Protocols and Formats Working Group Teleconference -- 01 Oct 2015

<trackbot> Date: 01 October 2015

<richardschwerdtfeger> meeting: W3C WAI-PF ARIA Causus

<richardschwerdtfeger> https://www.w3.org/WAI/PF/wiki/Meetings/TPAC2015/ARIA#Potential_Agenda_items

<richardschwerdtfeger> https://lists.w3.org/Archives/Public/public-pfwg/2015Sep/0262.html

<richardschwerdtfeger> this is the agenda:

<richardschwerdtfeger> https://lists.w3.org/Archives/Public/public-pfwg/2015Sep/0262.html

<MichaelC> +MichaelC

<jamesn> scribe: jamesn

<mck> http://rawgit.com/w3c/aria/mck_issue633/aria/aria.html#grid

Grid role?

MK: suggest that people look at the email and reply on the list rather than taking meeting time
... will take until TPAC to get screen reader vendor feedback

<clown> current master: http://rawgit.com/w3c/aria/master/aria/aria.html#grid

MK: In terms of agenda time I feel it would be better to do this later

RS: appendcount and prependcount.... do we really need these?

MK: want to get feedback from screen reader vendors as to whether they can do without them

RS: I'm still not convinced that we really need them

ACTION-1361

ACTION-1361?

<trackbot> ACTION-1361 -- Matthew King to Suggest new text for the application role -- due 2015-06-11 -- OPEN

<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1361

ISSUE-740?

<trackbot> ISSUE-740 -- Deprecating gridcell for grids and use cell -- open

<trackbot> https://www.w3.org/WAI/PF/Group/track/issues/740

MK- want to talk about in the context of ISSUE-633

ISSUE-633?

<trackbot> ISSUE-633 -- listbox and tree may contain only static items; badly need interactive widgets that can contain interactive typed items -- open

<trackbot> https://www.w3.org/WAI/PF/Group/track/issues/633

I think I have addressed this. If you look at the new gridcell text in the draft i just listed means issue-740 is overcome by events

MK: what this is saying is that if you "the cell elements of a grid have role gridcell"

<clown> http://rawgit.com/w3c/aria/mck_issue633/aria/aria.html#gridcell

MK: and if you follow the link to gridcell it says it can be focussable, editable.... i intended that to be part of the very first paragraph to make it mroe obvious to people.
... those are the things that are distinctly different about grids from tables
... rowheader and columnheader raises another issue. Makes mapping algorithm difficult. Becuase a colheader can either be a widget or not a widget. It seems to me like we need a header cell role which maps to a TH
... cells are not focusbale or selectable. THs are a form of a cell
... if th maps to colheader the supported states and properties include things that do not apply to a TH
... are we good with that

RS: I think we can do that

MK: the you could potentially have done the same thing for cell

RS: everything can get focus in HTML or SVG

MK: selection is a different thing'

RS: the net is that unless anyone disagrees is that close 740 as overcome by events
... I think we could put text in the docuement to address that in the context of a table colheader and rowheader are not selectable

RESOLUTION: close issue-740 to be addressed by issue-633

ACTOON-1361?

<clown> action-1361

<trackbot> action-1361 -- Matthew King to Suggest new text for the application role -- due 2015-06-11 -- OPEN

<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1361

MK: push to next week

ARIA 1.1 Action draft and exit criteria

<clown> action-1513

<trackbot> action-1513 -- Michael Cooper to Set up ARIA 1.1 Requirements draft -- due 2015-08-16 -- OPEN

<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1513

<clown> action-1719

<trackbot> action-1719 -- Michael Cooper to Draft aria 1.1 cr exit criteria -- due 2015-09-09 -- OPEN

<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1719

MC: exit criteria already on your agenda
... requirements do not need to be discussed at TPAC
... exit criteria - blocking time to work on b4 tpac

RS: move requiremnet due date out

new figure role

<clown> action-1653

<trackbot> action-1653 -- Ann Abbott to Find out what chinese equivalent to ctrl+a is, to help with internationalisation guidance in the ARIA APG 1.1 -- due 2015-06-22 -- OPEN

