Protocols and Formats Working Group Teleconference -- 11 Jun 2015

<trackbot> Date: 11 June 2015

<richardschwerdtfeger> meeting: W3C WAI-PF ARIA Caucus

<richardschwerdtfeger> http://lists.w3.org/Archives/Public/public-pfwg/2015Jun/0022.html

<scribe> scribe: MichaelC

ARIA WG charter

-> http://www.w3.org/2015/04/draft-aria-charter Draft ARIA Charter

mc: should be going to AC in next week or so

has been approved by W3M

going in separate surveys for granular results

Label role

action-1422?

<trackbot> action-1422 -- Joanmarie Diggs to Add label role (text from Matt King) -- due 2014-05-05 -- OPEN

<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1422

jd: have put edits in a branch, feedback so far from Joseph

not convinced we need the role

rs: not sure either but would have helped with captions

jd: caption isn´t a label, there are key differences

rs: we don´t share that view then

jd: there´s text saying you don´t have to use it but can if you like

not sure that´s strong enough reason for a role

rs: agree, but others wanted it to map to HTML figcaption and stuff

<Zakim> clown, you wanted to summarize my concerns with lable role.

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

<clown> An element containing content that labels another element

<clown> <span role="label" id="checkLabel" aria-label="Allow incoming

<clown> connections"></span>

js: my main concern is conflict between containing content, and examples that might not have content

which is all fine but the definition isn´t right then

either need to remove content from the definition, or not allow aria-label / aria-labelledby

jd: agree

this is an inheritance issue

rs: aria-labelledby is global

js: which means it can be applied to any element

would be weird to disallow it, but I don´t know the use cases enough

ack

rs: think aria-label aria-labelledby should stay global

easier to remember

no point making exception for a somewhat edge case

use cases for label:

caption

figcaption

but you could handle the labeling at the host language mappings level

to form a virtual labelledby

<richardschwerdtfeger> ack

js, jd, rs: this is useless

<clown> action-1422?

<trackbot> action-1422 -- Joanmarie Diggs to Add label role (text from Matt King) -- due 2014-05-05 -- OPEN

<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1422

<jamesn> no - I agree it is useless

jn: the rationale was an attempt to get a 1:1 mapping between HTML 5 elements and ARIA elements

but don´t think it buys us anything in practice

cs: useful from code clarity point

fe: yes, simplifies some SVG cases

rs: if something is labelled by something else, you can put label role in the platform mapping, don´t need author to do it

js: I heard CS say it makes it easier for author to know what´s gonna happen

rs: have never seen this in the wild

jn: not having it doesn´t stop authors from using HTML label

mc: figcaption and caption label their parents; label labels something else

jn: not invalid to have a label in HTML that labels nothing

mc: probably because of limits of DTD expression

fe: would be weird if a label for referenced something that said it´s labelledby was something else

rs: we should kill it

it only helps some behind the scenes computation

mc: expect HTML-AAM will have to map in way to not require ARIA label role, but potentially map to a platform feature nonetheless

does this change any previous decisions to try to provide a comprehensive ARIA?

rs: best to provide roles *where they are useful*

js: SVG?

fe: it´s all groups and stuff

rs: we could make target of an aria-labelledby to defined as a label, but don´t see it as important

fe: could @@

rs: we could create a label role for SVG

fe: hoped to piggy-back off of this in core ARIA

but not a big deal

rs: more useful to map to text

fe: usually you care about the text, not the symbol

but that may be an SVG-specific issue

in a chart you could have axis with major and minor tick marks, don´t want all of those to become labels

rs: propose that target of aria-labelledby has no mapping

fe: does that stop you from getting to it?

rs: no

jd: if it´s not mapped, AT can´t find

rs: in above example you would ignore the aria-labelledby

<text role=¨label¨ id=¨text¨>label</label><circle aria-labelledby=¨text¨/>

no, not that, I´m getting at that it´s not meaningful to put aria-labelledby on a label

don´t know why someone would, but...

