- From: Alexander Surkov <surkov.alexander@gmail.com>
- Date: Thu, 29 Jan 2015 14:18:19 -0500
- To: Richard Schwerdtfeger <schwer@us.ibm.com>
- Cc: Joanmarie Diggs <jdiggs@igalia.com>, "W3C WAI Protocols & Formats" <public-pfwg@w3.org>
- Message-ID: <CA+epNsd09nzpDmsa=Ja6-zobwvN8C7ewmzH=EPp8oCoK=+rGgQ@mail.gmail.com>
Hi, Rich. I believe that everything but "false" is mapped to INVALID state on role="checkbox", but you are right and we do a check when aria-invalid is exposed as text attribute, so if the value is "grammar" then we expose "grammar", if "foo" then "true". My concern is the whitelist approach makes the browser to do extra work, which probably is needed by no one. This concern of course is applicable to aria-invalid of text containers. If we have a chance to improve things now then why we'd need to follow the old pattern? No whitelist approach is more flexible and performant. If you add a new value into the spec later then browser implementation doesn't need any single change, it just works. Anyway AT will process values it supports and will ignore everything else (treating them as 'true' value). Thanks. Alex. On Thu, Jan 29, 2015 at 1:31 PM, Richard Schwerdtfeger <schwer@us.ibm.com> wrote: > In response to your first paragraph you already do this for aria-invalid: > > > *https://www.w3.org/WAI/PF/testharness/testresults?testsuite_id=1&testcase_id=587* > <https://www.w3.org/WAI/PF/testharness/testresults?testsuite_id=1&testcase_id=587> > > Why would we do things differently for aria-current? > > > Rich Schwerdtfeger > > [image: Inactive hide details for Alexander Surkov ---01/28/2015 08:01:09 > AM---I do not support this bit: "Any value not included in th]Alexander > Surkov ---01/28/2015 08:01:09 AM---I do not support this bit: "Any value > not included in the list of allowed values must be treated by > > From: Alexander Surkov <surkov.alexander@gmail.com> > To: Joanmarie Diggs <jdiggs@igalia.com> > Cc: "W3C WAI Protocols & Formats" <public-pfwg@w3.org> > Date: 01/28/2015 08:01 AM > Subject: Re: [REVIEW REQUEST] Action-1548: Add aria-current attribute > ------------------------------ > > > > I do not support this bit: "Any value not included in the list of allowed > values must be treated by *user agents* > <http://rawgit.com/w3c/aria/master/aria/aria.html#dfn-user-agent> as if > the value true had been provided" because it forces the browser to > validate the value and that's performance downgrade even if there's no > consumer for aria-current. So I would exclude words "by user agents" or > added "or assistive technologies" instead. Also I would replaced "must" to > "should" to make the attribute value list extensible. > > Also I think I don't get aria-current listbox example, in particular this > part " whereas *aria-selected* > <http://rawgit.com/w3c/aria/master/aria/aria.html#aria-selected> would be > used to indicate the page of results the user had selected to be displayed > next". Also it would be good to explain connection with focus. Is there any > known app on the web implementing this use case? > > Thanks. > Alex. > > On Thu, Jan 22, 2015 at 4:48 PM, Joanmarie Diggs <*jdiggs@igalia.com* > <jdiggs@igalia.com>> wrote: > > Hi all. > > The aria-current attribute is now in master. Please review. Thanks! > Done: *http://rawgit.com/w3c/aria/master/aria/aria.html#aria-current* > <http://rawgit.com/w3c/aria/master/aria/aria.html#aria-current> > > --joanie > > >
Attachments
- image/gif attachment: graycol.gif
Received on Thursday, 29 January 2015 19:18:48 UTC