Accessibility Guidelines Working Group Teleconference

10 Dec 2019


AlastairC, Jennie, jon_avila, JakeAbma, JustineP, MarcJohlic, janina, Raf, stevelee, Detlev, JohnRochford, Chuck, Laura, mbgower, johnkirkwood
Bruce, John, DavidF
JakeAbma, Jennie, Detlev


<JakeAbma> scribe: JakeAbma

<Jennie> Scribe: Jennie

<JakeAbma> scribe: JakeAbma

Conformance Challenges update

AC: any updates?

Janina: number of issues ready for feedback and if fine to be closed

<alastairc> Latest draft: https://w3c.github.io/wcag/conformance-challenges/#introduction

<stevelee> link?

Janina: will change heading structure for document, Michael helps us with that

Peter: really love folks to do is to mention which issues are something to really talk about
... and if some parts need to be changed before first draft publishing

<alastairc> Any issues now, post issues in github. Then survey for next week, specifically on FPWD issues.

<janina> Editor's Draft: https://w3c.github.io/wcag/conformance-challenges/

Peter: work on this for this week, new survey will be made for next week then

WCAG 2.2 Accessible Authentication https://www.w3.org/2002/09/wbs/35422/accessible-auth/results

<alastairc> https://docs.google.com/document/d/1J3NFw6NPyj7QGddBtRmagrtS-x4t9BWan8PYEdSpMZM/edit



<Detlev> Link in surves and IRC are the same I believe: 4th version


AC: lets start with survey one: https://docs.google.com/document/d/1dk4pYxq9vxj2E1GfeH6RGWvqvtCyGfLGDVFopNf8B_U/edit

John: will we keep touch in there or delete?

AC: touch has overlap with gestures, so might be deleted

John: if path based gestures are in the air, then if needs to be in the text
... no overlap with hands in the air, as it is not ' touch'

<JohnRochford> Wow, great examples, Jennie!

Jennie: I recommentd keeping touch
... multiple ways of haptic feedback and ways of responding to this feedback, for the purpose of tranfering information

AC: will haptic cover both?

Jennie: not sure from AT perspective
... I like the concept of possible inputs and outputs, but there's also the decoding aspect

<stevelee> +1

Jennie: we need to keep cognitive and coding in there

Raf: there have been some concepts of rith on the phone


<alastairc> tapping / rythm

AC: ok, happy to keep touch in

MG: we have keyboard and pointer gestures SC already, some issues should be added to odality SC not authentication
... would like a modality SC

<Detlev> +1 to Mike

<alastairc> Note: Entering information into an input field does not in itself require a particular modality as it could be accomplished by keyboard, voice input, or any other accessibility supported method. The success criteria is triggered when a particular modality is required by the content, for example, requiring a password triggers the cognitive function test modality, entering the password in an input field does not.

Steve: the cognitive challenge should be important

AC: but biometrics are also important, and the reason we added them

MG: we're talking about web here, it's strange to come up with biometrics right now

AC: will undermine the modality approach right now

Steve: separation might be a good thing, not all in single list

Jennie: when I hear cognitive load, that is different than cognitive function test
... I agree we can remove input / output components, focus on the cognitive part, make it more discreet

<Zakim> mbgower, you wanted to say does anyone have the SC wording from the 2.1 proposed SC?

<alastairc> https://github.com/w3c/wcag21/issues/23

<JohnRochford> Previous version, from 2016: An alternative user authentication is available for users who are unable to use the primary user authentication method, unless it can be shown that all users have access via the primary method. This alternative user authentication method does not rely upon the user's ability to do any of the following: memorize character strings, including memorizing correct spellings; or perform calculations, such as including correctl[CUT]

<JohnRochford> entering numbers and letters from a character string; or speak; or reliably produce gestures; or recognize characters presented on screen, and then enter them into an input field.

AC: summary not to rely on cognitive function test , unless there's an alternative

John: biometrics is the only one without cognitive test of some sort

<JohnRochford> Correction: Steve, not John, said that.

