- From: Alexander Surkov <surkov.alexander@gmail.com>
- Date: Wed, 24 Dec 2014 09:45:39 -0500
- To: Bryan Garaventa <bryan.garaventa@ssbbartgroup.com>
- Cc: "W3C WAI Protocols & Formats" <public-pfwg@w3.org>
- Message-ID: <CA+epNsevZSZJtVT1QCvS1CDY9rk===euVup7n0R3Tu5GWFCaoA@mail.gmail.com>
Hi, Bryan. Firefox implementation on this issue is controversial, for example <input type="checkbox" aria-disabled="true" aria-checked="true"> is exposed as disabled, not checked checkbox I think. Personally I think ARIA should yield to native markup when in conflict. Thus I wouldn't fix <input type="checkbox" aria-checed="mixed"> because it would conflict with HTML5 intermediate checkboxes. I'm curious if spec agrees on it. Thanks. Alexander. On Mon, Dec 22, 2014 at 6:50 PM, Bryan Garaventa < bryan.garaventa@ssbbartgroup.com> wrote: > Hi, > This came up for me today, and I'm not sure what the proper answer for > this is. > > I understand that native host semantics typically trumps ARIA markup for > form fields. > E.G Using readonly instead of aria-readonly, using disabled instead of > aria-disabled, using checked instead of aria-checked... > > Now, in the case of a standard checkbox control, the native attribute, > checked, has no native tristate value since it is a boolean. > > So, is it actually valid to use aria-checked='mixed' on a native checkbox > control that includes the checked attribute? > > E.G > > <input type="checkbox" checked aria-checked="mixed" title="Control Name" /> > > Which, by the way, doesn't actually work in IE or FF. > > Only the following does reliably: > > <span role="checkbox" aria-checked="mixed" aria-labelledby="mcb1" > tabindex="0" class="mixedStateCheckbox" ></span> > <span tabindex="-1" id="mcb1"> Control Name </span> > > Thanks, > Bryan > >
Received on Wednesday, 24 December 2014 14:46:06 UTC