W3C

Accessibility Guidelines Working Group Teleconference

25 Jul 2017

Agenda

See also: IRC log

Attendees

Present
AWK, KimD, JakeAbma, Laura, ChrisLoiselle, JF, steverep, jasonjgw, MikeGower, Greg_Lowney, Melanie_Philipp, Makoto, dboudreau, Detlev, wayne, WayneDick, chriscm, lisa, bruce_bailey, MichaelC, Rachael, kirkwood, marcjohlic, Pietro, jon_avila, alastairc, JMcSorley, David-MacDonald, Kathy, JanMcSorley, Katie_Haritos-Shea, Elledge, Mike
Regrets
Denis_Boudreau, Glenda_Sims, Jake_Abma, EA_Draffan, Bruce_Bailey, Mike_Elledge, Chris_Loiselle, Kim_Patch, AWK, Joshue
Chair
MichaelC
Scribe
alastairc

Contents


<MichaelC> s/In content implemented using markup languages, the conventional name of conventional user interface components can be programmatically determined. (AA)//

<scribe> scribe:alastairc

Target size https://www.w3.org/2002/09/wbs/35422/target_size/results

Announcements

<MichaelC> Silver meetup

MiichaelC: When at TPAC in Nov, there will be a bay area meetup on the Silver work. Expect people would like to participate, details closer to the date, it is on Thursday evening.

MichaelC: Any other announcements? Nope.

Target size https://www.w3.org/2002/09/wbs/35422/target_size/results

Personalization https://www.w3.org/2002/09/wbs/35422/LatestPersonalization/results

<david-macdonald> In content implemented using markup languages, the conventional name of conventional user interface components can be programmatically determined. (AA)

MichaelC: 8 comments, 4 supporting. A few against, comments?

Jason: The viability depends on the definitions, there's a lot of work there. The acceptability / effectiveness will depend on detailed work that hasn't been done. A good horizontal review would be desirable soon, get it into the draft would be a good way to do that. Still has issues that won't be resolved for a few weeks.
... For example, 'conventional ui components', then compare to the HTML5 spec equivalent (auto-completion), there's a more rigorous approach. That's a good example to use.

<lisa> new titile is Purpose of controls:

MichaelC: Bruce suggested a title change, Marc has issues with the words listed as definitions and whether they are examples or not?

Lisa: The title was meant to be "Purpose of controls".

<david-macdonald> hit refresh for latest version of answers

Lisa: It will help to achieve personalisation, support the title change.
... We do need to harmonise, we know about the HTML5 and ePub examples.

MichaelC: Bruce isn't here, but there is a proposed title change.

<lisa> +1

MichaelC: Does anyone object to putting it into the draft?
... i think Marc Johlic's comment was similar to Jason's, but do people present object to putting it out for review?

gowerm: I don't think we should put this out for review, it seems like we're just putting it out.

<Alex_> is the audio dead or just me?

<david-macdonald> my audio is fine

Lisa: We had a call to try and address everyones concerns. The thing under scrutiny is the definitions now. We had some quite different opinions on the call, and everyone was happy with the core SC text. The definition is there form the last couple of years, it isn't new.

<Detlev> is it going to be purpose of controls of personalisation?

<Detlev> Or

<Pietro> * audio on WebEx by browser is good

David: Echo Lisa, there was quite a bit of scrutiny, understanding the definitions need work.

<MichaelC> ac: definitions were to address JF objections

<MichaelC> list things in use today

<MichaelC> the subgroup liked, just needs cleanup

<Zakim> MichaelC, you wanted to ask how much time to clean up defs and to say defs are new from SC perspective

<Detlev> "Purpose of controls" won't reflect the personalisation issue if this is core

<lisa> from 2 weeks ago

MichaelC: Whilst the terms in the definitions have been used elsewhere, they are new to the SC. How long will it take to clean them up? If it's a week lets come back to it then?

<Alex_> p

Kathy: how does this support personalisation? We have a list of these in English, will that be for all of the languages?

