W3C

- DRAFT -

Accessibility Guidelines Working Group Teleconference

24 Jul 2018

Attendees

Present
AWK, MichaelC, Greg_Lowney, jemma, marcjohlic, Rachael, gowerm, Chuck, Laura, Katie_Haritos-Shea, Kathy, jon_avila, kirkwood
Regrets
Brooks, Glenda, JF, Detlev, Jake, Alastair
Chair
AWK
Scribe
mike_elledge, rachael

Contents


<Mike_Elledge> I'll scribe

<jemma> scribe:mike_elledge

TPAC Registration reminder https://www.w3.org/2002/09/wbs/35125/TPAC2018/

awk: Cheap seats are avaialbe until next Tuesday. How many signups?

mg: some issues about registration with credit card

mc: more registrants, but some on call have not yet.

ak: will try to have phone connection, but terrible. Remote attendees probably found challenging.

Techniques for approval & reminder to reviewing draft techniques: https://github.com/w3c/wcag/pulls?q=is%3Apr+is%3Aopen+label%3ATechniques

<jemma> I plan to register.

ak: Hopefully ppl can coem in person.

zakim: next item

TPAC Registration reminder https://www.w3.org/2002/09/wbs/35125/TPAC2018/

ak: variety of techniques here and task forces, but dont' have new ones ready to go.

dm: Auto fill is ready for preliminary review
... has a pull request in github

mg: If try to assign own reviewers github doesnt' do well

mc: has to do with how notifications are set by ppl
... clear when I get it, but hard to tell apart

<daivd-macdonald> https://github.com/w3c/wcag/pull/437

dm: tagged for review 437

mc: labeled with "techniques"

dm: will do that

mc: will show up on list now

Pull 436

ak: 436 is the first on the list
... mike gower, aria role of log. what's needed? Need for review?

mg: yep. just labelled. no comments yet.

ak: this is 413.

<laura> Does it need 5 +1s for WG review?

ak: ppl should assign to themselves so they can be reviewer.

mg: +1 as reviewer also

ak: putting so i can be reviewer.

mg: wondering if working samples working okay in one browser combination but nto others. Do we enter bugs or list them.

ak: chat log. what's not working.

mg: jaws firefox will not actually read log
... add component and it is silent.

dm: aria live not supported in all situations.

ak: in past, since there are techniques with vrying levels of support, have technology notes. also have tech support notes to see for different techniques.

<MichaelC> User Agent Support Notes for WCAG Techniques

<MichaelC> (not yet migrated to 2.1 framework)

ak: if support is low, then probably not worth advocating. but if at or browser bug, just document it.

mg: fyi. role of log is poorer support than aria alert.

<MichaelC> ARIA implementation data

ak: all techniques wind up in one file for aria. have to make connection in technique as we built it.

dm: do we want to discuss this technique now?

ak: 436--log.

<jemma> https://www.w3.org/TR/wai-aria-1.2/#log

dm: yes. aria log-role. lumberjacks!

ak: need to get ppl to look at it first. are there specifics or concerns?

dm: text looks good. example 1 seems to give undefined. 2 talks about aria additions, but it's on by default.

ak: browser

<jemma> https://developer.mozilla.org/en-US/docs/Web/Accessibility/ARIA/ARIA_Techniques/Using_the_log_role

dm: latest chrome on mac giving "undefined"

ak: any thoughts or comments? 435 is the technique, 436 is the examples.

dm: write comments under pull requests?

ak: put them there.

dm: right there on bottom.

ak: everyting to misspelling to different appraoch to description, other details

432

ak: seems like it has had more attention. Highlight?

mg: More attention. hopefully will be ready for review soon. Pink flamingo!

ak: 413, status role. Great to have ppl look at it. main purpose is getting tasks done now.
... should be helping with techniques. need more eyes on it.

417

ak: Jon avila on call

ja: started it.

ak: looks like a bunch of reviews, have talked about it. nothing in last week. challenge was...

ja: don't know what hang up was

ak: usng viewport units. had concerns around whether could always be failure. longer title to capture issue. technique using viewport of vw units won't always fail.
... can we refine failure so does always fail? and how?

ja: last time was okay with suggestions to keep viewport units in there rather than general failure. viewport units do always fail. but can have other examples on page.
... if have those other things then have to have otehr consieerations. otherwise ppl won't be aware of potential failures.

ak: tried to avoid appending failures.

<AWK> https://www.w3.org/TR/WCAG-TECHS/

mg: or use "solely" in name?

ak: a possibility as well.
... pull request 417
... as example, f48 useing tabular unit, might fail but could have correct tabular info as well, or in way that passes. Don't like this example, just has text

ja: doing something only by using it elsewhere. Can say "only by using viewport units". not a precedent. don't want long name, but want to keep it.

