Re: ARIA semantics for secondary navigation

On May 20, 2009, at 1:46 AM, Simon Pieters wrote:

> On Wed, 20 May 2009 09:40:30 +0200, James Craig <jcraig@apple.com>  
> wrote:
>
>> On May 19, 2009, at 11:44 PM, Schnabel, Stefan wrote:
>>
>>> Who says that a text node referenced by "aria-labelledby" MUST/ 
>>> SHALL/HAS
>>> TO be visible?
>>
>> No one, but there isn't much point in using aria-labelledby if you're
>> gonna hide the label.
>
> I can think of a couple:
>
>  * If the element is a heading element, the heading would still be  
> part of
>    the document outline and thus help people who navigate the  
> document by
>    headings instead of by landmarks.

True, but only if you use a certain modalilty-specific methods of  
hiding the content. For example, positioning it off-screen would still  
allow accessible heading navigation, but using the CSS 'display:none'  
or 'visibility:hidden' would not.

>  * For users who don't use software that exposes the ARIA  
> information and
>    also doesn't apply CSS (e.g. Lynx), the label would be just as  
> useful
>    as for those to whom the ARIA information is exposed.

Does anyone regularly use CSS-incapable browsers like Lynx anymore?  
It'd be nice to see some numbers.

Another argument might be that a user has a user style sheet that  
always exposes heading, but I there isn't any data I know of that  
suggest many people have set up user styles at all. Much less in this  
way.

>  * Search engines who don't look at ARIA markup would still get the  
> label.

Some search engines may considering it cloaking. Even so, search  
engines are getting smarter and smarter every day, and the hidden  
headings of navigation or other landmark sections aren't what they are  
looking to index. Unless you have an actual product called "main  
menu," who cares if your site shows up in search engine results for  
"main menu." That's meta data for the user, not the search engine.

Received on Wednesday, 20 May 2009 21:44:57 UTC