Jason: To answer Michael's question, not sure how long it would take, but I think it would take a while, several weeks of work probably? Needs to be specified closely. Also, on the public review side, the reason here is we need more (external) review to progress.
... we can include notes to indicate the issues. This is on the side of external expertise.

gowerm: I really believe that this is a new version, it should have the same scrutiny as everything else. To early for a CFC.

<MichaelC> ac: would like to put out with ednotes

<Alex_> Michael, you dropped me off the queue

<MichaelC> worth getting feedback on concepts

<david-macdonald> can't hear you Alex

<Alex_> i'll have to dial back in

Lisa: In terms of list being new, it was in the last draft as well, in the definitions available from the SC. It was the same list. It's not so new. It was similar to the second bullet in the earlier draft. It was tweaked to address everyone's concerns.
... alastair spoke to how it was addressing personalisation, so just asking for part of it, just one section of it.

<Zakim> MichaelC, you wanted to test ¨We´re almost there but need improved definition wording to approve¨

MichaelC: we have 5 min left on this topic, can we consider another approach to next steps? Not sure we have solid consensus for this yet, and it wouldn't be published this week in any case. Let's clean up on the definitions we know need work already, give it a chance to be socialised a bit more.

Alex_: I don't think this is ready, we need some more time, we need a new survey to look at carefully. Let's wait.

Kathy: Clarification, we have conventional names for three different categories, or is it English based?

markjohlic: Are we saying that these are the only words that could be used?

Lisa: We'll need to harmonise them with HTML5 spec, but we wanted a very clear scope so people knew what to do.
... these are programatic names, can use coga-semenatics, or through controlled vocabulary, or other ways of meeting them. Could do it in any language right now. Could be in the text label as well.

Kathy: There are other conventional names, this is a US-English based list.

MichaelC: Encourage people to address comments on i18n and clarity, and get people to look at it in the next week, go for a decision next week. Is that a plan for the moment?

RESOLUTION: Partially clean up definitions and bring back next week.

Target size https://www.w3.org/2002/09/wbs/35422/target_size/results

<steverep> Can we please make sure to get the version desired to be reviewed in GithHub? Thanks.

<lisa> thanks :)

Detlev: Wanted to clarify, Andrew is SC manager and we had comments from Greg about links in lists, where they get very long. It's a trade off between target size & usability. Thought we should allow for lists to be closer together.
... the text in the survey doesn't take that into account. My comment in the survey includes some text for that, maybe not good text but initial idea.
... another thing addressed in the response was about the CSS hack technique which extended inline target sizes. That lead to the proposal to remove the AAA SC.
... There were concerns about older pages which wouldn't meet it, so there are exceptions, including for blocks of text.

MichaelC: Does addressing that affect your view on the AAA SC?

Detlev: I'm not sure we should keep it. I'd like to keep it, but removing it addresses StephenR's comment.

MichaelC: 3 people didn't want to remove the AAA sc.

<lisa> +1 to cathy

Kathy: A lot of people would prefer to keep it, but could live with removing. Concern that at AAA you don't have links in paragraphs. But isn't AAA's purpose to show things that are desirable, but may not always be achievable? Yes, there is extra work, but that was the point.
... that said, if it helps to get the AA SC in, then people would prefer that.

McihaelC: So the AA version isn't accepted?

Kathy: No, was removed previously.

jasonjgw: I think my comment has been covered: AAA is for things that are not always possible but desirable, so I don't think the objection is reasonable.

<Zakim> steverep, you wanted to say that if we keep it, then acknowledge not using CSS negative margin technique

steverep: I'm ok with leaving the AAA version in, I'm just not happy with the CSS negative margin technique being in, that has issues. Two big issues: can't guarantee it wouldn't overlap, and messes with focus indicators.

david: There's an exception for blocks of text, so the negative CSS example isn't needed.

Kathy: It wasn't a technique, it was an example to show it is possible. Not required, no technique coming for that.

David: The missing thing is about the stacks & menus, left-nav bars. I think we need to go to 22px for stacked/grouped links in that scenario. (at AA).

MichaelC: Can we resolve the AAA SC?

