See also: IRC log
<trackbot> Date: 14 January 2014
<bgaraventa1979> Bryan Garaventa
Update re: PF decision HTMl5 @hidden plus aria-hidden='false' (https://www.w3.org/Bugs/Public/show_bug.cgi?id=18574)
<MichaelC> http://www.w3.org/2001/12/zakim-irc-bot.html
<scribe> scribenick: clown
(https://www.w3.org/Bugs/Public/show_bug.cgi?id=18574)
JS: Mostly for Cynthia's
benefit.
... We discussed the relationship between aria-hidden and HTML5
hidden attributes
JS See bugzilla: https://www.w3.org/Bugs/Public/show_bug.cgi?id=18574
JS: @hidden should have a weak
mapping to aria-hidden=true
... It was decided that there is a weak mapping.
... related outcome is this construct:
<div hidden aria-hidden="false">
JS: That means it is not visible, but is exposed at the a11y API level.
/
CS: People shouldn't be adding
things to the a11y API.
... I would call this a bad practice.
... It's valid HTML, but is open to discussion.
JS: Does Microsoft support this?
CS: Because it works in IE
doesn't mean we think it is a good technique.
... This is an author error.
http://www.html5accessibility/tests/hidden2013.html
CS: I would caution authors
against doing that.
... Unless there is a really good reason to do that.
JS: It functions the same way as offscreen content: not seen, but spoken by a screen reader.
CS: I think that offscreen is a
bad idea as well.
... It causes problem in the text only version.
... Also, the main stream tester doesn't see it and it isn't
tested.
BG: In general I agree that
offscreen text is a bad idea, but sometimes it is very
helpful.
... A aria table with editable grid cells, sometimes a
string.
<cyns> It causes the same problem as the old 'text only' pages. Adding a separate UI for screen-readers that is not visible to mainstream developers and testers runs the risk of that UI not being kept up to date, tested regularly. Also for it to break in production without anyone noticing
BG: the only way to communicate to an SR user in a cross platform way was to use offscreen text.
<cyns> That's true of both offscreen and aria-hidden
BG: Don't the html5 hidden and aria-hidden=false cancel each other.
<cyns> It should be a last resort technique, but many developers go to it too quickly.
JS: So, the hidden attribute
means it's not rendered on screen and not seen, but the
aria-hidden=false means it is nonethelss exposed to the a11y
API.
... I think this changes the meaning of aria-hidden to
something like "expose/don't expose" at a11y api level.
<cyns> Issue 2: <div hidden aria-hidden=false style='display:none'> each of these methods of hiding things from screen readers is fine on its own, and I'm ok with all of them existing. Using more than one on the same element is likely to introduce bugs, and authors should be cautioned against doing so.
JS: the original meaning had something about "perceivable" in it.
CS: the original specification of hidden was that it was display:none.
JS: But mozilla requested a fuller specification of this technique so they can think about it, and respond.
CS: I would like to see the same in order to respond.
DB: I'm still with Matt with the idea of getting a new attribute for this kind of thing.
JS: agreed that there is a weak mapping. There never was a strong mapping.
CS: I'm fine with that. I'm not sure about the implications.
DB: I agree. For best practices, this should be a fallback.
action-1320?
<trackbot> action-1320 -- Joseph Scheuhammer to Consider mapping the "offscreen" api properties in the situation of aria-hidden="false" on non-rendered elements. -- due 2013-12-23 -- OPEN
<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1320
JS: if we go forward with above
construct, how do we map it to a11y apis.
... I suggested looking into the offscreen properties.
... This is what i found:
https://lists.w3.org/Archives/Member/w3c-wai-pf/2013OctDec/0079.html
JS: MSAA has
STATE_SYSTEM_OFFSCREEN
... UIA has IsOffscreen
CS: But, this isn't a back door
for screen readers.
... It was for small screens where some of the content was
scrolled offscreen.
DB: what about occlusion.
CS: Not the same thing, but there
may be another property for that.
... I'm not sure "offscreen" was for things inside a window,
but for windows that were offscreen.
DB: If it scrolled offscreen is it "offscreen"?
CS: I'm not sure. I have to look.
BG: What about layering?
DB: It should be similar to the occlusion question.
BG: Like in a carousel where you have a layering effect.
DB: Visible by not legible?
... Includes things like translucent, and overlays.
BG: yes.
DB: It's funny that as time goes on there are more contexts that don't fit our original states.
JS: I had difficulty with finding an offscreen state for ATK/AT-SPI
DB: I think that's a combo of STATE_INVISIBLE and STATE_SHOWING.
JS: I believe Alex said something
along those lines.
... It would be clear STATE_INVISIBLE and clear
STATE_SHOWING
zakim take up item 2
JS: <general discussion of agenda for face-to-face>
DB: Is there a call next week?
JS: If we have a call, what will
we talk about?
... There is not meeting next week.
s/ \///
This is scribe.perl Revision: 1.138 of Date: 2013-04-25 13:59:11 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/PAI/API/ Succeeded: s/original state/original states/ WARNING: Bad s/// command: s/a11ybroken.dequecloud.com\/// Succeeded: s/a11ybroken/ / Succeeded: s/a11ybroken/ /g Succeeded: s/.dequecloud/ /g Succeeded: s/.com/ /g Succeeded: s/occulsion/occlusion/ Succeeded: s/html5accessibility /html5accessibility/ Found ScribeNick: clown Inferring Scribes: clown Default Present: verytired_David_Bolter, Joseph_Scheuhammer, Bryan_Garaventa, Michael_Cooper, cyns, Cooper Present: verytired_David_Bolter Joseph_Scheuhammer Bryan_Garaventa Michael_Cooper cyns Cooper Agenda: http://lists.w3.org/Archives/Public/wai-xtech/2014Jan/0005.html Found Date: 14 Jan 2014 Guessing minutes URL: http://www.w3.org/2014/01/14-aapi-minutes.html People with action items: WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option.[End of scribe.perl diagnostic output]