Re: opacity, image size, and screen reader behavior

For 3., this would be a CSS validation issue, not an HTML validation 
issue. Whether they should check for this sort of thing or not...not too 
sure it would make a big real-world impact, since visually the opacity 
won't change on the child element, so presumably a (sighted) web 
developer would not do it (by mistake only perhaps) since it wouldn't 
have a discernible visible effect anyway?

P
--
Patrick H. Lauke

www.splintered.co.uk | https://github.com/patrickhlauke
http://flickr.com/photos/redux/ | http://redux.deviantart.com
twitter: @patrick_h_lauke | skype: patrick_h_lauke

------ Original Message ------
From: "Phill Jenkins" <pjenkins@us.ibm.com>
To: "WAI Interest Group" <w3c-wai-ig@w3.org>; public-act-r@w3.org
Sent: 26/07/2021 19:48:53
Subject: opacity, image size, and screen reader behavior

>opacity is an attribute that is often used to hide a UI element from 
>sighted users.
>This is equivalent (or similar) to the 'height' case: an element with 0 
>px height is hidden, but 1 px and greater may be 'visible.'  see 
>discussion on 
>https://stackoverflow.com/questions/9056855/how-does-css-opacity-affect-accessibility
>      <div style='opacity:0;'><img src="about:blank" /></div>
>
>Some informal screen reader testing shows that:
>- VoiceOver on Mac Desktop: opacity:0 is ignored, but image 'alt' is 
>announced if present.
>- VoiceOver on Mobile:  opacity:0 and smaller than 5x5 are ignored, but 
>image greater than 5x5 px 'alt' is announced if present.
>- JAWS & NVDA on Windows Desktop: opacity:0 is ignored, but any images 
>greater than 10x10 px 'alt' is announced if present
>
>1. Is there an updated/maintained reference that documents this 
>behavior for combinations of screen readers with browsers with opacity 
>and/or image size?
>2. Is there a list of best practices that are recommended?
>
>3. Also, the HTML spec says opacity cannot be overridden by a child 
>element. Some accessibility checkers do not check for invalid attempt 
>of child elements trying to override parent, should the checkers 
>(and/or code validators) call out this invalid HTML? not sure if it is 
>needed - your thoughts?
>
>__________
>Regards,
>Phill Jenkins
>See the new IBM Equal Access toolkit and accessibility checker at 
>www.ibm.com/able <https://www.ibm.com/able>
>pjenkins@us.ibm.com
>Accessibility Executive
>IBM Accessibility
>
>