RE: ARIA as stop-gap (was Re: Next steps for the ARIA syntax discussion)

Leif -

Ah, yes, I do see the differences now between our ideas. It is important to
note that my proposal would still allow ARIA within the HTML itself, which
is where I believe it rightfully belongs; it is more of an attempt to get
designers/developers to be thinking of ARIA as something that is at the same
level as CSS, and that the two go hand-in-hand.

I think we're all reaching out here, trying to find a way to make the ARIA
spec more useful that a mere attribute would make it. I still think that it
would be extraordinarily helpful to figure out a mechanism to allow a link
between CSS class and ARIA "role".

Simply put, designers treat classes/selectors/IDs/etc. in CSS as semantic
entities, but sadly, CSS by itself conveys no semantic information. They
define a class or a selector or whatever that can *only* be used in certain
situations, and if you see it on the screen it quite clearly fulfills a
particular role, but the machine has no idea what that role is (since it is
imitating the purpose of another widget through CSS and JavaScript).
Allowing a connection between CSS and ARIA is the best (in my mind) way of
bridging that gap, at the expense of having the two specs slightly
entangled. Given that no one uses CSS outside of HTML, and that no one will
use ARIA outside of HTML, I don't see that as a problem anyways.

J.Ja

-----Original Message-----
From: public-html-request@w3.org [mailto:public-html-request@w3.org] On
Behalf Of Leif Halvard Silli
Sent: Wednesday, June 04, 2008 5:30 PM
To: Justin James
Cc: 'Anne van Kesteren'; elharo@metalab.unc.edu; public-html@w3.org;
wai-xtech@w3.org; www-tag@w3.org
Subject: Re: ARIA as stop-gap (was Re: Next steps for the ARIA syntax
discussion)


Justin James 2008-06-04 19.42:

> I think you hit the *exact* same1 realization as I did yesterday. This is
> the heart of my proposal, which is to let the class (or ID, or whatever
the
> appropriate CSS definition is) contain the ARIA information.


Actually, no. I suppose you refer to this proposal [1]:

> div.checkbox {aria-role: checkbox;} 
> <div class="checkbox">Blah blah blah</div> 
> 
> The div would be treated as if @aria-role="checkbox" had been 
> specified. [...]


You move the aria behaviour out of HTML and into CSS.

I keep everything inside HTML, with a special syntax for the 
single aria attribute that I proposed (to replace aria-*).

Class="Checked-true" tells a User Agent nothing. Whereas 
aria="checked-true" tells what is needed and gives a hook that can 
replace class="checked-true" entirely and easily - it is DRY [2].

Your proposal would not make the ARIA values available as 
selectors. And some ARIA properties can list more than one ID, and 
it is unhandy to have such info in a separate document.

[1]http://www.w3.org/mid/036901c8c5c9$4d656c50$e83044f0$@com
[2]http://c2.com/cgi/wiki?DontRepeatYourself
-- 
leif halvard silli

Received on Thursday, 5 June 2008 08:13:00 UTC