PFWG-ISSUE-554 (Was: aria-invalid is confusing)

Since aria-invalid is a global attribute, there is no difference between the default value and undefined, except in the case of inheritance.

FWIW, this would represent a change in the way we do value inheritance though. To my knowledge, only aria-relevant has its value inherited from an ancestor element, though this would be useful for a few other attributes, including aria-required and aria-invalid. I've raised a WG issue for ARIA 1.1. If we inherit values though when the attr is undefined on an element, then the default value only would only apply to the root element in the document… I think.

ISSUE-554
https://www.w3.org/WAI/PF/Group/track/issues/554


On Feb 4, 2013, at 5:13 PM, Alexander Surkov <surkov.alexander@gmail.com> wrote:

> Hi, James.
> 
> It seems it's usual practice to ignore the attribute having empty or
> undefined value as it wasn't presented at all. Otherwise it looks
> good.
> Thank you.
> Alex.
> 
> 
> On Tue, Feb 5, 2013 at 5:17 AM, James Craig <jcraig@apple.com> wrote:
>> Would this edit resolve the confusion?
>> 
>> From:
>> "if the attribute is not present, or its value is false then the default value of false applies."
>> 
>> To:
>> "if the attribute is not present, then the value is inherited from the parent element."
>> 
>> 
>> On Feb 4, 2013, at 2:43 AM, Alexander Surkov <surkov.alexander@gmail.com> wrote:
>> 
>>> Hi. The spec says
>>> (http://www.w3.org/WAI/PF/aria/states_and_properties#aria-invalid):
>>> 
>>> "For future expansion, the aria-invalid attribute is an enumerated
>>> type. Any value not recognized in the list of allowed values MUST be
>>> treated by user agents as if the value true had been provided. If the
>>> attribute is not present, or its value is false, or its value is an
>>> empty string, the default value of false applies."
>>> 
>>> In particular this means that
>>> 
>>> <div aria-invalid="true">
>>> <span>1st span</span>
>>> <span aria-invalid="false">2nd span</span>
>>> </div>
>>> 
>>> Both "1st span" and "2nd span" text ranges should have "invalid:false"
>>> text attributes because "if the attribute is not present, or its value
>>> is false then the default value of false applies". It doesn't seem
>>> this behavior in the context of the "1st span" text range is evident
>>> and actually wanted.
>>> 
>>> Thank you.
>>> Alex.
>>> 
>> 
>

Received on Tuesday, 5 February 2013 01:37:02 UTC