<Zakim> mbgower, you wanted to say the yahoo method is just a prompt

<johnkirkwood> what about the interpretation of strange looking characters/letters

MG: landscape is changed a lot, and re-authentication is much easier, we should focus on the cognitive part

<alastairc> If a method of authentication relies on any one of the following modalities, at least one other method must also be available that does not have cognitive function tests and also does not rely on the same modality.

<alastairc> https://docs.google.com/document/d/1dk4pYxq9vxj2E1GfeH6RGWvqvtCyGfLGDVFopNf8B_U/edit#

<JohnRochford> Alternative SC text: If a method of authentication relies on any one of the following modalities, at least one other method must also be available that does not have cognitive function tests and also does not rely on the same modality.

<jon_avila> I've heard of phishing scams where that bad person call someone and then says they are with the X company and that they will send a text messages to your phone to confirm they are real and to answer yes. When they answer yes in the text message the bad person is then logged in and the then repeats information from the persons account to gain credibility.

<alastairc> If authentication relies cognitive function tests, then there is another method of authenticating.

<Jennie> It is also format - not just the number of characters. E.g. space included between numbers?

<Jennie> Scribe: Jennie

Detlev: Rereading text pasted above.
... relying on one of the following modalities...

<mbgower> And is the ability to copy/paste considered to be covered by "transcribing characters"?

Detlev: the question is when does the method of authentication rely on a modality.
... e.g. only biometric.

<JohnRochford> Most common password is a blank one: https://www.forbes.com/sites/daveywinder/2019/09/21/the-most-common-and-insecure-password-revealedits-not-123456-or-admin/amp/

Detlev: does this mean only cases where there is no password or other keyboard accessible way of authenticating.
... is that our focus?

<JakeAbma> +1 to Detlev, as my comment in the survey is exactly the same

Detlev: does rely mean that is the only way of authenticating?

Alastair: Put aside things that are already accessible. Think about remembering a password, transcribing information from one place to another.

Detlev: that seems to mix modality of sensory input and output with the cognitive pieces

Steve: Isn't the idea now to replace the word rely?

Alastair: The feedback I'm hearing is it is hard to understand the categorizations, especially on the web.
... We are more restricted with what websites can do to authenticate people.
... What I heard proposed is to say "If it relies on cognitive function tests"...

Detlev: so we would need to more clearly define what is a cognitive test?

Alastair: Yes, there is a definition of it.

<JohnRochford> Cognitive function test text: Cognitive function tests: Include memorizing character strings (except for your own name, email address, or national identification number); using correct spelling; performing calculations or puzzles; and transcribing characters into an input.

<mbgower> I put a comment into the doc which actually breaks it down into JUST measures of cognitive function

Alastair: but that's the principle.

Steve: yes, I'm happy with that.

Alastair: in some senses, we would not be covering certain future situations that modalities might help with,
... but it might make it a more understandable SC for the current environment.

<JakeAbma> new text still says " relies on any one of the following modalities"

John R: Yes, and along the way we got feedback that said we should include these other modalities

scribe: I think removing them is a much better idea.

Alastair: Mike has put an idea into the Google doc.

Mike G: it sounds like we don't want any of these to be done.

Alastair: If people have any further comments, and we will need to update the Understanding Document
... Does anyone object to pairing it back to the original intent?

<mbgower> +1 pare it back

Alastair: John if you can have a look at Mike's comment, that's a good starting point.

John R: I will do so.

Alastair: Any other comments?

WCAG 2.2 Confirmation before submission https://www.w3.org/2002/09/wbs/35422/wcag22-confirm-before-submission/results

Alastair: this is a new one.

<alastairc> https://docs.google.com/document/d/1pcg6ixAfuwlo6jb2tkZBGTDhF0fAiO49h21E6HCbQ6I/edit

Alastair: I see Jake did lots of comments in the document.