<clown> +1

<Stefan> +1

<richardschwerdtfeger> Proposal: do not put role of “label” in the ARIA spec. as is to prone to author errors

<clown> +1

<richardschwerdtfeger> +1

<joanie> +10000000

<fesch> + 1/2

<Stefan> +pi

<clown> + i

RESOLUTION: do not adopt the label role in ARIA core

<bgaraventa1979> +1

<jamesn> +1

Localized role descriptions

<richardschwerdtfeger> ttps://rawgit.com/w3c/aria/roledescription/aria/aria.html#aria-roledescription

<clown> action-1417?

<trackbot> action-1417 -- Joanmarie Diggs to Add edit for ISSUE-636: localized role description -- due 2015-06-12 -- OPEN

<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1417

rs: Markku´s gonna love this

various: sorting what this means

<clown> <a role="button" aria-role-description="attachement-button"> ...

jd: you would use this to provide clarification to the user about the intent of the role

mc: kind of like an unofficial sub-role

jd: get AT to announce something helpful

but needs to be on a meaningful role to begin with, or the AT might overlook entirely

rs: so require ARIA role when this is used?

jd: we tried that, but found use cases for use on elements that have implicit ARIA semantics

so came up with rule that their must either be ARIA role, or have implict role mapping

??: will authors understand that?

mc: you have to be pretty knowledgeable to know why to use this to begin with, so maybe it´s safe enough

fe: no, it´ll be a shiny toy that people will play with in all sorts of ways

jd: is the concern about what the prose says, or how it says it?

<richardschwerdtfeger> Authors MUST only use aria-roledescription on elements that equate to a valid WAI-ARIA role (have an implicit WAI-ARIA role semantic) or or have a valid WAI-ARIA role applied. If neither condition is met then then the aria-roledescription MUST NOT be exposed by user agents.

<richardschwerdtfeger> Authors SHOULD only use aria-roledescription on elements that equate to a valid WAI-ARIA role (have an implicit WAI-ARIA role semantic) or or have a valid WAI-ARIA role applied. If neither condition is met then then the aria-roledescription MUST NOT be exposed by user agents.

<clown> If neither condition is met then user agents MUST NOT expose the aria-roledescription.

<richardschwerdtfeger> Authors SHOULD only use aria-roledescription on elements that equate to a valid WAI-ARIA role (have an implicit WAI-ARIA role semantic) or or have a valid WAI-ARIA role applied. If neither condition is met then user agents MUST NOT expose the aria-roledescription.

<richardschwerdtfeger> RESOLUTION: apply the new text above to replace the second paragraph in the specification.

jd: this needs more review cycles, but if approved I´ll close the action / issue

will do on the list

rs: let me know if it needs to come back in a meeting

Keyboard shortcut help

action-1642?

<trackbot> action-1642 -- Richard Schwerdtfeger to Create aria-keyshortcut proposal to group -- due 2015-06-11 -- OPEN

<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1642

issue-711?

<trackbot> issue-711 -- Expose Keyboard Shortcut help information to ATs -- open

<trackbot> https://www.w3.org/WAI/PF/Group/track/issues/711

<richardschwerdtfeger> https://lists.w3.org/Archives/Public/public-pfwg/2015Jun/0029.html

rs: HTML trying to create an access key

this issue goes back 10 years

we knew the accesskey attribute had problems

proposed an access element

brought to XHTML WG

but the scene changed

but we´ve been talking about access keys ever since, but not managed to get anything done

good reasons to take up

want to be able to remap keys

on user side, in order not to conflict with AT

and associate description

but that´s just for moving focus

<Stefan> +q

but now we´ve got shortcuts on elements that have a role and label, and activate

so need the key to activate not just focus

but ARIA doesn´t allow that

because it doesn´t define mainstream UI

so doing this in HTML with access key makes sense

but no progress in years

so my proposal is to put stuff in ARIA spec

and make it optional but hopeful that host language would say ¨go ahead and implement in mainstream UI¨

