Accessible Rich Internet Applications Working Group Teleconference

25 Feb 2016

See also: IRC log


Deborah_Kaplan, JaEunJemmaKu, Janina, Joanmarie_Diggs, Joseph_Scheuhammer, Rich_Schwerdtfeger, fesch, Michiel_Bijl, matt_king, LJWatson, MichaelC, Bryan_Garaventa


<richardschwerdtfeger> meeting: W3C ARIA Working Group Meeting

<richardschwerdtfeger> https://lists.w3.org/Archives/Public/public-aria/2016Feb/0398.html

<scribe> Scribe: Stefan

<richardschwerdtfeger> https://lists.w3.org/Archives/Public/public-aria/2016Feb/0398.html

<richardschwerdtfeger> https://lists.w3.org/Archives/Public/public-aria-admin/2016Feb/0007.html

issue 1009 CFC

<richardschwerdtfeger> https://lists.w3.org/Archives/Public/public-aria-admin/2016Feb/0007.html

Rich: any input?

Joanie: Joseph proposed change, was that made?
... should I wait?

<clown> is this the url: https://rawgit.com/w3c/aria/issue1009/aria/aria.html#aria-details

Rich: go pull it

<clown> ?

Joanie: I will merge

<clown> https://rawgit.com/w3c/aria/issue1009/aria/aria.html#aria-details

Rich: OK I made the change
... next thing is heartbeat drafts

<clown> action-2031

<trackbot> action-2031 -- Richard Schwerdtfeger to Update ARIA 1.1 Abstract and Introduction to accommodate SVG in addition to HTML -- due 2016-03-01 -- PENDINGREVIEW

<trackbot> http://www.w3.org/WAI/ARIA/track/actions/2031

Rich: Michael 2031 and merge it?

Michael: i'm not ready yet

<clown> https://rawgit.com/w3c/aria/action2031/aria/aria.html

Rich: when is the date?

Michael: have to look to editors minutes

<clown> https://lists.w3.org/Archives/Public/public-aria/2016Feb/0383.html

Michael: not this week .. next week
... by March 11th

<MichaelC> Current publication timelines

Rich: Joanie ready next week?
... we have dpub graphics etc. should we put all in there?

Michael: q for the group, personal opinion is better to point to intermediate resource, reccommended: aria intro

<MichaelC> ARIA Intro

Rich: anybody issue with it?

<richardschwerdtfeger> https://rawgit.com/w3c/aria/action2031/aria/aria.html

Still action 2031

Joseph: theres a huge list of docs ...

Here were no objections to Rich's question

Matt: people needs to be aware of aria versions there are looking at

<MichaelC> 1.1 specific part of the ARIA intro

Rich: some refs will be removed
... we need cFC for non-normative text .. but action is closed .. objections? None

<clown> https://rawgit.com/w3c/aria/action2031/aria/aria.html#ua-support

Rich: section 1.3 has old link to imp guide

<clown> https://rawgit.com/w3c/aria/action2031/aria/aria.html#at_support

Rich: needs some polishing -> action to Michael

<richardschwerdtfeger> ACTION: Michael Search for old links and update them in the ARIA 1.1 spec. [recorded in http://www.w3.org/2016/02/25-aria-minutes.html#action01]

<trackbot> Created ACTION-2033 - Search for old links and update them in the aria 1.1 spec. [on Michael Cooper - due 2016-03-03].

<richardschwerdtfeger> ACTION: Rich Remove reference to UAAG and reference EO ARIA 1.1 Suite of documents [recorded in http://www.w3.org/2016/02/25-aria-minutes.html#action02]

<trackbot> Created ACTION-2034 - Remove reference to uaag and reference eo aria 1.1 suite of documents [on Richard Schwerdtfeger - due 2016-03-03].

RESOLUTION: close 707 and action 2031

<clown> issue-707

<trackbot> issue-707 -- Update ARIA 1.1 Abstract and Introduction -- open

<trackbot> http://www.w3.org/WAI/ARIA/track/issues/707

ARIA Test Coordination Call to

ARIA Test Coordination Call #

<clown> http://doodle.com/poll/g4kixuskyfvgmm24hakmvpig/admin#table

Michael: checking the invites ..

Rich: we can do this by email
... after the call
... there is a separate list

<MichaelC> Webex for next monday┬┤s meeting: https://mit.webex.com/mit/j.php?MTID=m7e86f0230b1b7a09167aa46b4be2800a meeting number 647 399 010

Michael: forward to testing group


1490: Combobox (20 minutes *max* discussion) Matt

<clown> action-1490

<trackbot> action-1490 -- Matthew King to Propose spec text edit for issue-610: comboboxes should allow complex children elements -- due 2016-02-03 -- OPEN

<trackbot> http://www.w3.org/WAI/ARIA/track/actions/1490

<mck> <div role=combobox aria-owns=list> <span role=textbox> </span> </div> <ul id=list role = listbox> </ul>

Matt: we have consesus around 1.1 structural for combo
... composite widget, has a textbox, popoup element shown as listbox and siblings
... clear to NVDA and FS on board from API perspective
... advantages for 1.1 structure: 1. you can determine value of combo since there is textbox element 2. screen reader has ability to make both parts seprately percievable