<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1653

<clown> action-1633

<trackbot> action-1633 -- Shane McCarron to Work with rich on updated process for creating new role value taxonomy extensions -- due 2015-05-14 -- OPEN

<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1633

<richardschwerdtfeger> https://www.w3.org/WAI/PF/wiki/ARIAExtensions

RS: had a discusssion about extending for dpub stuff

SM: please close it - it is done

RS: dicsussions i had with SteveF - i posted a defn to the list

<clown> https://lists.w3.org/Archives/Public/public-pfwg/2015Sep/0033.html

RS: in HTML5 there is a figure. the issue we have is when you have a figcaption this really is a description not a label for the figure
... but there are times when the label is embedded in the description
... is some special role for figcaption needed to allow you to do this

MK: could this be a label in the figcaption?

RS: Steve was asking do we need a figcaption role
... when yu have a figure role it would also have a describedby to point to the description

Stefan: what is the difference between figcaption and a role of figcaption

<clown> http://www.w3.org/TR/html5/grouping-content.html#the-figcaption-element

RS: we could mao that to a figcaption but it is really the description for the figure which precedes it

MK: we decided not to have cpation for table as we have aria-labelledby and aria-describedby

Stefan: would still have to connect them using escibedby so defeats the purpose
... is there an exmaple?

RS: HTML?

Stefan: a code example which is causing the issue

RS: have figure followed by figcaption. trying to address the mapping under the surface

<clown> https://lists.w3.org/Archives/Public/public-pfwg/2015Sep/0033.html

<richardschwerdtfeger> https://lists.w3.org/Archives/Public/public-pfwg/2015Sep/0154.html

figure: A perceivable section of content which supports the main

document, and should be easily perceivable regardless of its

position in the layout. A figure might contain a graphical

document, an image, or other content such as code snippets or

example text. Authors SHOULD provide a reference to the figure

from the main text, but the figure need not be displayed at the

same location as the referencing element."

A figure MAY have an associate label or caption, or an

associated description.

Assistive technologies SHOULD enable users to quickly navigate

to figures. Mainstream user agents MAY enable users to quickly

navigate to figures.

We differentiate label and descriptions as one being short and the

other being long. Steve would like for us to have a figcaption role

but what I think we really need to do is state that Figures must have

both a visible label and a description. In HTML the description can

be implied. I would like to discuss this with the group and also see

whether Steve things that we should require an aria-labelledby on the

<figure> element or at least an aria-label. I can see where the

implied description could be adequate and we might just want to have

an aria-label on it.

We need to resolve this before this becomes a formal proposal.

RS: if we have figcaption do we assume it is always a caption?

FE: would a figcaption show up in the list of figures or something else

RS: they are using it for dual purpose

<richardschwerdtfeger> <figcaption>The Stata Center</figcaption>

RS: I would have thought you would want the label in the list of figures

<img src="bubbles-work.jpeg"

alt="Bubbles, sitting in his office chair, works on his

latest project intently.">

<figcaption> Bubbles at work</figcaption>

</figure>

<clown> http://www.w3.org/TR/html5/grouping-content.html#the-figcaption-element

MK: I would have reversed the text in that bubbles example
... what any AT does is up to them
... we don't dictate their UI

FE: if you make a table of contents you want it to be predictable

MK: we do use TOC in the spec - but we provide the information - they decide what to do about it

RS: in SVG drawings when you have a figure - you may not have a description. It would be down in the SVG itself

<mck> scribe: matt_King

<mck> RS: if we are gong to have figure role, we can not mandate they have a describedby except through authoring guidelines

<mck> Joseph: In SVG, do you have to title and desc?

FE: text can also be in there so could have aria-labelledby

<mck> Fred: there could be aria-labeledby and a title as well

FE: could also be a title for the graph inside the SVG so could be an aria-labelled by in the SVG

<mck> Joseph: if using straight SVG and no ARIA, what do you use?

MK: if using straight SVG then use title

<mck> RS: then use title in svg; it will create label.

