WAI-PF ARIA Authoring Practices Guide Taskforce

29 Feb 2016

See also: IRC log


JamesNurthen, Jemma_Ku, Michiel_Bijl, AnnAbbott, jemmajaeunku, Bryan_Garaventa


<scribe> scribe: MichielBijl

Jump for joy to celebrate leap day 2016!!!

Review latest revision of landmark section proposal (John) https://rawgit.com/jongund/aria/master/practices/aria-practices.html#aria_landmark

JN: Ann + Jon, can you take us through this?

AA: We did a whole lot of editing here.

JG: There is no table of contents anymore

JN: looks a bit shorter too, which is good
... Saw a proposal, JG would create some sort of example

JG: That's too come
... People working on patterns can help

JN: Much easier to understand
... How are we on normative language?

AA: Where are you looking?

JN: Step 3 “name a specific region”

AA: That's from the spec

JN: Okay, then I guess we can do that.
... It's not the same for all of them is it?

AA: I'd have to dive into each one

Relevant text from APG: If a specific landmark role is used more than once on a web page, it must have a unique label.

JG: Not in role=region

JN: Can we change that to a should?

JG: Yes

AA: then we don't have to back it up with the spec right?

JN: That's reasonable.

<annabbott> Step 3 > first bullet: change "must" to "should"

JN: Should someone create their own aria-label instead of referencing a visual label?

AA: No point in adding verbosity to screen readers

MB: I agree

LW: Duplication can get a bit tiresome

JN: I'm okay with that

JG: Leave it the way it is?

JN: Yes
... Design patterns of the banners
... Why does that banner need to have a unique label?

Bullet four: If a page includes more than one banner landmark, each should have a unique label (see Step 3 above).

JN: Happy leaving it like this

MB: I'm fine with that too

JN: Fine with complementry

JG: Is bullet two of 2.3.1. banner, Techniques row in the table true?

LW: Might be true, might be easier to say it's only a banner role if it's scoped to the body element

<jamesn> If not a descendant of an article or section element role=banner, otherwise No corresponding role

<jamesn> https://www.w3.org/TR/html-aria/

MB: So would a header element within a <div role="document"> get role=banner?

JN: Not sure
... This is written backwards

LW: Been battering Steve about that.

JG: Should be restrictive
... We want to make it harder to create role=banner landmarks
... Mailed Steve Faulkner about it, he agreed.

JN: Want to test this for accuracy

JG: There was a testing meeting earlier today

JN: This is HTML 5 testing

LW: Yes, host language defines how this is defined
... That's why I suggested to take down the focus on HTML in this document
... Focus more on ARIA

JG: Not a service to developers if we ignore HTML

LW: Not ignore it but this is ARIA practices, not HTML practices

JN: I agree, we need a way for people to find the correct HTML information

AA: Does the HTML spec state that you get role=banner with this??
... Do we want to ask developers to look at three documents?

JN: Realistic I kind of agree with everyone, not sure we have the bandwidth to maintain.
... Steve has already documented this in his doucment
... Either we have to maintain it ourself, or point at Steve's document

AA: if you use the header element and you put role="banner" on it, what happens?

Link to W3C validator: http://validator.w3.org/nu/

<jamesn> Can we just link to https://www.w3.org/TR/html-aria/#header

JG: Why do we want to make it harder for developers?
... I disagree

JN: Can we link directly to HTML in ARIA document?

LW: There is a lot of information about this out there

JN: consensus seems to be to de-emphasise the HTML 5 information
... Is that what you're suggesting Léonie?

LW: Yes

AA: I need to agree with Jon
... We're trying to make a one stop shop here

JG: Let's not disconnect the information from the relevance.

<LJWatson> The table here could be included as a simple reference for HTML... https://www.paciellogroup.com/blog/2013/02/using-wai-aria-landmarks-2013/

LW: 2.1 should be moved to the bottom of this section
... It's the APG, not the HPG
... I'd be happy to create a comparative version

AA: I just know that day-to-day I work with people not involved with standards work
... They are clueless about this information or where to find it

LW: That's why I'm suggesting what I'm suggesting
... Put it in a single place

AA: They might only come for role=banner
... We have a problem in contentinfo
... Talking about footer
... In all three bullets for techniques.

JN: What's the problem?

JG: Footer creates contentinfo

AA: Oh, sorry

JN: Okay
... Looking at 2.3.4. for form
... Not sure why we have the first bullet point

JG: It's from the specification

“Whenever possible, controls contained in a form landmark should use native host semantics:”

JN: Can we move that to the bottom?

JG: Yes

JN: Maybe not
... If the same set of links is repeated on the page
... Don't think I agree with that
... If they contain the same information they should have the same label

MB: I agree with that

JN: Only issue is that users might think they are looping through the same element or have returned to the top of the list of elements
... copy 1, copy 2 just sounds weird

JG: Is there any guidance on redundancy?

JN: Don't know what the best practice is, but don't like copy n as part of the label?
... Maybe top navigation, side navigation etc

AA: That's a violation of WCAG 2.0

JN: Hmm, not sure (?)

JG: Léonie what would you like to see?

