Re: integration testing of xml:ID

On Tuesday, April 19, 2005, 8:16:53 PM, Dan wrote:


DC> This section "Impacts on Other Standards" is great...
DC>   http://www.w3.org/TR/2005/CR-xml-id-20050208/#impact

DC> I see
DC>  "Informed readers that C Impacts on Other Standards will be removed
DC> before xml:id becomes a recommendation"

DC> Hmm... well, it would be great to remove the "Impacts..." stuff
DC> because it's all taken care of, but looking at a draft
DC> implementation report...
DC>   http://www.w3.org/XML/2005/01/xml-id-implementation.html

DC> ... I don't see, for example, a test corresponding to
DC> this ID selectors test...
DC> http://www.w3.org/Style/CSS/Test/CSS3/Selectors/current/xhtml/tests/css3-modsel-15.xml

I agree that there is a grave risk of a 'falling between two stools'
here. If XML core says its not their job to test xml:id in CSs
selectors, and CSS says its not their job to test xml:id in CSS
selectors, then it is not tested.

DC> Before removing the "Impacts..." section, please at least
DC> add a link from the implementation report to something
DC> from the CSS WG that says "yes... good question... we'll
DC> look into that."

As a minimum requirement.

DC> I can't tell if the intent is that the CSS specs get revised
DC> to refer to xml:ID or if CSS implementations are to independently
DC> pick up xml:ID support or what. The current spec is kinda waffly,
DC> as far as I can tell...

I recall asking that question too (about DOM getElementById as well)

DC> "Document languages may contain attributes that are declared to be of
DC> type ID." is pretty waffly
DC> http://www.w3.org/TR/1998/REC-CSS2-19980512/selector.html#id-selectors

DC> I haven't looked into DOM, XPath, etc. as closely, but the concern
DC> applies.





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