Detlev: Happy to keep the AAA if others are not objecting.
... for the stacked links, there are concerns around i18n and different text directions, so need to be careful of language.

gowerm: I assume we are focusing on the draft SC language, aren't the techniques the focus?

<Detlev> +1 to keep AAA

<steverep> +1 to Michael on implementability means there must be viable techniques

MichaelC: Yes, so long as there are techniques to fulfil it. We shouldn't block SC due to techniques, so long as we're happy they are feasible.

<Zakim> MichaelC, you wanted to ask who can´t live with keeping AAA version

MichaelC: a number of people have said they're happy to keep the AAA. Would anyone object to keeping the AAA version?

<gowerm> yes

<lisa> +1

MichaelC: ok, so we have a tentative decision to not drop the AAA version. I think we can circle back to the AA version and it's exceptions.

<Detlev> Correct

+1

MichaelC: we have a new exception from Detlev, what's the summary of that?

<Kathy> +q

David: When you have a stacked set of links, like a left hand menu, increasing to 44px tall will create a lot of scrolling, so want to make an exception for that, 22px x 44px minimum.

Kathy: One of the things we need to add, addressing Detlev's comments, and the i18n aspect: When there are groups of controls that will cause an increase in scrolling... if we had something to quantify the scrolling, we could then drop the size to 22/44px. There are situations with 44px across and horizontal scrolling we need to be careful of.

<kirkwood> +1 to Kathy

Detlev: We have discussed the figure +10 controls, but it's problematic. There are different screensizes, content management systems with varying numbers of items in each menu, it's quite difficult to put in a fixed figure.

<kirkwood> +1 to DetLev

Detlev: need to define a stack grouping, rather than a number of items.

David: Just don't know how you could evaluate it against a set screen size. Would love to get to it that way, just not sure how.

<david-macdonald> +1

Kathy: Don't want to have something with 22/44, when could be 44/22. How about at least 44px on one side?

<kirkwood> +1

<Kathy> targets in stacked lists have at least one dimension that is 44 pixels and the other is at least 22 pixels

steverep: We could use a cap on the total number of pixels in a list? E.g. 320px.

<Zakim> MichaelC, you wanted to say need to circle back on refined def for new exception?

<Detlev> +1 Fine with me

MichaelC: Does someone have a wording ready for today? Kathy's at xx.55:27.

<david-macdonald> +1

<Kathy> +1 fine with me

<marcjohlic1> +1

MichaelC: Any objections?

<Pietro> +1

Alex_: I think 'stack' means up/down, rather than horizontal?
... in that case stack isn't the right word.

MichaelC: That's an open issue.

<david-macdonald> targets in grouped lists have at least one dimension that is 44 pixels and the other is at least 22 pixels

<Kathy> targets in horizontal or vertical lists have at least one dimension that is 44 pixels and the other is at least 22 pixels

steverep: Is the goal to not impede the authors ability to layout stacks?

<gowerm> How about just "targets in lists have at least..."

<kirkwood> if concern is to limit scrolling maybe we should just say that?

Kathy: don't want the user to have to scroll a menu. E.g. 10 links at 44px high is 440px, very high. We see it in vertical situations, but applies in both directions.

<Kathy> targets in a list have at least one dimension that is 44 pixels and the other is at least 22

<Detlev> @steve: It's too prescriptive

steverep: Need to define a total pixel dimension that doesn't impede it, and then tell author to maximise the dimensions within the list. cutting in half may not be necessary, it could be 32px high.

gowerm: I'd prefer not to have this exception because, for menus, it is a vital control. We have mechanisms for people to make things larger. Many of the exceptions are not for critical controls, whereas a menu is a critical control.

<Kathy> grouped targets with more than 5 items have at least one dimension that is 44 pixels and the other is at least 22 pixels

david: I think we need to explicitly say this is an exception. Making groups of links min of 44px is difficult. Language around the size of items could be different across evaluated. It's on every page, on every website.

MichaelC: I think we've agreed to keep the AAA version, and most people are happy with the AA version. But some work on wording needed for everyone to be happy.

