W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > January to March 2013

Re: ARIA role restrictictions in HTML5

From: Steve Faulkner <faulkner.steve@gmail.com>
Date: Thu, 21 Mar 2013 09:42:33 +0000
Message-ID: <CA+ri+VkOwo_tHNmg3N+t2AbcQPqUb=RYa0CY=27zOvKf_7TDeg@mail.gmail.com>
To: Bryan Garaventa <bryan.garaventa@whatsock.com>
Cc: David Ashleydale <david@randomlife.com>, Jonathan Avila <jon.avila@ssbbartgroup.com>, w3c-wai-ig@w3.org
Hi Bryan,

ARIA is implemented in browsers, once implemented features are very rarely

The hope is that with the addition of new native HTML controls in HTML5,
developers will use these instead of rolling their own, but from experience
we know that even basic controls such as buttons that have been around
since HTML 2.0 are still not used in some circumstances (For example the
Gmail UI contains  literally hundreds of controls but the vast majority are
built from divs and spans with ARIA added).

So while it is obviously better that native controls are used as they have
role,state, property information built in as well as keyboard interaction,
this is often not the case.

Also it should be noted that HTML5 fills in some of the gaps, but does not
fill in all of the gaps and that will be the case for the forseeable future.

I wrote this a few years ago (needs updating), but it is still the case:
HTML5 and the myth of WAI-ARIA redundance

with regards

HTML 5.1 <http://www.w3.org/html/wg/drafts/html/master/>

On 20 March 2013 19:29, Bryan Garaventa <bryan.garaventa@whatsock.com>wrote:

> **
> My concern is that ARIA works well right now when properly implemented.
> Many others have thought the same, and have implemented  ARIA within web
> applications across the web, not just on single page implementations, but
> have built ARIA support into CMSs as well.
> In short, ARIA is now entrenched in the web, and it will likely never
> leave it regardless what the standards are.
> So it would be good to know whether ARIA recognition will ever be pulled
> out of browser and Assistive Technology support, sort of like pulling out
> the rug from all of these implementations, making previously accessible
> components suddenly inaccessible?
> ----- Original Message -----
> *From:* David Ashleydale <david@randomlife.com>
> *To:* Bryan Garaventa <bryan.garaventa@whatsock.com>
> *Cc:* Jonathan Avila <jon.avila@ssbbartgroup.com> ; w3c-wai-ig@w3.org
> *Sent:* Wednesday, March 20, 2013 12:07 PM
> *Subject:* Re: ARIA role restrictictions in HTML5
> It's funny -- I always thought that ARIA would be kind of a preview for
> HTML 5. That the ARIA attributes would become part of the HTML spec.
> But it doesn't seem to be turning out like that.
> David
> On Wednesday, March 20, 2013, Bryan Garaventa wrote:
>> **
>> I agree, anchor elements work well for this purpose, especially for
>> graceful degradation with radio buttons.
>> Regarding buttons, I often see A tags styled as buttons for form
>> submission elements.
>> Not having the ability to put role="button" on such elements to aid
>> screen reader interaction, would impair accessibility, not enhance it. The
>> same is true for Toggle Buttons, and Checkboxes.
>> A tags are also used for Listbox Option elements, which is also used to support
>> graceful degradation.
>> This brings me to a question I've been wondering about.
>> Is HTML5 supposed to replace ARIA, or will they work together? In other
>> words, will components built using current standards compliant ARIA still
>> be valid ten or twenty years from now?
>> ----- Original Message -----
>> *From:* Jonathan Avila
>> *To:* w3c-wai-ig@w3.org
>> *Sent:* Wednesday, March 20, 2013 7:25 AM
>> *Subject:* ARIA role restrictictions in HTML5
>>  I was looking at the latest draft version of the HTML5 specification
>> and noticed in the implicit aria semantics table it indicates that only a
>> limited set of ARIA roles can be used with certain elements such as the
>> anchor element to conform to the HTML5 specification.  Specifically you
>> could not use a role of button, radio button, etc. on anchor elements.
>> This seems problematic but makes good semantic sense.  One advantage of
>> using anchors with hrefs for diverse ARIA roles is there is some
>> progressive enhancement support.
>> http://www.w3.org/html/wg/drafts/html/master/dom.html#sec-implicit-aria-semantics
>> Jonathan
Received on Thursday, 21 March 2013 09:43:47 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 13 October 2015 16:21:47 UTC