<AWK> https://www.w3.org/TR/WCAG-TECHS/F84.html

ja: have generic technique already: 1.4.4 would only be caught under another one, like f80, but is very specific instance

ak: looking at f80, non-specific link, so there's failure unless you do something extra. which is what we're saying here.
... basic html page uwing vw units will cause probs.

dm: outcome-based, if another call can pass.

ja: ppl with concerns on call today/

ak: dk
... propose as title? failure due to ppl relying on sizing text with viewport units without...

mg: could say "alone"

<laura> Survey comments from last week: https://www.w3.org/2002/09/wbs/35422/techs2/results#xq1

ak: if syaing in viewport units alone could use other means to adjust to another vw unit..
... add note that title needs lengthening to add qualifier
... test procedure, is this a failure that only applies to desktop?

ja: yes, in mobile pinch just scales (magnify).

ak: care only if in browsers with text re-sizing without zoom

khs: need that qualifier as well

<Zakim> Greg, you wanted to say we should not rely on behavior of mobile vs. desktop

gl: dislike relying on assumption that mobiel will always scale without reflowing, while desktop will.

ja: agree. if use commands it reflows.

ak: procedure needs to be more generic to not be just desktop

<Greg> We should use the phrases "scale with reflow" and "scale without reflow" rather than mobile and desktop.

mg: agnostic.

ak: get rid of pixels? very specific.

r: yes, makes easy to test. wld rather mobile setting, or 1280, so not so specific

mg: haven't tested, but imagine features for enalrging text willcome into play

ja: havent' tried with viewport units, but may come into play with mac

ak: question about #2, window sizes. Do we need different tests for diff devices. Or pick a window size and work with it?

mg: thought only time using 1280 when describing reflow or 400% zoom. in this case is expansion is beyond conatainer so doesnt' seen to apply

ja: challenge always taht no resolution set. at 200% would have different results depending upon resolution.

ak: makes me concerend that if we specify pixel size for container it isn't part of the success criteria.
... adding criteria.

ja: not such high resolutions when wcag came out. kind of overlooked as issue wrt testing.

ak: one of teh questions is whether to have specific pixels, and what to do with different screen sizes if we do.

<daivd-macdonald> test page http://davidmacd.com/test/lvm-text-sizes.html

ak: #4, says any of following methods...resizing text no longer belong? using zoom and nothing esle?

ja: do one or the other.

ak: could change to a number 5. 4 with first two bullets, then check if onpage controls to re-size text, then check results. wld capture if nothing else has been done.

ja: problem to leave all three in?
... even though title says it, test procedure still allows it.

ak; maybe it does work this way

dm: can confirm that 500% page doenst increase on mac, but on mobile will go larger on safari and iphone 10

ak: on desktop doesn't change text size. same size at 500%. should be a failure. going smaller stays same size. Chrome 65 at 500% at same size as 25%.
... when use command + on Safari does work

ks: expanding screen desktop the text size increases. but not chrome.

<Greg> It does not resize in Vivaldi.

dm: safari does, but chrome does not increase.

ak: need to make sure that we have close to full complement of user agent support as failure

<kirkwood> gets smaller and larger in chrome on mac

gl: vivaldi another browser where does nto change size.

ak: vivaldi based on chrome engine?

gl: don't remember

<AWK> Need to add user agent notes for this technique

dm: Edge, chrome and firefox all honor that. Mobile does not. Windows does. IE too.

<AWK> Need to add live example

dm: just safari on desktop doesn't resize text.

ak: Chrome and windows fail as expected.
... need to provide pixel width and make it less about desktop. also add in user agent notes, and live example.
... ppl not opposing this as a failure.

dm: awesome failure!
... kudos jon!

ak: some progress. let's look at 389

389

ak: technique for sticky headers and footers. positive technique? but some comments. jake, alistair,

ja: its the landscape issue.
... this can happen whether in landscape or not. If zoom in may see short window where you can see body. only a line or two.
... originally a landscape issue, bec not as much height, but applies else where now.
... wcag doesnt' say have to have sepcific height line, but makes it very difficult to read.

ak: orientation sc failure?

ja: not really orientation, more text resize, 1.4.4. although strictly speaking nto a failure.

ak: could have header sufficiently large but faile in landscape. could fail even if line of text not as zoomed in.

ja: less than one line of text is possible. usually a few lines. but such a small view into page hard to keep track what's going on. Usability issue.

ak: advisory for 1.4.4?

ja: could apply to orientation if page cannot be operated. but would be extreme case. could be 1.4.4. but in less severe cases would be advisory in stead of failure.

ak: any other comments?

dm: in general it was difficult to make it work with sticky headers and dropdown. gets messy. certainly a place for this.
... releasing sticky headers when there is a ceratin header would make sense

ak: needs ppl to review it. additional changes taht can be suggested. Alistair is making changes as well.
... is phrased as a postiive technique, ie. hwo to avoid failure. doesn't say which SC it is for--which it should.