FE: if you loaded SVG as a document it will do that

<mck> RS: it is like the name calculation for html

clown: alt is specific to the image element

rs: figcaption - looking at it. I don't know if it affects the positioning. Usually descriptions would be underneath it

MK: seems like making a decision based on a sample size of 1
... :)

RS: I don't know if it goes at the beginning. All have seen figures with really big descriptions after them
... the most sommon thing is that figcaptions are descriptions

FE: in a regular document - with a list of figures what do you see...

MK: thinking dpub - see a list of figures, usally figure 1 etc.... about a sentence long. Often those captions are relatively long - and the figures do not have a concise label
... in a list of figures would see the labels and the description

<clown> http://rawgit.com/w3c/aria/master/html-aam/html-aam.html#h-figure-and-figcaption-elements

MK: would rather add a caption role and in the context of figure is a figcaption and in table it is a table caption
... we could point at the <P> and use aria-describedby and it would have the effetc but that element doesn't get the caption role

clown: it remains a paragraph inside the table
... not talking about real html table markup

mk: an opportunity to make the table picture more robust

stefan: what is the benefit to a cpation role?

rs: do we want a cpation role and do we want a label role... don't want to open up the label can of worms as people will ask why we dont have it elsewhere

stefan: say you have a figure and link using aria-labelledby wouldn't it be an AT issue to connect the 2

rs: if you put a label inside a figure then the text would appear on the screen

MK: if a figure role on the div and have aria-describedby pointing to a paragraph
... is there no text element in scg
... why not just add a text element and point to it using aria-describedby

RS: the text appears within the figure

stefan: if the only issue is visibility can just add a hidden attribute to it

rs: svg put it in the container can have the text within the svg itself

<clown> <div><h4 id='figcaption">Figure caption text here</h4><svg> … <somesvgelement aria-labelledby="figcaption">…</svg></div>

fe: they are not going to put the description inside the svg

MK: we came up with a table sollution which didn't need a cpation role
... seems a perfect parallel
... if the issue is mapping it doesn't seem like we need it

RS: Steve asked that we have a separate role for figcaptions
... reasoning he gave is that it acts as a label and a decription

MK: I think a figure that has aria-describedby pointing to something has a description, shouldn't require a label - if it has both then a UA can ecide which to use for whatever purpose it has
... seems like label and description should be good for figures
... it is up to them if they have labels or descriptions or both

<richardschwerdtfeger> https://lists.w3.org/Archives/Public/public-pfwg/2015Sep/0154.html

RS: will tell Steve we don't want a seperate caption - if you have it use describedby

<jnurthen> issue-740?

<trackbot> issue-740 -- Deprecating gridcell for grids and use cell -- closed

<trackbot> https://www.w3.org/WAI/PF/Group/track/issues/740

<jnurthen> action-1672?

<trackbot> action-1672 -- Michael Cooper to Mark aria-grabbed, aria-dropeffect as deprecated and provide reference definition of derprecated -- due 2015-08-16 -- OPEN

<trackbot> https://www.w3.org/WAI/PF/Group/track/actions/1672

<jnurthen> MC: I will put on an editors agenda to work out how it should work

<richardschwerdtfeger> The group agreed that these should be deprecated. We just need wording that these attributes are deprecated and that alternative will be sought in later version of ARIA that takes into account HTML drag and drop functionality.

<jnurthen> issue-704?

<trackbot> issue-704 -- Should there be an ARIA method to indicate the primary button? -- open

<trackbot> https://www.w3.org/WAI/PF/Group/track/issues/704

<MichielBijl> +1

<jnurthen> RS: coga people will want this

<clown> +1

<MichielBijl> +2

<jnurthen> rs: going to have an aria function?

<jnurthen> rs: a help button - need to know what the purpose of a button is

<jnurthen> aria-default="true/false"

<MichielBijl> input type=button/submit

<MichielBijl> role=button?

<clown> but not role=button aria-pressed

<MichielBijl> scribe: MichielBijl

Matt: if there is more than one default button, there is no benefit to coga
... it should be a relationship

clown: what if you're in a modal and that has a default button

Matt: then you're in a different window

Protocols and Formats Working Group Teleconference

01 Oct 2015

Attendees

Contents

Grid role?

ACTION-1361

ARIA 1.1 Action draft and exit criteria

new figure role

Summary of Action Items

Scribe.perl diagnostic output