W3C

- DRAFT -

Accessible Rich Internet Applications Working Group Teleconference

12 Jan 2017

Agenda

See also: IRC log

Attendees

Present
Joanmarie_Diggs, Rich_Schwerdtfeger, MichaelC, Janina, JF, Joseph_Scheuhammer, Léonie, matt_king
Regrets
Chair
Rich
Scribe
Janina

Contents


<Rich> https://lists.w3.org/Archives/Public/public-aria/2017Jan/0023.html

<Rich> https://lists.w3.org/Archives/Public/public-aria/2017Jan/0023.html

<scribe> scribe: Janina

Updates

rs: Rich will do core HTML mappings review

Testing & Testable Statements

rs: We're stuck! Between IA2 and UIA bridges

jd: Don't know the details, working on ATK+AT-SPI2 and have started, just barely though, on the ATTA for AXAPI

rs: Believe most are in Firefox
... Don't know about Edge
... Will start filling in the test harness
... Would appreciate assistance from the group!

<Rich> https://www.w3.org/wiki/ARIA_1.1_Testable_Statements

rs: Anything else we can do accelerate the auto test harness?

jd: Written out testable statements would really help
...

rs: Believe most JFW is done
... Need to check with Domenic re Chrome

mk: If we have questions re specific tests, where should we discuss that?
... Mailing list? Git?
... I've questions about grid
... unclear whose testing problems I'm seeing fall into ... ARIA? HTML?
... Role=grid on table, and nothing on cells ...

rs: So what does a td convey?

mk: Hard for me to tell ...

mk Working off of observed screen reader behavior

jd: I can examine, if you like. Should be similar/same

mk: So, wondering whether we should be testing propogation
... Seems grid is unique
... No other ARIA role does anything like that

rs: May not be just browser, may be AT

<clown> https://rawgit.com/w3c/aria/master/core-aam/core-aam.html#role-map-gridcell

jd: IA2 has only role=cell

<clown> https://rawgit.com/w3c/aria/master/core-aam/core-aam.html#role-map-cell

jd: Believe we need testable statements for anything that has changed since 1.0, including changed behavior

stefan: Wondering about Edge support ... Do we have? What's situation with JFW plus Edge

rs: We will test Edge, we have mappings
... Whether JFW supports Edge isn't our department
... NOte that Edge does not expose DOM
... As to where to ask -- list is OK, and on the call is OK.

<Rich> https://www.w3.org/TR/html51/semantics-embedded-content.html#the-picture-element

Embedded Content for ARIA 1.2

rs: We can go through these, but first some examples ....

<Rich> <style>

<Rich> img {display: block; margin: 0 auto;}

<Rich> </style>

<Rich> <picture>

<Rich> <source

<Rich> media="(min-width: 650px)"

<Rich> srcset="images/kitten-stretching.png">

<Rich> <source

<Rich> media="(min-width: 465px)"

<Rich> srcset="images/kitten-sitting.png">

<Rich> <img

<Rich> src="images/kitten-curled.png"

<Rich> alt="a cute kitten">

<Rich> </picture>

rs: Note the multiple source support
... Best fit wins
... img is the fallback
... But, we don't want to map multiple sources -- It's only one image

jd: So what's exposed to AT should be the rendered image

rs: Wondering where alt is ...

mc: Believe it's from the fallback img
... So don't need new ARIA feature

rs: Need to decide whether we expose Picture and Source

<clown> https://w3c.github.io/html-aam/#el-picture (no corresponding aria role)

mc: Suspect you'd map Picture
... So authoring error if img fallback with alt not provided

mk: Would no alt on imate and ARIOA-Label on Picture be illegal?

<JF> +1 to Matt's suggestion - aria-label on the <picture> element

<MichaelC> mc: HTML mappings needs to say that img embedded in picture isn´t mapped

rs: What happens with tab index on a descendent

mc: Should take the parent

<MichaelC> because the picture itself is mapped to img

mc: Still not sure we've anything to do here
... We're not in a special situation here

rs: Question is what to convey to user

<MichaelC> the text alternative algorithm should say get text alt from embedded img (preferred) but otherwise from aria-label on picture or something

jd: I believe we map to role=img

jn: Can't you also have text?

mc: If yes, could add to accessible mapping calculation

js: NO,

mc: So all HTML AAM, and not our spec

<clown> https://w3c.github.io/html-aam/#el-picture

mc: Note AAM isn't complete, and this issue may not have been considered

js: Yes, we're asked to review

rs: Not sure what for tab index
... ARIA attribs on Picture allowed, but not mapped

js: No, not according to AAM

<Rich> Allowed ARIA state and property attributes:

<Rich> Global aria-* attributes

<Rich> DOM interface:

js: Maybe I'm looking at old

<clown> https://www.w3.org/TR/html-aria/#allowed-aria-roles-states-and-properties

rs: Not sure what with focus

<clown> https://w3c.github.io/html-aria/#allowed-aria-roles-states-and-properties

mc: Suspect it's underdefined

rs: Map Picture to Img

mc: NO, we do nothing
... Nothing to do in ARIA, but much to do with the AAM, which is not us

rs: Assume content below picture, one of which is the image. Don't want to map both
... In a web component?

