namespace plus a tag soup parser.

I'm familiar with the concept.

AvK> None of these three have support for xml:id at the moment.

One might ask what stops them adding it, but you go on to partially
answer that.

AvK>  I know
AvK> there were plans for Opera 8, but it didn't make it. For Mozilla
AvK> there are also some unresolved issues that prevent developers from
AvK> implementing it.

Pointer to those unresolved issues? Are they specific to Mozilla, or
would they prevent others implementing xml:id?

AvK> Also, although there is a relation between xml:id and CSS

yes...

AvK> that shouldn't really result in a "UA that implement CSS and XML
AvK> should implement xml:id." Better would be to require XML UAs to
AvK> implement xml:id and adopt xml:id inside the official XML
AvK> recommendation.

which would mean that id assignment would be reported, but would not
require, nor test, that CSS selectors were picking up that id
assignment.




-- 
 Chris Lilley                    mailto:chris@w3.org
 Chair, W3C SVG Working Group
 W3C Graphics Activity Lead