<david-macdonald> Are we close enough for consensus in a draft? Only 3 meetings left before cutoff?

<gowerm> yes

<Detlev> +1

<david-macdonald> +1

<gowerm> +1 to note in editor's draft

Kathy: as we have to get outside feedback, and passed it once, can we include it with a note in the draft?

<marcjohlic1> +1 to editor's draft w note

<david-macdonald> grouped targets with more than 5 items have at least one dimension that is 44 pixels and the other is at least 22 pixels

MichaelC: For the public draft it wouldn't appear for a while anyway. Do people want to add it to the editors draft?

<david-macdonald> +1

+1

<Detlev> +1

RESOLUTION: Keep the AAA version of target size

Alex_: How do you mean open for exception.

<lisa> +1 to keeping the tripple a

<david-macdonald> grouped targets with more than 5 items have at least one dimension that is 44 pixels and the other is at least 22 pixels

<gowerm> +1

RESOLUTION: Accept AA version with editors note about exception for groups of links

Confirm Important Information https://www.w3.org/2002/09/wbs/35422/confirm-impt-info/results

Lisa: this started as a modification of 3.3.4, so there does appear to be overlap. However, some of the choices in there don't help people with cognitive issues.
... there were three items that seemed acceptable, and Mike suggested the basis for the current wording.

gowerm: The existing error wording has: this is one of the three possible ways of doing things. Anytime you are entering data to complete a transation, you get a read-only view to confirm.
... normally you can cancel and go back and edit it. This is to enable people to have a simple read-only version to understand before submitting.

MichaelC: Is this intended to be merged during normalisation?

gowerm: could stay separate, allows for a different scenario.

Lisa: but then 3.3.4 becomes redundant, think it could be merged.

jason: my comments are in the survey, but want to focus on definition of transaction, and what is required in providing a read-only summary, and one of the way to implement would be a confirmation step users have to proceed through. That's a sig advantage to some people, but disadvantage in real-time situations for other people with AT.
... allowing people to review the submission is an important principle, but the proposal doesn't spec the circumstance it is appropriate for.

MichaelC: I think Jason's comments circle around defining a transaction, if that were clear then concerns about being too broad might be addressed. I thought it was clear, but seeing that it's fairly generic now.

Jason: that's about half of them, the others are what needs to be provided. E.g. could a tick-box undermine the read-only aspect? Also, imposing a confirmation step (one way of implementing it), will slow some people down in time-sensitive situations.

gowerm: some (a lot of) of that can be addressed in understanding. For example, github gives you a preview (read only) version, so that's a way to implement it. The transaction definition is a good idea, but we do have quite a few SCs that use it already.

Rachael: Can we setup a separate call for that definition?

MichaelC: Noted.

Mike_Ellege: please talk about the read-only aspect?

gowerm: The intent is that it's read-only, approve/cancel, final review with minimal complications.

david: This could be a huge ask, we want to balance benefit with change, as all forms are affected.

MichaelC: That ask depends on the definition of 'transaction', we need to work that out.
... Alex mentioned this is an unrealistic change to an established SC. If it were to merge into an SC, then it does raise the alarm for Alex's comment.

gowerm: I don't have a problem with taking the wording from 3.3.4, and if you had a single-A version, you could meet this without meeting 3.3.4. So this is a more specific verison.

MichaelC: If we're clear it is separate that affects the evaluation.

Alex_: Agree with moving 3.3.4 to single A, but not sure how you define a transaction that doesn't make it a huge requirement.
... 3.3.4 is established, that can move to single A. Having this as a separate thing on top of 3.3.4 just isn't doable, except perhaps a triple-A.

Lisa: It might be more useful to go to other topic?

MichaelC: Who shares Alex's concern this is too big of an ask?

<KimD> +1 to David's comments

David: There's a trade-off, not a clear win.

<Detlev> I thought some new SCs might be merged anyway ?

<laura> +1 to David.

Rachael: If it were scoped to multi-page forms, would that change the answer?

<marcjohlic1> That sounds more reasonable - multi-page req

<lisa> good idea

<Detlev> Multiple steps rather than pages?

