Re: aria-describedat

On Thu, Mar 29, 2012 at 12:59 PM, david bolter <david.bolter@gmail.com> wrote:
> On Wed, Mar 28, 2012 at 5:05 PM, John Foliot <john@foliot.ca> wrote:
>
> Hi John.
>
> I thought of you jovially when I wrote my email :) The idea for a new name
> is to avoid the baggage/politics, stop energy, lack of usage/support, or
> whatever has prevented 'longdesc' from blossoming in the past as well as to
> make it sound (and be) more general purpose. I note Silvia might think this
> is a weakness and that's fair. Regarding the ARIA route, my biggest concern
> with having an ARIA attribute affect browser UI for everyone out-of-the-box
> is that I believe part of the beauty of ARIA is that it is purely annotative
> semantics that can be added to describe existing UI without interfering with
> that UI. This is a nice promise to make to people who toil to make custom
> web UI work across browsers. I spent a year doing this in dijit (dojo UI)
> and adding ARIA (my primary role) never broke anything thank goodness.
> Ultimately I personally think this helps uptake of ARIA and I'm aware we
> don't all agree. In my world ARIA is a non disruptive way to make custom web
> UI accessible. I think the behavior we want for something like @duck_soup
> should live as an attribute outside ARIA territory.
>
> If the majority think we should open ARIA up to driving web UI for everyone,
> and break the promise above I think we'll be making a mistake.

Just a note: you've made me change my mind. I agree, the new attribute
needs to be a HTML attribute and not aria. Browser should additionally
expose the link to a11y APIs and thus to screenreaders, but since it
introduces user interaction in the browser for everyone, it should
indeed be a browser attribute.

Cheers,
Silvia.

Received on Thursday, 29 March 2012 20:33:22 UTC