Re: [g-a-devel] Do we really need 'press' and 'release' AtkAction types in addition to 'click'?

[Joanmarie Diggs <joanied gnome org>]

Hey Peter.

On Mon, 2011-06-20 at 13:11 -0700, Peter Korn wrote:
> Joanie,
> 
> "pressed" should be equivalent to a mouse-down-but-not-released, which
> often triggers visual feedback from the widget.  At least for
> programmatic testing, this may be useful to test for.

To play devil's advocate: If that is the *only* reason we include it,
does it really belong?

In other words, I think it's totally cool that there is so much overlap
between what AT users require in order to use their computers, and what
automated testing frameworks need: It means the latter group does not
have to reinvent the wheel. And it means there are more eyes looking at
our a11y support.

But I personally do not think it's "our job" to support things we have
no use for just because a testing framework might want it.

>   I don't have an immediate accessibility use case to hand...

Yeah, I cannot come up with one either. If no one else can, my vote is
to toss these out. It's hard enough to get everything AT users require
implemented correctly, not to mention consistently across toolkits. Why
muddy the waters with things AT users don't need?

--joanie