<mck> <div role=combobox aria-owns=list> <span tabindex=0 role=textbox>value being edited </span> </div> <ul id=list role = listbox> focusable options with tabindex -1 </ul>

Matt: next part of discussion: focus managenment and ara-activedescendant: if focus is on textbox (DOM focus)

<mck> <div role=combobox aria-owns=list> <span tabindex=0 aria-activedescendant=opt1 role=textbox> </span> </div> <ul id=list role = listbox> <li id=opt1 tabindex=-1> </li> </ul>

Matt: when listbox expanded and activedescendant: hopeully will be agreement tthat this above will be ok
... : difference between 1.0 and 1.1 now descendat of parent within the tree .. this actually works in all browseers .. questuiion is formalizing this when textbox in combo has focus and uses aria-active descendant

Rich: you have owns on combo but activedescendant on textbox?

Matt: right

Cynthia: what about aruia-controls to relate the parts

Matt: this does the composite container concept
... activedescendant is on textbox because of DOM focus
... is also therer
... this way it works

Joseph: it works by accident
... it creates isse for what's is focusabe
... descendendt of descendants doesnt happen here

<bgaraventa1979> +q

Cynthia: do we need this?
... then we need to rewirte aria-activedescendant definitrion

Rich: aria-controls on combobox that controls textbox

Matt: messy from authoring perspective

<fesch> +1 to it being messy for authors

Matt: by this pattern we know what behaviour we want from combo box
... author should NOT do work of user agent

Rich: descendant is descendant of item that has focus
... textbox on behalf of combo has focus

Matt: one of two pssibilities: ativedescendant on textbox is proxy for combo

<clown> <div tabindex=0 role=combobox aria-owns=list aria-activedescendant> <span role=textbox>value being edited </span> </div> <ul id=list role = listbox> focusable options with IDs </ul>

Matt: other thing is with DOM focus
... both are equally good .. from semantic perspective .. for implementation I prefer first way
... activedescendant on textbox

Brian: seen many implementations .. you cannot guarantee parent-child relationships on combos .. style trums everything

Matt: : this structure fixes the problem

Brian: aria-controls will better fix that

Matt: bad idea frm API perspective saays AT developers

Cynthia: implementation just moves the focus

<clown> http://w3c.github.io/aria/aria/aria.html#aria-activedescendant

Matt: AT dwevelopers ignore activedescendant when not in context .. so no problem for them
... in IE, when you mouseover menus, activedescendants is pointing to elements within menu and not related to elememt with dom focus .. all AT ignoring that

Cynthia: when are they using the DOM?

<fesch> this is a brittle pattern

<clown> <div tabindex=0  role=combobox aria-owns=list aria-activedescendant="option5"> <span role=textbox>value being edited </span> </div> <ul id=list role = listbox> focusable options with IDs </ul>

Cynthia: as UA, I don't want resolve parent-chuild relationships

<Zakim> clown, you wanted to point out that tabindex=-1 on the list options is unneccessary in an activedescedant scenario.

<clown> <div tabindex=0  role=combobox aria-owns=list aria-activedescendant> <span role=textbox>value being edited </span> </div> <ul id=list role = listbox> focusable options with IDs </ul>

Joseph: tabindex=0 on combo and aria-activedescendant on it

Matt: causes lotta issues in edit case

Joseph: not when textbox is a span and JS will put the chars in it

Matt: has huge impact on JS for non-editable content (??)


<Zakim> fesch, you wanted to say this is way to technical and tricky for authors

Fred: tis is a brittle (?) pattern
... sorrta magic

Matt: we are not confused we need to make descision

Rich: for a container as a widget, can we say active child ?
... as long as you have a place in a container widget

<richardschwerdtfeger> ack

Cynthia: UIA Comb box is a container with a edit and a list and expand/collaps pattern

Rich: time is over

<cyns> UIA combobox docs https://msdn.microsoft.com/en-us/library/windows/desktop/ee671622(v=vs.85).aspx

Rich: do people mind if Matt + Joseph change definition of activedescendant?

Cynthia: make the spec match

Matt: update action related to this and bring on an actual revised branch / update based on discussion .. but after next week

Rich: activedescendant for the widget .. you can do and say that

Brian: we should avoid examples where it won't work

Joseph: core AM docs .. looking at it

Rich: moving on

Action 2021: Draft text for default max and min spinbutton values

<richardschwerdtfeger> https://rawgit.com/w3c/aria/action-2021/aria/aria.html#spinbutton

<clown> action-2021

<trackbot> action-2021 -- Joanmarie Diggs to Draft text for default max and min spinbutton values -- due 2016-02-18 -- PENDINGREVIEW

<trackbot> http://www.w3.org/WAI/ARIA/track/actions/2021

Joanie: I added a normative statement for UA's

<joanie> Authors must set the aria-valuemin, aria-valuemax, and aria-valuenow attributes. If missing, user agents must set the values of these attributes using the rules for the equivalent native host language element.