388

ak: textbox reflow.
... jon, jake, have looked at it. also does not seem to say which SC is failing.

l: put it automatically.

ak: reflow. additoinal context for the group?
... haven't looked at this one. Why two procedures?
... read vertically vs. horizontally. would be good to merge into one procedure.

ja: trying to clarify in original text. for horizontal text just need horizontal reflow.

<scribe> scribe: rachael

ak: This needs some people to look at it. Please add yourself as a reviewer

Topic 387

387

<jon_avila> I have to drop

ak: Tech media query reflow
... There is just one procedure. Comments?
... Three people have said yes its ready to go. We need a few more. Please review it if able to get it to 5 for next week's call

386

Tech spacing override

<gowerm> https://github.com/w3c/wcag/pull/386

ak: This is the one we've been going back and forth on.

laura: I updated the description and examples to make it more like a technique.

ak: Is there a link to examples?

<laura> https://rawgit.com/w3c/wcag/tech-spacing-override/techniques/css/spacing-override.html

Laura: What authors should do and an example to show how to do it.

dm: Is there any algorithm to recommend? It would be very helpful.

laura: Alastair looked into it and no.

dm: The bookmark show if it fails but it won't tell people what to do. Can we come up with a way to recommend changes based on the container?

<gowerm> Ensuring that text containers resize when text resizes

mg: I suggested not to pick up the 1.4.4 language but to make this the technique. (See above)

<laura> Mikes’s proposal: https://github.com/w3c/wcag/issues/384#issuecomment-407406967

mg: What I was suggesting what to take the preamble, peel off the techniques from that. Gave example .Create generic techniques that work well with existing techniques. This is captured on github.

Laura: You can use em and still mess it up.

mg: If you have a generic technique about the containers, then you can use the other techniques in addition

<jemma> jemma: it makes sense to me - container resizing goes with text sizing.

dm: I'm not sure that makes sense. It really is all about the containers. I want to hear about the characteristics of the containers, not the text. How can you mess up on boxes?

Laura: Those should be a failure technique.

mg: I am suggesting the same. The general technique wording I pulled out is: Ensuring that text containers resize when text resizes

dm: What would the techniques then be?

ak: The techniques we have do this. They all size or do not size the container.

<gowerm> https://github.com/w3c/wcag/issues/384

ak: Its all relative. The examples you have are good. Its a question of how to package it. I haven't had a chance to look at it, but can.

mg: Issue 384 is tracking this separate from the pull request.

ak: Mike has engaged with this as well as others. Anyone else who wants to be involved speak now or jump into the review process becuase we need to wrap it up.

mg: The content here is really good. Its a matter of packaging and getting a few good techniques.

ak: Topic: 380

380

Ensuring single pointer usage.

<gowerm> https://github.com/w3c/wcag/pull/380

ak: Everything seems to be fairly positive. Jon avila is asking about drag and drop. We need to find resolution around 403, which we did so we need to revisit this with that resolution in mind.

This is ready for a little more work but assuming the drag and drop is addressed, what we need is more people to look at it.

437

Tech autocomplete

<gowerm> https://github.com/w3c/wcag/pull/437

dm: Simple technique to show how to implement the autocomplete feature.
... Technique and some examples. I've incorporated comments from two people to date

ak: Basically going back to what we're trying to accomplish with this technique and how to do it. I focused on it helping providing a way for anything to get access to the form fields.
... Looking at the procedure, it says check to see if the form fields have complete values. Some of our techniques have leading text, "for form fields that identify the user..." We need to make sure the fields also have a corresponding autocomplete value
... The mapping has to be called out. Its not every form field.

dm: I can put that in a sentence leading up to the steps.

ak: That would be good.

dm: This is not necessarily a functional test. This is more an inspection test. In 2.0 we have more inspection tests.

<gowerm> For each form field that collects information about the user and corresponds to an autocomplete field described in the HTML 5.2 specification, confirm that:

ak: We need more reviewers for this. Is it a nit to call out the HTML 5.2 specification?

dm: Lets use the wcag 2.1 pointers.

AK: Lets go to the next item unless there are more comments?

zakim take up next item

Feedback on process reminder https://www.w3.org/2002/09/wbs/35422/processfeedback/

<Ryladog> KHS: Do we need to identify in the test procedure to place focus i forst feild and begin to type? AWK suggests that is best in User Agent notes. DM says that is functional testing

Errata update

dm: Should we be booking for TPAC?

ak: Book hotels now. If you can, book flights early.

