Re: ARIA roles added to the a element should be conforming in HTML5.

2009-10-22 17:18, Leif Halvard Silli skrev:
> Well, HTML doesn't allow ARIA yet, so that's fine, isn't it? ;-) It
> would also be interesting to check whether people actually use CSS in
> order to misuse <h1> ...

I have seen it in misguided SEO... I was a while ago, though.

But this is getting off topic.

When I started out developing I was warned about browser bugs and 
inconsistencies for button. I do not know if it was correct advice, 
since in those early days I tended to believe everything... True or not, 
such advice tend to live on many years beyond their usefulness*

The a-element actually has been chosen for accessibility reasons in many 
cases, since it was the only way authors (like me) who distrusted 
buttons, could make an object keyboard focus-able. We did not get 
tabindex=0 until recently...

Basically Stevens fear is that while old content can be improved by 
evangelizing ARIA, such evangelizing efforts will be hindered if ARIA is 
seen as a problem, when validating.

I do not share his fear, since I believe that authors who are reachable 
for evangelism are also reachable for "proper semantics" evangelism. And 
sticking to my taking a teachers perspective on things I think that the 
use of proper button elements should be encouraged, and that 
element-ARIA-role mismatches errors might be a learning aid.

Let me just repeat my two conditions that I have stated separately in 
other e-mails:

1. The validator warning must be carefully phrased to not discourage ARIA.

2. Turning links into buttons for hijax or other uses of scripts as 
progressive enhancement might be a valid use case, which means that a 
dynamically added ARIA role of button should be allowed, even when it is 
not allowed to be hard coded into the HTML.


-- 
Lars Gunther
http://keryx.se/
http://twitter.com/itpastorn/
http://itpastorn.blogspot.com/


* I still see people recommend that one puts HTML comments around CSS 
code or JavaScript, to "hide it from older browsers..."

Received on Thursday, 22 October 2009 23:09:05 UTC