Protocols and Formats Working Group Teleconference -- 30 Jun 2014

<trackbot> Date: 30 June 2014

<richardschwerdtfeger> meeting: W3C WAI-PF ARIA Caucus

<Birkir> Birkir Gunnarsson the the secret 919 code

<Birkir> Raleigh, North Carolina!

<Birkir> like the Life Aquatic discussions, great movie too ;)

<Birkir> Next time, I prommise, not quite set up this time.

<Birkir> I promise not to scribe in Icelandic .. and next time will be next Monday, promise

<richardschwerdtfeger> scribe: jongund

RS: Intro to the specs

JS: We will publish heart beats in we want to be updating intro, we need the newer thinking
... Please look at the intro, looking for volunteers to look at the intros

RS: I have to do it for SVG
... Things have chnaged form when we started ARIA
... I am willing to take a stab for ARIA spec and Core and I will be doing it for SVG
... I will not being doing it for HTML5

MC: We will have several fingers in that pot, but if rich can do a first pass that will be a big help

RS: Are there specific points you want in the intros

<MichaelC> ACTION: schwerdtferger to draft introductions for ARIA 1.1, Core-AAM 1.1, and SVG-AAM 1.1 [recorded in http://www.w3.org/2014/06/30-aria-minutes.html#action01]

<trackbot> Error finding 'schwerdtferger'. You can review and register nicknames at <https://www.w3.org/WAI/PF/Group/track/users>.

MC: I have a few things, write from the perspective of someone who is not familiar with the technology or the problem, help people understand the problem and solution

RS: We are looking at application wide changes

<trackbot> #aria is associated with #pf.

<MichaelC> ACTION: RS to draft introductions for ARIA 1.1, Core-AAM 1.1, and SVG-AAM 1.1 [recorded in http://www.w3.org/2014/06/30-aria-minutes.html#action02]

<trackbot> Created ACTION-1470 - Draft introductions for aria 1.1, core-aam 1.1, and svg-aam 1.1 [on Richard Schwerdtfeger - due 2014-07-07].

RS: We are trying to harmonize the integration of accessibility around graphics and HTML5
... ARIA 2.0 is way off in the future
... Addressing things that we were not able to address in 1.0
... We are trying to fix all these particular issue

JS: We need a requirements document, issues that we are working on, when we are done we point to a problem statement ...
... looking for some to help with that
... Did I miss anything

MC: Two important roles we want to consider, how does the spec meet its requirement
... JG: I would be willing to help with that

<MichaelC> ACTION: jon to start list of requirements for ARIA 1.1 [recorded in http://www.w3.org/2014/06/30-aria-minutes.html#action03]

<trackbot> Created ACTION-1471 - Start list of requirements for aria 1.1 [on Jon Gunderson - due 2014-07-07].

MC: A single requirements document or separate documents for each spec, need to make a conscious decision

JS: JRG think about that and give a recommendation
... When do we want to publish a new heart beat, we need to get ahead of TPAC, early october

RS: About 3 months, about every quarter
... any objections?

MC: We need to keep that in our minds...

JS: We need to test early, when we want to do that

RS: To that point, I would like to get some feedback, for role="none" I have been filing bugs

CS: I would prefer something other than a wiki

JS: Concerned about about maintanence

CS: Bugzilla

RS: Bugzilla for all the other bugzilla

<Birkir> assign someone to file issues against browsers/assistive technologies (or to oversee such a process)?

CS: A bug for each browser and linked ...

RS: Do you want to do it?

CS: When do we want to start testing?

RS: We want to get the defects in the browser

CS: Let me talk to IE team on what would help them

RS: You can talk to mozilla on the list, JC is out for 3 weeks

CS: I can talk to IE team and ping david

RS: I can't do anymore right now

JG: I have some people who can help with testing

MC: It sounds like we have some people to create test cases

RS: FOr role=presentation, we can do role="none"

JS: That is an east

MC: test harness copy some stuff over...

JG: Question about nested tables and role=presentation

RS: Can JRG create test cases

<Birkir> If they are accessible, I would be happy to help out with test case writing.

JG: We can try it

<Birkir> I am new to this process.

<MichaelC> ACTION: cooper to work with Jon Gunderson and his team on creating test cases for for ARIA 1.1 [recorded in http://www.w3.org/2014/06/30-aria-minutes.html#action04]

<trackbot> Created ACTION-1472 - Work with jon gunderson and his team on creating test cases for for aria 1.1 [on Michael Cooper - due 2014-07-07].

<MichaelC> ACTION: cynthia to prepare recommendation on automation of tracking of bugs filed with various implementers [recorded in http://www.w3.org/2014/06/30-aria-minutes.html#action05]

<trackbot> Created ACTION-1473 - Prepare recommendation on automation of tracking of bugs filed with various implementers [on Cynthia Shelly - due 2014-07-07].