<joanie> Implicit Value for Role: Default for aria-valuemin is the same as the equivalent native host language element.

<joanie> Default for aria-valuemax is the same as the equivalent native host language element.

<joanie> Default for aria-valuenow is the same as the equivalent native host language element

Matt: HTML AAM has mapping?!
... for an author is simple: spec the value

Joseph: spin button in SVG ? Rich: no

Matt: should spec value that exists in real wold

Cynthia: we should implicit role for value

joseph: HTML5 spin button does not require valuemin and max
... can be anything
... temperature max .. can be infinity
... HTML5 is saying ... ya don#t NEED to spec this

Cynthia: UA's say there must be maximum

<Zakim> cyns, you wanted to say the UIA control pattern for combobox has a container with an edit and a list, like Matt's structure

Matt: if the author doesn't spec min/max .. will we get inconsistent behaviour among browsers?

Cynthia: If platform API reqquires integer then we have an issue

JOanie: can we move on because I have more to think about and research, and we have a guest to discuss aria-linktype?

joseph: if platforms require values, the DOM isn't .. acc api is sayin zero .. mismatch

Matt: no SR does minmax yet for spinbutton, for sliders they do

Action 2006: Draft proposal for new aria-linktype role Joanie

Rich: we have discussion with linktypes witrh Jamwes Craig
... then subrole discussions came
... linktype go around is for epub .. do we want that? also in APG?

Deborah: funct. req. is: couple of key things, 1. why taxonomy ? we want to be able to say reading system for different stuff
... role we want make sure that vocab will cause AT make browsers to do the right thing

Rich: can call main role alink, then it is also a glossary reference (typed link)
... subrole in UIA?

Cynthia: annotation control pattern can be used but we don't have subroeles
... we have a proposal for annotations - almost done

Rich: aria is an annotation as well

<cyns> UIA annotation pattern https://msdn.microsoft.com/en-us/library/windows/desktop/hh448769(v=vs.85).aspx

Cynthia: can be mapped

ich: don't think so .. in middle of doc you cannot see eg glossary references

Rich: we have no vehicle witrh remote refs
... you can do localized role description

Cynthia: gets pretty overloaded - may work
... we do this for lotta other stuff

Rich: you can put role link yielding control type link but not properties, right? Cynthia: no

Cynthia: aria-banana will be shoved to aria propertis in UIA API
... what is the proposal?

Rich: role equals glossary refernd
... role=link linktype=foo
... keep separation in the mapping itsself

Matt: if yu make a link with, hoe many different roles? (??)
... different types, biblio, footnote, glossary etc. .. list shold be expandable who does this?

Deborah: we have a comprehensive list and will maintain that (?)

<fesch> Lisa Seeman wanted COGA to coordinate with linktype and shortcuts

<clown> richardschwerdtfeger: ^

<dkaplan3> Curent list is at https://rawgit.com/w3c/aria/action-2006/aria/aria.html#aria-linktype

Rich: we will make sure that AT will be able to differentiate the link types

Matt: we will take these and make them part of core?

Rich: do people want this?

<MichielBijl> +1 to footnote



Rich: there seem to be giid arguments for that

<AmeliaBR> Some of this functionality could be handled by role-description. The benefit of having standard link types is that the name can be localized by the user agent/AT.

Rich: proposals seem a bit to overlap

<joanie> AmeliaBR: Not just localized, but programmatically processed with specialized logic.

Matt: mainstream browsers need support these particular roles

RESOLUTION: DPUB link types preceed as standard roles

<clown> AmeliaBR, joanie, James Craig reject the approach using role description.

(in their name spaces)

Summary of Action Items

[NEW] ACTION: Michael Search for old links and update them in the ARIA 1.1 spec. [recorded in http://www.w3.org/2016/02/25-aria-minutes.html#action01]
[NEW] ACTION: Rich Remove reference to UAAG and reference EO ARIA 1.1 Suite of documents [recorded in http://www.w3.org/2016/02/25-aria-minutes.html#action02]

Summary of Resolutions

  1. close 707 and action 2031
  2. DPUB link types preceed as standard roles
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.144 (CVS log)
$Date: 2016/02/25 19:00:32 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.144  of Date: 2015/11/17 08:39:34  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/ACIDENT/accident/
Succeeded: s/diffwrence/difference/
Succeeded: s/nrmative/normative/
Succeeded: s/HTML AAAM/HTML AAM/
Succeeded: s/on?/on because I have more to think about and research, and we have a guest to discuss aria-linktype?/
Succeeded: s/shrubed/shoved/
Found Scribe: Stefan
Inferring ScribeNick: Stefan
Present: Deborah_Kaplan JaEunJemmaKu Janina Joanmarie_Diggs Joseph_Scheuhammer Rich_Schwerdtfeger fesch Michiel_Bijl matt_king LJWatson MichaelC Bryan_Garaventa
Found Date: 25 Feb 2016
Guessing minutes URL: http://www.w3.org/2016/02/25-aria-minutes.html
People with action items: michael reference remove rich

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.

[End of scribe.perl diagnostic output]