<MichaelC> TPAC agenda

mg: Best option I found is to fly into Paris and take a train.

mc: We don't have an agenda. The recommended hotels are on the main page.

ak: Has anyone else booked a flight? I was considering Geneva

dm: Has anyone crossed that border? Geneva is close

<Ryladog> Crowne Plaza Lyon - Cite Internationale, I got a good dea

ak: I have, it was trivial

k: I have my hotel, the crown plaza. The prices were reasonable. The prices for Europe go down as we get closer to august. I am also taking additional time in London

mc: There are 3 hotels on the site right by the meeting location. One has a negotiated booking link that expires 20 Sept. They can run out of spaces. There are other hotels suggested

<gowerm> https://www.w3.org/2018/10/TPAC/Hotels-Transportation.html

TPAC Registration reminder https://www.w3.org/2002/09/wbs/35125/TPAC2018/

Errata update

ak: We found a bunch of editorial errata and we'd like to make sure we publish that because that is what we've done for 2.0 in the past. Some are embarrassing spelling mistakes. All editorial.

mc: We already approved some editorial changes that are in the editors draft. They should show up in the errata page.

<MichaelC> WCAG 2.1 Errata

<MichaelC> In the Introduction, several (but not all) references to "WCAG 2.0" should be "WCAG 2.1".

<MichaelC> In the 0.5.2 Numbering in WCAG 2.1, the words "critera" and "ccriteria" should be "criteria".

<MichaelC> In Success Criterion 1.4.10 Reflow, the first note had a supernumary "Note" indicator which should be removed.

<MichaelC> In 5.3.1 Required Components of a Conformance Claim the editorial note "In WCAG 2.0 this was a dated URI, which may need to be adjusted when this becomes a Rec." should be removed.

<MichaelC> In the definition for keyboard interface, the second (of three) note should be an example of the first note, leaving only two actual notes.

<MichaelC> In the definition for technology, the third note should instead be an example.

<MichaelC> In 7. Input Purposes for User Interface Components, the word "county" should be "country".

mc: The page is currently blank because we haven't approved any. Bullets above are proposed
... Andrew asked about the last one because it is an error copied from html. It is in the editors draft. We can decide to retain the error from html or change it.

ak: Is it actually an error?

mc: I think it is.

ak: there is a telephone number with county code...

mc: I don't think counties have telephone codes.

ak: It isn't in the errata for the html spec.

mc: It hasn't been filed.

ak: we should let the html folks know so they can add it to their errata.

mc: Will add to html github issue list.

ak: This is changes to the normative doc. We should send this list around. Can we create a pull request for the errata page?

mc: Its managed separately. I would need to set up a seperate issue.

<marcjohlic> +1 to cfc'ing them

ak: Do people agree these are editorial enough that we can run a cfc?

<laura> +1

+1 to cfc

ak: does anyone think otherwise?
... We will handle this as a cfc

RESOLUTION: MC will run the list by the group as a CFC

Feedback on process reminder https://www.w3.org/2002/09/wbs/35422/processfeedback/

ak: We have a survey. I checked about 2 ours ago and only 1 person had responded. We really want everyone's feedback and thoughts on. Please take some time to provide feedback on this.

mc: Closes mid-late august

mg: We will be opening the formal window next Monday through 24 august. If people are concerned that people will be away during that time, please let us know. The survey is open now if you want to send responses early.

<MichaelC> Filed issue on HTML https://github.com/w3c/html/issues/1571

AOB

ak: Anything else to ask, raise, discuss?

trackbot, end meeting

Summary of Action Items

Summary of Resolutions

  1. MC will run the list by the group as a CFC
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.152 (CVS log)
$Date: 2018/07/24 16:45:55 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.152  of Date: 2017/02/06 11:04:15  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00)

Default Present: AWK, MichaelC, Greg_Lowney, jemma, marcjohlic, Rachael, gowerm, Chuck, Laura, Katie_Haritos-Shea, Kathy, jon_avila, kirkwood
Present: AWK MichaelC Greg_Lowney jemma marcjohlic Rachael gowerm Chuck Laura Katie_Haritos-Shea Kathy jon_avila kirkwood
Regrets: Brooks Glenda JF Detlev Jake Alastair
Found Scribe: mike_elledge
Inferring ScribeNick: Mike_Elledge
Found Scribe: rachael
Inferring ScribeNick: Rachael
Scribes: mike_elledge, rachael
ScribeNicks: Mike_Elledge, Rachael
Found Date: 24 Jul 2018
People with action items: 

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


WARNING: IRC log location not specified!  (You can ignore this 
warning if you do not want the generated minutes to contain 
a link to the original IRC log.)


[End of scribe.perl diagnostic output]