LW: Same label sounds good

AA: Wouldn't that be confusing if you go through the list of landmarks?

LW: Not sure it would matter if I was looking for navigation
... Different version numbers would make me curious as to what the difference is

JG: So better to use same label for each?

LW: Yeah

JN: We often have next/prev buttons, and we label them the same as they have the same function, even if they appear multiple times on the same page

AA: Might be hard for checking tools

JN: Yeah, but we're not writing for checking tools

AA: No, I know, but Jon asked what I thought.

JK: What is the use case for multiple navigation?

MB: Maybe lists of links to different pages on a website in the header that reappears in the footer?

*scribe passed out*

<jemma_> *lol*

JG: If it's not clear to ourselves, we're not clear to others

JN: We have something now that is helpful to people in general
... Great work Jon and Ann, thank you

JG: Teresa Boers helped out a lot

<annabbott> Correction: Teresa Boers

<jongund> JN: We just start discussing tabpanels

<jongund> JN: checking bugzilla for issues

Jump for joy to celebrate leap day 2016!!!

Complete review of section 2.31 Tab Panel http://w3c.github.io/aria/practices/aria-practices.html#tabpanel

<jongund> JN: We have been through some of it before

<jamesn> https://www.w3.org/Bugs/Public/show_bug.cgi?id=28993

<jongund> JN: We dicided to do that (pageup and page down)

<jongund> JN: We have been through the top section and made changes, and we removed stuff from...

<jongund> JN: MB did you lookup the control+up arrow stuff

MB: I can't recall agreeing to look at that

<jongund> MB: un intelligible

<jongund> MB: Iam still here

<jongund> JN: We should not say you have to do it

<jongund> JN: We decided to drop control+PageUp and PageDn

<jongund> JN: People can do it, but we won't recommend a set of controls

<jongund> AA: Droping control+up arrow?

<jongund> JN: You shift+tab back to the tablist

<jongund> AA: You are saying

<jongund> AA: Once you are in the tabpanel, shift+tab will take you tot he tab if you press it enough

<jongund> LW: That works in software interface

<jongund> JN: If someone whats a quicker way, they can do so

<jongund> AA: As long as it is discoverable

<jongund> JN: The discover ability is going to depend on the app

<jongund> AA: There is lots tab panels, and tabs, and people are not trained on them

<jongund> LW: Thats why it is important to emulate other software patterns (i.e. Windows)

<jongund> JN: You can still add the functions, and in some cases it is highly desirable

<jongund> JN: Not a good design flow, but adding a keystroke they don't know will help

<jongund> AA: Agreed

<jongund> JN: Do you have any implemenations

<jongund> JG: We have an example, but it gets really complicated, I am glasd its gone

<jongund> JN: There was an issue in Github on this as well

<jongund> JN: Why ALT+Delete and not just Delete

<jongund> JN: At Oracle when you are on a tab ou can just press delete

<jongund> AA: What has focus?

<jongund> JN: When focus is on the tab itself

<jongund> JN: That is why it is ALT+Delete in the spec, but we did not like it

<jongund> JN: Has anyone implemented it?

<jamesn> http://www.oracle.com/webfolder/technetwork/jet/uiComponents-tabs-addRemove.html

<jongund> LW: I have not seen it

<jongund> AA: No

<jongund> JN: Link is to an example of a deletable tab

<jongund> JB: Can you focus the x

<jongund> JN: You can only focus the tab and use the delete key to remove

<jongund> JN: I am happy to leave ALT+Delete as long with the Delete on the tab option

<jongund> AA: There is more text in there


<jongund> JN: Can we talk about not deleting tabs

<jongund> JG: I think we should have techniques

<jongund> JG: We can build some examples for the APG

<jamesn> https://github.com/w3c/aria/issues/206

<jongund> JN: SHould tab panels activate on focus? Was raised on Github

<jongund> JN: Some times we do not want the arrow key to change the tab

<jongund> JN: There are more complex web apps, that moving focus to a new tab is problematic

<jongund> AA: I agree

<jongund> JN: I agree and the people commenting on the bug agree to

<jongund> JN: When application does not respond in a timely way, it can be a problem

<jongund> JN: There are problems sometimes moving the focus

<jongund> JN: Allowing space or other key to navigate a tab

<jongund> JN: I think it is OK from a WCAG perspective either way

<jongund> LW: I think it is OK

<jongund> JN: I will write a proposal on this using space or enter

<jongund> LW: Thanks

<bgaraventa1979> yes

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.144 (CVS log)
$Date: 2016/02/29 19:31:22 $

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/contain this information/state that you get role=banner with this?/
Succeeded: s/Trisa/Teresa/
Succeeded: s/*lol/*lol*/
Found Scribe: MichielBijl
Inferring ScribeNick: MichielBijl
Present: JamesNurthen Jemma_Ku Michiel_Bijl AnnAbbott jemmajaeunku Bryan_Garaventa
Got date from IRC log name: 29 Feb 2016
Guessing minutes URL: http://www.w3.org/2016/02/29-aria-apg-minutes.html
People with action items: 

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

[End of scribe.perl diagnostic output]