W3C

- DRAFT -

Protocols and Formats Working Group Teleconference

12 Oct 2015

See also: IRC log

Attendees

Present
(no, one), JamesNurthen, JemmaKu, jongund, AnnAbbott, MattKing, Raghu, Michiel_Bijl, IanPouncey, Bryan_Garaventa, LJWatson
Regrets
Chair
SV_MEETING_CHAIR
Scribe
jamesn

Contents


<trackbot> Date: 12 October 2015

<scribe> Meeting: ARIA APG TF

<JemmaKu> + jemmaku

<JemmaKu> https://cdn.rawgit.com/a11ydoer/practices/master/examples/checkbox/checkbox-1.html

Discuss status of example development.

2 template examples:

https://cdn.rawgit.com/a11ydoer/practices/master/examples/coding-template/template2.html

https://cdn.rawgit.com/a11ydoer/practices/afc179d3cf5ae3496d46080b216ab6751613c16e/examples/coding-template/template3.html

<JemmaKu> https://cdn.rawgit.com/a11ydoer/practices/master/examples/checkbox/checkbox-3.html

<scribe> scribe: jamesn

Pattern work status page update https://github.com/w3c/aria/wiki/Aria-Authoring-Practices-Patterns-Status

MK: no changes
... working on menu right now but didn't finish it
... last week we left the conversation - we are not done with that

<JemmaKu> me/waves hello to Michiel

MK: was looking for another bug on the menu pattern - 1 is out of order

jn: I will submit the other bug

checkbox example review (John)

<JemmaKu> https://cdn.rawgit.com/a11ydoer/practices/master/examples/checkbox/checkbox-1.html

<JemmaKu> https://cdn.rawgit.com/a11ydoer/practices/master/examples/checkbox/checkbox-3.html

JN: space doesn't seem to toggle the all condiments checkbox

JK: can look into fixing

MB: why are we using divs rather than checkboxes
... i think we should encourage people to use the native element

JG: If i am building a checkbox out of SVG then you don't have them
... you are not suggesting having a role on a checkbox are you?

MB: no
... suggesting having a mix of html checkboxes and aria checkboxed like i did with buttons

JG: waht is the advantage of that

MK: if using a native checkbox you don't need any aria at all
... I guess that is true of button too
... could make a toggle button or a menu button out of an html button

MB: we should encourage people to use native elements if they can

MK: if you were using actual checkboxes what would you use aria for

MB: in my view we should show both sidesa

JG: could put a note in it

<MichielBijl> Note: When using HTML use the <button> element. It is recommended that authors not re-purpose other elements to create buttons.

MK: we put comething like that in link

JG: could add some standard ones to the mixed example
... in this case standard checkboxes are used and role=checkbox was used on the mixed one

MK: we would end up putting that note in everything that has an HTML equivalent

JG: would make examples more complex if have to explain why there are mixes of things

BG: see a lot of devs that put role=checkbox on checkboxes
... when aria checkboxes are handy is on touchscreen. People make checkboxes where the only thing visible is the label, but there is no visible checkbox there
... on desktop can use off-screen standard checkbox. on mobile that doesn't work so well

MB: we encourage people to use native elements.

JG: we are not trying to teach people how to use a standard checkbox. there are other places for that

MK: thinking a little about the points steve makes in using aria in html docs
... it is the aria authoring practices guide
... we do want to help people using aria in html do so correctly
... you are starting to sway me Jon. Because there are places that there is redundancy in aria.
... if put <main> role=main is still enouraged in some places
... i don't know if that is still a good idea

MB: i view it as a place to send devs to learn how to do a design pattern

MK: we are trying to cover the aria roles

JG: we should do the roles we want people to use
... the main thing of documenting the patterns with an example for each role. Then see if have bandwidth to work on ant-patterns

MK: already have a plan for which ones we are going to do
... i don't think we need a plain html checkbox example, but a note would be good
... in the notes column - could add a note that "role checkbox is not applied to the standard checkboxed but is to the aria checkbox"

BG: could be useful to note which roles have implicit native semantics
... there are others which do not have native semantics

MK: i wonder if we should incorporate the implicit semantic language into this document someplace
... either do this or refer to it in steves document
... Jon if you can do the 3rd example that way for the mixed checkbox and add the note as described

IP: role presentation doesn't make an attribute optional
... aria-describedby attriebutes don't have an element with that ID

http://www.w3.org/TR/WCAG20-TECHS/F38.html

role=presentation doesn't allow you to skip alt="" according to the html validator

JN: WCAG allows it

MK: I want to get the language about implicit semantics into the APG in appropriate places
... dont know yet if it would be good to have a section about implicit semantics and reference it or come up with a short phrase

LW: once is easiest

+1

LW: can also link to the mapping documents from that explanation
... i think it is worth taking a handful of sentences to expalin it

JG: checbox 2 will reference aria-describedby for a grouping element

the aria-describedby will say what group the checkbox belongs to

Example template and example coding practices

https://cdn.rawgit.com/a11ydoer/practices/master/examples/coding-template/template2.html

<JemmaKu> https://cdn.rawgit.com/a11ydoer/practices/master/examples/coding-template/template2.html

https://cdn.rawgit.com/a11ydoer/practices/afc179d3cf5ae3496d46080b216ab6751613c16e/examples/coding-template/template3.html

<JemmaKu> thanks, James

MK: MB also brought this up

JG: the basic compoennts of bothe templates is the same

JK: let me copy the new link for template 3

<JemmaKu> https://cdn.rawgit.com/a11ydoer/practices/afc179d3cf5ae3496d46080b216ab6751613c16e/examples/coding-template/template3.html

MB: I would go for the checkbox example as I dont like tables

JN: tells people what is implemented and what isn't

MK: tells people which decisions are made and what aren't

JG: if we can reinforce things it is good

AA: i agree with Jon

<JemmaKu> I fixed the coding.

