- From: Orion Adrian <orion.adrian@gmail.com>
- Date: Fri, 16 Dec 2005 13:58:58 -0500
- To: www-style@w3.org
On 12/16/05, Anne van Kesteren <annevk@opera.com> wrote:
> On Fri, 16 Dec 2005 13:37:40 +0100, Orion Adrian <orion.adrian@gmail.com>
> wrote:
> > However, he is right in that you can apply semantic behavior through
> > CSS, and not just a look.
>
> Not in the way he suggests you do it.
And in the way I'm suggesting it?
> > myheading
> > {
> > role: heading;
> > }
> >
> > And this is perfectly accessible, or at least as accessible as XHTML
> > is when you have CSS. XHTML just implies certain role behaviors.
>
> It is not. Your examples are not supported by any UA. XHTML is supported
> by most.
Accessibly requires knowledge of at a minimum. Since no UA knows about
it, of course it isn't as accessible in actuality, but the proposal
does allow for accessibility if supported, which was what was in
question if I understand correctly.
> Also, this would require some additional document where XHTML already
> contains this kind of meaning. Not to say I really dislike your "proposal"
> though. Quite often there seems to be some need to let CSS specify
> semantic hints/semantics.
I think the fundamental problem here and why everybody wants to
populate CSS with junk is because they want the matching mechanism.
They want to be able to say, all these like this have these
properties.
It's part of the reason I'd like CSS to be broken apart into it's core
components and standardized seperately. Specifically the matching
algorith + selectors - certain pseudo-classes; the box model and the
properties that can be applied to it; and the grammar.
--
Orion Adrian
Received on Friday, 16 December 2005 18:59:39 UTC