mc: Don't know, but it's all auto

rs: Unclear how this is used in a component

mk: Don't get the difference

mc: Believe a component doing anything fance will be script based, and so this is up to the scripting

rs: So, an inconsistency in the mapping ...

mc: Noting Picture is a grouping, you're only picking one image, and how the alt gets to the user is underdefined
... Don't think we need to go to this level of situation handling

rs: So we have to identify this inconsistency to Web Platforms

jf: Wonders what happens when fallback is missing

mc: We can't fix all errors

jf: Does it still end up something like image1765.jpg?

mc: That's not specified, of course

rs: Wondering where my issue is ...

js: Core AAM

<clown> https://lists.w3.org/Archives/Public/public-aria/2017Jan/0006.html

js: Above contains links on where to file issues

<Rich> ACTION: Rich As part of HTML-AAM review: picture maps to img. name computation starts with aria-labelling on picture followed by alt text on the embedded img and last would be the URL of the chosen source. The image child is not mapped [recorded in http://www.w3.org/2017/01/12-aria-minutes.html#action01]

<trackbot> Created ACTION-2132 - As part of html-aam review: picture maps to img. name computation starts with aria-labelling on picture followed by alt text on the embedded img and last would be the url of the chosen source. the image child is not mapped [on Richard Schwerdtfeger - due 2017-01-19].

<Rich> action-2132?

<trackbot> action-2132 -- Richard Schwerdtfeger to As part of html-aam review: picture maps to img. name computation starts with aria-labelling on picture followed by alt text on the embedded img and last would be the url of the chosen source. the image child is not mapped -- due 2017-01-19 -- OPEN

<trackbot> http://www.w3.org/WAI/ARIA/track/actions/2132

rs: Continuing -- Iframes ... Do we need?

mc: Unsure

rs: Don't believe we need

<clown> https://w3c.github.io/html-aam/#el-iframe

mc: Suspect this would now be scripted

mk: Think no corresponding role is perfect, we want iframe to do nothing. Alternative is none

mc: Need access to content

mk: None preserves, and anything inside the iframe

jn: Can navigate by iframes

rs: Could be confusing in a mashup

jn: Yes, sometimes

mk: Agree we don't need a new role for it
... Perhaps mapping to "none" might be a worthy discussion with AT
... Maybe if we have title -- just like we treat section

jd: Are we pruning iframes from the tree?
... Just want to avoid nested doc frames

rs: Leave as is?

jd: Yes, no ARIA role, but other expose may be useful

RESOLUTION: No new role for Iframes in ARIA

Embed Element

<clown> https://w3c.github.io/html-aam/#el-embed

jd: Believe we've morphed the application role recently
... Believe Embed is like 1.1 Application -- Custom control like

<Rich> https://www.w3.org/TR/html51/semantics-embedded-content.html#the-embed-element

<Rich> Allowed ARIA role attribute values:

<Rich> application, document or img or presentation.

rs: Don't believe there's anything for us to do that adds value
... Believe we have what we need

RESOLUTION: No new role needed for Embed

mc: Also for Object?

rs: Yes

RESOLUTION: No ARIA role needed for Object either

Next Week

rs: Pick up with Video and Audio

RESOLUTION: No ARIA role for Param element

https://www.disneymoviesanywhere.com/support

rs: So, for next week.

Summary of Action Items

[NEW] ACTION: Rich As part of HTML-AAM review: picture maps to img. name computation starts with aria-labelling on picture followed by alt text on the embedded img and last would be the URL of the chosen source. The image child is not mapped [recorded in http://www.w3.org/2017/01/12-aria-minutes.html#action01]
 

Summary of Resolutions

  1. No new role for Iframes in ARIA
  2. No new role needed for Embed
  3. No ARIA role needed for Object either
  4. No ARIA role for Param element
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.148 (CVS log)
$Date: 2017/01/12 19:01:42 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.148  of Date: 2016/10/11 12:55:14  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/at-spi/ATK+AT-SPI2 and have started, just barely though, on the ATTA for AXAPI/
Succeeded: s/Also the Apple mappings//
Succeeded: s/eMBED eLEMENT/Embed Element/
Found Scribe: Janina
Inferring ScribeNick: janina
Default Present: Joanmarie_Diggs, Rich_Schwerdtfeger, MichaelC, Janina, JF, Joseph_Scheuhammer, Léonie, matt_king
Present: Joanmarie_Diggs Rich_Schwerdtfeger MichaelC Janina JF Joseph_Scheuhammer Léonie matt_king
Agenda: https://lists.w3.org/Archives/Public/public-aria/2017Jan/0023.html
Found Date: 12 Jan 2017
Guessing minutes URL: http://www.w3.org/2017/01/12-aria-minutes.html
People with action items: as html-aam maps of part picture review rich

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


[End of scribe.perl diagnostic output]