RS: Anything else JS?
... Issue 658
... This is clean up
... Annotation role....., MS proposal for annotation
... We have already done this, can we close, any objections?
... Resolution to close 568
... Is Joespeh S on today?
... Lets hold off until the next weekend
... Issue 493

<richardschwerdtfeger> https://www.w3.org/WAI/PF/Group/track/issues/493

RS: Why is closed

MC: I am not sure this working, JC closes the issue when an action is created to resolve the issue

RS: Issue 522
... This is from JN

<jamesn> issue-522?

<trackbot> issue-522 -- Name Computation 2A - bullet 3 confusion -- open

<trackbot> https://www.w3.org/WAI/PF/Group/track/issues/522

RS: Is aria-labelledby and aria-label .... not marked as presentational....

JN; This is how other people are reading it, people are not getting the child elements for role="presentation"

JN: If something is marked presentational it should be used in accessible name calc

CS: Firefox gets the name from going through the accessible tree, not the DOM

JN: The name calculation says what should be done, and the presentation should not block the name calculation

CS: it seems the description is unclear

MK: They decided to put all the content of a table into one element, so a screen reader is treating the table as a paragraph
... this is a miss understanding on how to implement role presentation

CS: Should this be part of the implementation guide, maybe a flow chart or several examples

<richardschwerdtfeger> http://www.w3.org/TR/wai-aria/roles#textalternativecomputation

JN: There is not the same understanding even with people whoa re members of this group

RS: Let's say we have a label on something, and the button is marked presentational, the name has to be applied to what..

<bgaraventa1979> +q

JN: creating a an example .....

MK: that is one thing that is not clear, we need to give something a role ...

CS: DIV has a role of group

MK: It does not

CS: It does on every browser
... Every windows browser

RS: I am sure that is not the case

MK: I thin your right rich
... If is suppose to be something other than what it was, .... the content still has to be there, and the that content should have a role, even though it is different

RS: The text content is there, but there is no role
... In SVG there is a lot of things that will have role="none"

<jamesn> in FF it seems to be role of section

RS: There are different roles for DIV in various browsers and OSs
... the text has a node
... Text in the browser has role of readonly text

MK: We do not say that specfically

RS: We don't have a role a mapping for text

MK: We don't have a role mapping for role=presentation

RS: Paragraph has roles of ....

CS: What are the roles for span

MK: It is interesting though ....

<jamesn> span is closest as spans can have child elements.....

RS: span has a number of different roles depending on OS and browser

<jamesn> ugh...

MK: we have a defect in FF they put all the text of a layout table as a single line
... For a able to be readable needs to have some type of separator for text in the different cells
... giving an example .....

CS: that is a data table it should not be a layout table
... You could make that a data table, another example.....

MK: In practice it does't work very well, we need a more paragraph model

CS: You have images, text ...., really should be absolutelely positioned...

<Birkir> If they are accessible, I would be happy to help out with test case writing.Divs would do the trick with a presentational table, they are blocklevel elements, spans make tables with role="presentation" near unreadable, that is true

MK: If they are spaced differently, the screen reader needs to know about the spacing too
... That's why the role of whats left matters

RS: Button doesn't have any semantic children
... The button objects goes away and you just have text
... On the table all the children go away....
... That makes sense too

MK: We did have a separate issue of buttons have popup menus

RS: Look at this as semantic sugar, when you take it away you just have content

JN: I reported it, brian was interpreting different from me
... Reading a e-mail response....

Brian: I know which one your talking about now..

JN: I requesting a clarification so people reading it for the first time have a better understanding it
... There is confusion even among current working group memebers

RS: Which aspect of the text could help with clarification

<richardschwerdtfeger> If aria-labelledby and aria-label are both empty or undefined, and if the element is not marked as presentational (role="presentation"), check for the presence of an equivalent host language attribute or element for associating a label, and use those mechanisms to determine a text alternative. For example, in HTML, the img element's alt attribute defines a label string and the label element references the form element it labels. See How to Specify

<richardschwerdtfeger> Alternate Text ([HTML], section 13.8) and HTML 5 Requirements for providing text to act as an alternative for images ([HTML5], section 4.8.1.1).

JN: We understand it now, so it is hard for us to identify the problem
... Maybe we say that the child text content does not go away and is still part of the accessibility API

MK: Can we simplify role=presentation now has a role=text or something else (e.g. equivalent), would that help

RS: We spent a lot of time writing this, reading through spec text....
... Do people understand the text?