do have valid use cases for remapping

<Stefan> Need Concept Parallel to that in HTML next

<Stefan> Must think of Other Agents than AT

<Stefan> Split into aria-hotkey and aria-accesskey

ss: should parallel the HTML design pattern

consider how UAs expose the hot keys

split properties to separate focus from activate

hotkey to trigger function, access key to set focus

rs: we weren´t charged to create ¨accesskey¨ in ARIA, just a shortcut

bg: there´s access key, and access key label

the label is supposed to grab native browser keys and stuff

rs: I don´t use accesskey because of how messed up it is

have some UA interest on this

do you all agree ARIA should not specify UA behavior?

<silence=consent>

so that´s a position that has not changed

Ok to put this in a branch for further review?

sounds like the next step, then we need explicit review from the UA folks

note that in the current approach, it would only be available to AT

but UAs could expose

jn: or extensions, or special script library

I´d like to take it further, and provide a description of what a shortcut does, not necessarily tied to an element

ss: I proposed that, response was could use aria-describedby

<clown> quoting Rich: "The AT will have access to the label and descriptions for the widget"

rs: I meant that enough information is available, not that you use aria-describedby specifically

ss: that requires an additional step

a descriptive string would be useful

rs: gotta figure out where to map it

adds to the set of descriptions we provide

js: aria-hint

fe: aria-help

jd: some platforms have an actions interface to get mappings

rs: CS will bring up control patterns
... I´d like to get review on the current first, then address whether to tackle a description

mc: agree with agressively seek review

<clown> issue-406

<trackbot> issue-406 -- Proposal for new aria-hint property. (Previously proposed as @aria-help) -- open

<trackbot> https://www.w3.org/WAI/PF/Group/track/issues/406

rs: let´s put this in a review branch, with a note about needing to get back to descriptions

action-1642: https://lists.w3.org/Archives/Public/public-pfwg/2015Jun/0029.html

<trackbot> Notes added to action-1642 Create aria-keyshortcut proposal to group.

<richardschwerdtfeger> https://www.w3.org/WAI/PF/Group/track/products/17

Actions review

rs: issues list is growing

mc: time for another triage to push stuff to 2.0

rs: yes, and we´re moving up the 2.0 timeframe anyways so won´t be forever

close action-1320

<trackbot> Closed action-1320.

<clown> action-1380?

<trackbot> action-1380 -- Joanmarie Diggs to #presentation should mention aria-hidden vs presentation role on raster and vector images in relation to ACTION-1379 -- due 2015-06-11 -- OPEN

<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1380

don´t think mapping the "offscreen" api properties in the situation of aria-hidden="false" on non-rendered elements will go anywhere

<clown> action-1379?

<trackbot> action-1379 -- Joseph Scheuhammer to TAC should cover rendered descendants from external sources: svg, iframe, etc. (Compose text and get final review form ARIA subteam) -- due 2015-02-13 -- OPEN

<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1379

<clown> action-1405?

<trackbot> action-1405 -- James Craig to For issue-526 modify spec. to allow non-global attrs to apply to host language elements having the correct aria semantics -- due 2015-01-22 -- OPEN

<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1405

<clown> action-1416?

<trackbot> action-1416 -- James Craig to Include definitions for terms used as first word of all attr definitions (defines, identifies, indicates) -- due 2014-04-07 -- OPEN

<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1416

<clown> action-1424?

<trackbot> action-1424 -- James Craig to Propose spec text for generic/general/??? role (computed role of html:div, html:span, svg:g, etc) and clearly explain explicit usage of this role is not common, and clearly explain relationship to group and none role. -- due 2014-04-21 -- OPEN

<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1424

- DRAFT -

Protocols and Formats Working Group Teleconference

11 Jun 2015

Attendees

Contents

ARIA WG charter

Label role

Localized role descriptions

Keyboard shortcut help

Actions review

Summary of Action Items

Scribe.perl diagnostic output