RE: aria-hidden and HTML hidden attribute (resurrecting an old issue)

Cynthia Shelly wrote:
>  We need to decide on the correct behavior for the markup below:
>
> HTML5accessibility.com test asserts that the paragraph (id=1) and the 
> button (id=2) should be visually hidden, but available in the AAPI. Does
> the ARIA WG agree?
>
> <p hidden aria-hidden="false" id=”1”>one</p>
> <input id="2" type="button" hidden aria-hidden="false" />

+1


>
> What about aria-hidden=false objects that are children of hidden or
> display:none? Should the paragraph (id=4) be in the accessibility tree?
>
> <div hidden id=3>two<p aria-hidden=false id=4>three</p>four</div>

This seems a bit tougher to me, because the parent (@hidden - an HTML5
attribute) which contains the aria-hidden child doesn't specifically state
it's impact with regard to children's states.
(https://www.w3.org/TR/html5/editing.html#the-hidden-attribute) I don't see
any harm however in suggesting the answer is true, in keeping with the first
two examples.

<div style="display:none;" id=5>seven<p aria-hidden=false
id=6>nine</p>four</div>  

Regarding display:none; the CSS Display Module Level 3 WD
(https://www.w3.org/TR/css-display-3/#box-generation) says ="none" means
"The element and its descendants generates no boxes.", which might suggest
that there is no box for the paragraph text of (id=6) to be hidden and
clipped into (?? overflow ="hidden" ?)
(https://www.w3.org/TR/CSS2/visufx.html#overflow)  and thus perhaps no, but
I think I may be overthinking this :-) 

(I also suspect the intent of an author here would be with the expectation
of true, if we accept the other 3 statements as true)

JF

Received on Thursday, 7 April 2016 04:14:07 UTC