Re: the market hasn't spoken - it hasn't bothered to listened [was Re: fear of "invisible metadata"]

Yes, you can actually do things like <span role="wairole:menuitem"> or 
<div role="main">

We haven't told the world about that since it's not really legal syntax, 
but we allow it. If HTML 5 legalized it, that would be great.
Unfortunately you can only use XHTML and wairole with it this way -- the 
"wairole" prefix is accepted to mean the standard WAI-ARIA roles, 
without needing an xmlns which wouldn't have been possible.

For me, reverse engineering means playing detective and spending a lot 
of time on frustrating little issues. Hopefully that would be minimal, 
and any implementors would instead get most of what they need out of the 
information I'm more than willing to share about the Firefox 
implementation. I do think running both automated and end-user tests and 
comparing output would be a useful way to harmonize implementations.

- Aaron

Simon Pieters wrote:
> On Wed, 27 Jun 2007 14:53:51 +0200, Aaron Leventhal 
> <aaronlev@moonset.net> wrote:
>
>>
>> Why do you say role has to be reverse engineered?
>>
>> W3C docs:
>> http://www.w3.org/TR/aria-role/
>>
>> Firefox docs:
>> http://developer.mozilla.org/en/docs/ARIA:_Accessible_Rich_Internet_Applications 
>>
>
> While I haven't looked into it in depth, I would expect that reverse 
> engineering the implementation in combination with reading the 
> documentation would lead to more interoperable implementations (and 
> find bugs in the documentation) than merely implementing the spec.
>
>> Also, role can be used in HTML in Firefox.
>
> Oh, sure, you can use MathML and SVG in HTML in Firefox too, if you 
> rebuild the DOM afterwards with scripting... Or am I missing 
> something? Can namespaced role=""s be used text/html in Firefox 
> without scripting?
>
>> [...]
>

Received on Wednesday, 27 June 2007 19:30:26 UTC