Re: 'role' should be property

On 2007-06-01 01:10:50 +0200 Maciej Stachowiak <mjs@apple.com> wrote:

> On May 31, 2007, at 4:05 PM, Leif Halvard Silli wrote:
> 
>> On 2007-05-31 18:41:20 +0200 Maciej Stachowiak <mjs@apple.com> wrote:
>> 
>>> Can we please take the discussion of changing CSS somewhere that  it  is 
>>> more on-topic, such as <www-style@w3.org>?
>> 
>> Right. But. I don't think you are in support of the role attribute?  So if 
>> we agree that it belongs in CSS instead, then we can skip the  role debate 
>> here also?
>> 
>> Also, Dmitry says «focus on 'class'» and «do less, but add more  meaning 
>> to the little you do». I think you say the same.
> 
> I'm personally not in favor of either the role attribute in markup,  or a 
> hypothetical role CSS property.


But then think about something more relevant for you, think about the scope and headers attributes of TABLE. Dmitry himself in of his letters use a table to illustrate his point.

In the parallell thread about the headers attribute, you raised the question about what implementations should be required to support. Well ... I am not saying that it is possible, but _if_ it were possible to move the things that scope and headers do *out* of HTML and into CSS, then suddenly it became much simpler to create HTML5 conformimg UAs, and to differentiate the UAs - they do not all have to support all CSS properties.)

I have man more points to make abotu scope and headers. I make them in the parallell thread. I'll just mention that it would be logical for headers to not only take ID's as values, but also class. 
-- 
leif halvard silli