RE: Question about aria-roledescription in the 1.1 spec

" But, you are correct.  Use of this attribute requires QA.  But, that's expected since the attribute is defined as, "... a human readable, author localized description of the role ...".  It's not possible to automate testing of human friendly strings that purport to describe the role more meaningfully than the role value itself."

I agree, and I know it's too late to do anything about this though my concerns already were noted when we spoke of this attribute months ago.

What I'm trying to do now is figure out what guidance to tell people when asked about it.

So a couple of things, the attribute says that the attribute overrides the perceived role, however Freedom Scientific has already told Rich (from what he said months ago please correct if wrong) that they will not stop the announcement of the native role within JAWS because this would be too confusing for their users regardless what the spec says.

So if I'm not mistaken, here we already have the most widely used screen reader saying that they will not do this, and some like NVDA and VoiceOver probably will suppress the native role, not to mention other discrepancies by other venders, so it's already forecasted to not work reliably before the horse is even out of the gate. So some ATs will suppress the role and others won't, and double speaking and redundant feedback for the use of this role will become common place for AT users, meaning that some developers will apply the role based on the feedback they get from the AT they are using, which is guaranteed not to be consistent.

At this time the only thing from a general guidance perspective I can say is don't use it at all because it's not going to be reliable.


Bryan Garaventa
Accessibility Fellow
SSB BART Group, Inc.
bryan.garaventa@ssbbartgroup.com
415.624.2709 (o)
www.SSBBartGroup.com


-----Original Message-----
From: Joseph Scheuhammer [mailto:clown@alum.mit.edu] 
Sent: Friday, September 30, 2016 10:27 AM
To: Bryan Garaventa <bryan.garaventa@whatsock.com>; Accessible Rich Internet Applications Working Group <public-aria@w3.org>
Subject: Re: Question about aria-roledescription in the 1.1 spec

On 2016-09-30 12:44 PM, Bryan Garaventa wrote:
> The use of aria-roledescription however is impossible to validate, because it can literally include anything, making it impossible to detect until after it hits production.
>

There are a couple of issues that can be validated:

  * the value should not be the empty string nor only whitespace characters.
  * User agents do not expose aria-roledescription if used on an element
    that does not have a valid role.

But, you are correct.  Use of this attribute requires QA.  But, that's expected since the attribute is defined as, "... a human readable, author localized description of the role ...".  It's not possible to automate testing of human friendly strings that purport to describe the role more meaningfully than the role value itself.  Perhaps, over time, certain role descriptions will become the norm for given roles in a context, and one could automatically check those descriptions.


--
;;;;joseph.

'Die Wahrheit ist Irgendwo da Draußen. Wieder.'
                 - C. Carter -

Received on Friday, 30 September 2016 18:06:37 UTC