<laura> single page would help.

david: wouldn't you end up with a lot of content for review?

<kirkwood> Multipage into one page would be helpful because it wouldn’t itme out on each page, david

<Alex_> has anybody used ERP system

<lisa> david, it is often, for example your tichet and hotel . o

<Alex_> everything is a transaction

gowerm: I see this as common practice now, for things which have legal implications, or deleting data. E.g. buying things from Amazon.

<kirkwood> +1 MikeGower aying not unusual, agreed

david: didn't realise it was limited to legal things, is that the case?

MichaelC: I heard that people were more or less favourable, but clearly the definition of 'transaction' needs refining. Also need to make sure the scope is clear, so it isn't every form. Is that ok?

Alex_: In enterprise areas, every action is a transaction, e.g. moving goods and materials.

RESOLUTION: Continue work on transaction definition and scope

MichaelC: There was a question about AAA proposals...

Consider some AAA?

<Alex_> sorry have to run

Lisa: Some SCs from COGA were rejected at AA, we were hoping they could be considered at AAA, then the information would be available in the spec.

MichaelC: Any reactions? Question for chairs soon.

<lisa> not problems with testability

jasonjgw: It is question for chairs, but it is question of what the issues were. If it's about scope to content that is ok, but testability is a problem regardless of level.

<Mike_Elledge> bye

MichaelC: Chairs will need to look at this, but not hearing objections, so let's figure out how it could work.
... hopefully a chair will be hear Thursday, thanks for approving one and reviewing others.

trackbot, end meeting

Summary of Action Items

Summary of Resolutions

  1. Partially clean up definitions and bring back next week.
  2. Keep the AAA version of target size
  3. Accept AA version with editors note about exception for groups of links
  4. Continue work on transaction definition and scope
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.152 (CVS log)
$Date: 2017/07/25 16:36:58 $

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)

Succeeded: s/In content implemented using markup languages, the conventional name of conventional user interface components can be programmatically determined. (AA)//
FAILED: s/In content implemented using markup languages, the conventional name of conventional user interface components can be programmatically determined. (AA)//
Succeeded: s/In content implemented using markup languages, the conventional name of conventional user interface components can be programmatically determined. (AA)//
Succeeded: s/w="w3c-wai-gl"//
Succeeded: s/spoke ot how/spoke to how/
Found embedded ScribeOptions:  -final

*** RESTARTING DUE TO EMBEDDED OPTIONS ***

FAILED: s/In content implemented using markup languages, the conventional name of conventional user interface components can be programmatically determined. (AA)//
Default Present: AWK, KimD, JakeAbma, Laura, ChrisLoiselle, JF, steverep, jasonjgw, MikeGower, Greg_Lowney, Melanie_Philipp, Makoto, dboudreau, Detlev, wayne, WayneDick, chriscm, lisa, bruce_bailey, MichaelC, Rachael, kirkwood, marcjohlic, Pietro, jon_avila, alastairc, JMcSorley, David-MacDonald, Kathy, JanMcSorley, Katie_Haritos-Shea, Elledge
Present: AWK KimD JakeAbma Laura ChrisLoiselle JF steverep jasonjgw MikeGower Greg_Lowney Melanie_Philipp Makoto dboudreau Detlev wayne WayneDick chriscm lisa bruce_bailey MichaelC Rachael kirkwood marcjohlic Pietro jon_avila alastairc JMcSorley David-MacDonald Kathy JanMcSorley Katie_Haritos-Shea Elledge Mike
Regrets: Denis_Boudreau Glenda_Sims Jake_Abma EA_Draffan Bruce_Bailey Mike_Elledge Chris_Loiselle Kim_Patch AWK Joshue
Found Scribe: alastairc
Inferring ScribeNick: alastairc
Agenda: https://lists.w3.org/Archives/Public/w3c-wai-gl/2017JulSep/0370.html
Found Date: 25 Jul 2017
Guessing minutes URL: http://www.w3.org/2017/07/25-ag-minutes.html
People with action items: 

[End of scribe.perl diagnostic output]