Re: From the HTML-WG about aria-hidden

Hi Leif,

A couple of quick points.  On 12-02-15 7:32 PM, Leif Halvard Silli wrote:
> ...
>>> With respect to aria-describedby the above proposes that
>>> aria-describedby trumps aria-hidden,
> It is ARIA 1.0 that causes aria-hidden to be trumped:
> 'Skip hidden elements unless the author specifies to use them
>   via an aria-labelledby or aria-describedby.'
> <>

One distinction here:  I was discussing whether the hidden element is 
exposed in the a11y tree.  The alternative text computation relates to 
whether hidden elements are included in the calculation of, well, 
alternative text.  The two are not the same thing.

For example:
<img aria-describedby="desc1" ...>
<p id="desc1" "aria-hidden="true" style="display:none;">Descriptive text</p>

There will be one accessible object in the a11y tree corresponding to 
the image, and that accessible will have a *description property* 
containing the string "Descriptive text".  That's because the hidden 
paragraph is not skipped here due to the description relationship, even 
though it is hidden.

However, there will be no accessible object in the a11y tree 
corresponding to the paragraph itself, because it is hidden.  Put 
another way:  the DOM contains two objects, one for the image and one 
for the paragraph.  The a11y tree contains only one object, for the image.

The issue I was exploring was whether the aria-describedby relationship 
forces an accessible object for the hidden paragraph into the a11y tree, 
even though it is hidden.

>>> and forces the rendering of the hidden element.
> No. Not rendering. It is just included in the 'accessible name' - like
> @alt text.

You're right; my mistake.  I meant to say "and forces the *exposure* of 
the hidden element",  in the sense of exposing an accessible object in 
the a11y tree.


'A: After all, it isn't rocket science.'
'K: Right. It's merely computer science.'
              - J. D. Klaun -

Received on Thursday, 16 February 2012 15:06:20 UTC