Re: CSS selectors and xml:id

On Monday, May 9, 2005, 1:01:00 PM, Anne wrote:

AvK> Chris Lilley wrote:
>> HST> Personally I would settle for DTD and W3C XML Schema in some detail,
>> HST> refer to 'other/external methods' and give xml:id as an e.g.
>> 
>> That  would be suboptimal, for example there would be no conformance
>> requirement and thus no xml:id samples in the test suite.

AvK> Really, why should the *CSS* test suite check for conformance on xml:id?

Where else would you expect CSS selector tests to go?

AvK> Really, theoretically you can implement just for HTML or for neither XML
AvK> nor HTML and still be conformant I believe.

But you would have a hard time proving it, if you implement neither HTML
nor XML.

AvK> Also, the arguments given that Amaya supports xml:id and that now
AvK> only one other browser has to support it to meet the requirements
AvK> is nonsense as Amaya hasn't implemented all of CSS.

I don't recall mentioning that it had? Implementing 'all of CSS' is
something as yet unattained, naturally, for any product.

AvK> Two interoperable implementations means that there are two
AvK> implementations which implement *every* feature of the
AvK> specification in an interoperable way. You can't say those two
AvK> implementations support that feature and those two support that, et
AvK> cetera. There have to be two implementations that support all of it
AvK> and are interoperable.

Right. And Amaya shows that adding xml:id support to an existing CSS
implementation is trivial, so (assuming the CSS implementors are, in
fact, interested in using CSS with XML) it would seem that they could
add it very easily.




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

Received on Monday, 9 May 2005 13:04:45 UTC