Brian: Developers don't understand the difference between the accessibility tree and the DOM

RS: The mis match between the DOM and accessibility tree

MK: There is more than that, must "expose" content, we don't say how that should be done

CS: The bulleted list is not the best to way to explain algorithm, pauedo code or input/output
... maybe some examples from good test cases
... There is too much open to interpretation

RS: So we can't put an exhaustive list in the spec, do we need to show an example
... One where you have a single table and show the new tree
... Nested data table in a layout table
... button with role presentation
... Do these sound like the three scenarios we need

JN: I am not going to comment until i see the new description
... We need get some new people to look at it

RS: is putting it in the role presentation for ..

JN: We need them in both places

CS: Even implementators

RS: I am worried about the spec getting to big

MC: We use supporting documents

RS: I am looking at bullet 3 again, this was difficult approach

<Birkir> aria accessible name calculation is also tricky in that the spec does not exclusively state if user agents should stop when they find an accessible name, or continue down the list.

CS: Give me an action

<Birkir> think a button with an aria-label and a title.

RS: Can you work on this with Joesph, because he wrote this section

CS: We can do it during the implementation call
... It is really important

<richardschwerdtfeger> ACTION: Cynthia Work with Joseph S. and David B. to rewrite text alternative computation for both the ARIA spec. and the Core Accessibility API mappings specification. [recorded in http://www.w3.org/2014/06/30-aria-minutes.html#action06]

<trackbot> Created ACTION-1474 - Work with joseph s. and david b. to rewrite text alternative computation for both the aria spec. and the core accessibility api mappings specification. [on Cynthia Shelly - due 2014-07-07].

MK: I now understand why we have problems in the browser, IE column, exposes role system ..., see general rules

RS: Who did that?

MK: That's what is says to do ...

CS: Part of the over all parent, a string within the windows

MK: Expose the content, means to put it in the tree

RS: We want to say the semantics goes away...

MK: If is is going to be in the tree, even plain text has some kind of role
... We want to make sure there is a sensible mapping, which we do not have

RS: Do you want put the examples in the spec, but the spec gets bigger

JN: the authoring practices can have more extensive examples
... Authors need to know what is going to happen

MK: I agree with you JN

CS: An input and output table is crucial...

MK: We need to train authors on how to use it properly

<richardschwerdtfeger> http://www.w3.org/TR/2004/REC-DOM-Level-3-Core-20040407/core.html

RS: Do you want to create that?

CS: I would like to have a collaborator
... What kind of drawing,I need to translate the thought into a picture

MC: I would like to help

CS: This would be good at the FTF meeting

RS: Creating action item...

<richardschwerdtfeger> ACTION: Cynthia work with Michael C. and Joseph S. to create up to 4 Accessibility diagrams showing the application (before and after) of applying role=“presentation”/none [recorded in http://www.w3.org/2014/06/30-aria-minutes.html#action07]

<trackbot> Created ACTION-1475 - Work with michael c. and joseph s. to create up to 4 accessibility diagrams showing the application (before and after) of applying role=“presentation”/none [on Cynthia Shelly - due 2014-07-07].

<MichaelC> trackbot, associate action-1475 with issue-522

<trackbot> action-1475 (Work with michael c. and joseph s. to create up to 4 accessibility diagrams showing the application (before and after) of applying role=“presentation”/none) associated with issue-522.

<MichaelC> trackbot, associate action-1474 with issue-522

<trackbot> action-1474 (Work with joseph s. and david b. to rewrite text alternative computation for both the aria spec. and the core accessibility api mappings specification.) associated with issue-522.

RS: We have 7 minutes left

<richardschwerdtfeger> http://www.w3.org/TR/wai-aria/roles#directory

RS: Any objects, content from the name from?

MK: Don't name from content

JN: What name from content, all the content
... For a table, it will be its contents, unless there is another label

MK: Table needs a caption ....

JN: Your right

RS: Any object, the contents value for directory?

MK: Do you delete that role in the table...

RS: I would be just the author, not from contents
... Action to editor

<richardschwerdtfeger> ACTION: jcraig remove contents from the list of Name From values for role “directory” [recorded in http://www.w3.org/2014/06/30-aria-minutes.html#action08]

<trackbot> Created ACTION-1476 - Remove contents from the list of name from values for role “directory” [on James Craig - due 2014-07-07].

JN: Role=row takes name form contents

RS: Like an e-mail program

<MichaelC> trackbot, associate action-1476 with issue-643

<trackbot> action-1476 (Remove contents from the list of name from values for role “directory”) associated with issue-643.

JN: Contents might be the right thing...

MK: If it shouldn't have a name it should only have a value

RS: Might break a product
... We can take it off line

rrsagent draft, minutes

- DRAFT -

Protocols and Formats Working Group Teleconference
30 Jun 2014

Attendees

Contents

Summary of Action Items