<JemmaKu> https://cdn.rawgit.com/a11ydoer/practices/master/examples/coding-template/template3.html

MB: should add a line which states that this is the support for this implementaiton

MK: on the checkbox example page there is a heading of simple checkbox
... right under there is coding exmample heading

JG: checknbox 1 simple checkbox - there would be no special name other than coding exmaple
... or would we assume that right after h1 the coding example would start'

MK: these are rendered checkboxes so coding example is confusing
... i dont think we need 2 H1s

JG: thinking about SEO putting stuff in the banner would help

MK: H1 in the banner is fine but next in the page should be an H2 - simple checkbox example
... dont know if we always want to number the examples - perhaps name them

JG: checkbox role: simple checkbox

MK: prefer checkbox pattern: simple checkboix

LW: where we have variants perhaps keep 1 per page and then have related patterns

MK: we have the cell in the table where we have examples

LW: also perhaps have related patterns linking to the other

MK: 1 pattern per page

JN: sounds like editorial work
... to add links if we decide we want them

JG: assumed that the example would start after the h1?

MK: fine to have an H2 after it before the example

JK: just an h1 in the main content - don't need a banner

JG: H1 will include checkbox pattern to mark the start of the example?

MK: wanted to put the code in the ewxamples on the agenda?

MB: want to look at the descrtiption in the checkbox example
... is there a name for this style of comments

JK: working on refactoring and the comments
... using jsdoc

MB: like we are using 2 spaces instead of tabs. a bit more compact

IP: commetns about js style
... I think argument for 4 spaced tabs as more commonly used and would make it easier to copy into wider sodebases
... variables that are global which we need to avoid
... all the functions are global which we need to avoid
... I would suggest we stick with semi-colons rather than not using semi-colons
... it will work but you need to understand the rules and not many who say they do actually do

MK: a lot of devs don't understand when it is ok to skip semi-colons so better to have everywhere

IP: reduces cognitive load
... I don't get the argument that it makes your file larger is a good one

+1

IP: in the case of groupboxstate = .... changed to lower case , should set the true or false, but there is a convention of creating a variable with values in upercase and then compare against your "constant" when setting and getting

<LJWatson> +1 to using ;

JK: should i submit code revioew to github and that would be the best way

IP: as we decide need to document them

<LJWatson> +1 to code reviews (friendly ones :)

MK: the purpose of it being on the agenda is to decide the conventions we wantr to have
... if we are going to have them then need to document them

<MichielBijl> Link to git issue: https://github.com/w3c/aria/issues/95

MK: are we going to have a long list of things

IP: can link to existing standarrds
... must pass a jshint lint test with certain settings
... github has their own standards etc.
... would be my approach to link out to things
... using jsdoc is a good decision
... if keep it very simple then can use external tools

MB: I have a csscomb file for everything i create
... for css files

IP: do we have anything on the wiki for how to contribute
... if so it should be there

MK: need exactly what you are saying on the wiki
... trying to go to the participation pages

<JemmaKu> I like git issue Michiel created. This would be very helpful for coding sinceI am not js developer yet

MK: seems like a michael thing that stuff in the pf pages are very out of date

LW: problem i found with github is that our work was in the main repo

<JemmaKu> having some standard coding practice doc would be very helpful.

LW: it still may be possible but need to work out how not to get in other peoples way

<JemmaKu> question to Ian

<JemmaKu> +q

MK: a page on how to contribute would be very useful

LW: more complex than it ought to be as we have multiple lines of work in the same repo

JN: should bring this up on the editors call

MB: before meeting ends
... was working on the alert example - when I started on the code example
... wanted to share the progress

<MichielBijl> http://s.codepen.io/Michiel/debug/0f9ce27964f35b55d3e7628d91874c2e.js

<MichielBijl> http://s.codepen.io/Michiel/debug/0f9ce27964f35b55d3e7628d91874c2e

MB: wanted some feedback
... no way to get a list of focusable elements...
... there is no way to get an array of focusable elements
... wondering if someone had a more reliable way of doing it
... jqueryui guys do it the same way

IP: dont think there is a native way

MB: the way I do it doesn't work.

<JemmaKu> https://developer.mozilla.org/en-US/docs/Web/API/Document/querySelectorAll

IP: does it in the order of your selector
... rather than write the script so it focuses on the first or last. could put a selector in a data attribute and then set focus to that

MK: should read the describedby so long as anything in the dialog gets focus
... the author can specify which element should get focus when it opens

IP: the things which read out a long describedby and then the element with focus. is this something you want to hear? or would you prefer to get a tab press first

MK: jaws will read the description last
... others may do it differently....title then description and then focused element
... always easy to do VO-F3 etc to find where the focus is

IP: you are happy with that
... I would set focus on the container

MK: I think it should always go to the first piece of content

BG: as far as screen reader behaviour goes. aria-describedby should only be annunced once in the dialog but Ie11 w/ JAWS is announcing it on each tab press

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.140 (CVS log)
$Date: 2015/10/12 18:30:55 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.140  of Date: 2014-11-06 18:16:30  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/default file/csscomb file/
Found Scribe: jamesn
Inferring ScribeNick: jamesn
Present: (no one) JamesNurthen JemmaKu jongund AnnAbbott MattKing Raghu Michiel_Bijl IanPouncey Bryan_Garaventa LJWatson

WARNING: No meeting chair found!
You should specify the meeting chair like this:
<dbooth> Chair: dbooth

Found Date: 12 Oct 2015
Guessing minutes URL: http://www.w3.org/2015/10/12-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]