Re: aria-activedescendant changes to the aria-implementation guide

Hi Andi,

> Please review section 4.3 [2] and let me know if I missed anything. 

First, a comment about this text:

> ... because WAI-ARIA 1.0 does not provide a mechanism to notify an
> application that the DOM has changed.

This wording it makes it sound like WAI-ARIA 1.0 is at fault.  It
isn't.  The reality is that there is no mechanism to notify a webapp
that the DOM has changed.  That deficiency is not specific to ARIA.

How about:

> ... because there is no mechanism to notify an application that the
> DOM has changed.

or:

> ... because there is no reliable mechanism to notify an application
> that the DOM has changed.

Secondly, is there a definition of "activation event" anywhere in the
document?  I think that would help.  I'm still fuzzy on what an
activation event is, although, in last Monday's meeting, James said that
is was essentially a click event.

Thirdly,  role checkbox is missing from the normative statements.  I
looked at the taxonomy to see if it inherits from something that is
listed, but it doesn't appear to.  The inheritance chain from sub to
superclass is: checkbox, input, widget, roletype.  I suspect the rule
that applies to menuitemcheckbox applies to checkbox.

Finally, there are two kinds of button -- push buttons and toggle
buttons.  Toggle buttons have role='button' *and*
aria-pressed='true/false'.  The section currently says that user agents
should not fire an activation event for buttons, but the reader might
think that's for push buttons only.  It might be useful to mention that
the same holds for toggle buttons.  (I assume it does because toggle
buttons are similar to checkboxes and radio buttons, and radio buttons
are listed in the SHOULD NOT section).

Hope that helps.

-- 
;;;;joseph

'I had some dreams, they were clowns in my coffee. Clowns in my coffee.'
                     - C. Simon (misheard lyric) -

Received on Wednesday, 27 April 2011 15:57:57 UTC