Jake: I did paste them in the Understanding also

Alastair: quite a few of the comments you made into changes in the document?
... was there anything in the survey you didn't change in the document or add?

Jake: I placed all comments in the document also.

Alastair: A few things from Steve

Steve: It was renamed, because it is about confirming before submission.
... the updated text says that.

Alastair: This looks like it expands on 3.3.6 which builds on 3.3.4.
... On the assumption that at the AA level it was scoped to legally binding things, and was narrower than it could have been
... Is it because there is a new user requirement that was not previously covered, or is it narrower than the previous one?
... or is it that the technology has moved on and it is easier to meet?

Steve: This came from one of the requirements we had in the COGA task force.
... I think the reason is it was felt it wasn't good enough for the cognitive use case.
... it is the "easily go back" piece so people can more easily make changes.

John K: Yes, that was the issue - the ability to get back to a specific place in a multi step process so one could correct information.

scribe: and not get lost cognitively in a multi step process. You can see the end results, and then go back to step 5 without going through the whole process.

Alastair: That's fine. It would be useful to point to whatever theme or research COGA had. I didn't see it.
... If the justification is it is not covering the cognitive need as much as we need to, that is fine.

John K: Yes, that is my recollection.

Steve: We will go through the Gap Analysis - pretty sure it is in there.
... we are currently mapping the requirements, so this may catch that.

Jake: What did we try to solve with this success criteria? The intent gives me the idea you need to be able at any step to be able to get to the others - that is clear.
... But in the text right now, it also mentions the overview before final submission, and once you have completed steps it says some information is added automatically.
... 1 of the issues I added into the survey is that there is another one for steps - Information in Steps, which is about autopopulating information.
... That could be deleted for this so it is only about being able to go backward and forwards, but that is also part of 3.3.6 and 3.3.4, so we need a good understanding of what we are trying to solve.

Steve: The summary of all entries is covered by 3.3.6?

Alastair: Yes, except that is AAA. And the other refers to legally binding transactions.

Steve: the traversing the steps has been crossed out - Jake's proposal.
... Keep the first one in because of the level.
... The second one is the crux of the issue - you can get back to edit.
... The 3rd one is overlapping - you are correct.

Alastair: I think that is David Fazio's.

<alastairc> The overlap could be with info in steps: https://docs.google.com/document/d/18MCz5XDsMmglcAe2j-HzQbpADpw_HtdVjDZ3EHX4-xk/edit#

Steve: the point of that one is that if you enter something once, you shouldn't have to enter it in another step.

Alastair: Yes. This is progressing fairly well.
... read text from link above).

<johnkirkwood> i think there may be a bit of cross over.

Alastair: This one is about making sure you can go back. There is potential overlap, but there is a difference.

Steve: Do we want 2 SCs?

Alastair: if one gets through and the other doesn't...
... Jake's made some suggestions. I would suggest not accepting the deletion of the final bullet just yet.
... My question is whether anyone has any potential objections with the approach to the success criteria text?

Jake: If we talk about the 1st bullet: a summary of entries in all steps is presented before final submission. This can be read in 2 ways.
... in every step, you see a summary
... or a summary is presented before final submission, like an overview page.
... I know the first page, in the Intent, that it is not about the overview page, so people don't need to wait for this.
... so people do not forget the issue.
... also a summary of all entries at all steps - needs to be reworded.

Alastair: I think we need to lean on how we describe the input
... For a funding application with 6 major sections, each one of which you want to define as a process. Each has its own summary page.
... but the entire process has its own summary. If you have each entry, it would be complicated.
... Imagine it is grant funding, this is the purpose, this is what we do with the grant, here are our financials.
... Then you get a summary of what you have entered. In this case i would want each section to be defined as a process.
... If we give examples this is what is needed.

<Zakim> mbgower, you wanted to say that Rachael and I drafted something very similar for each of the bullets last time

Mike G: I hate going back to 2.1, but we had an SC that was similar to the 1st bullet of this, and one for the exception of this.

scribe: I will locate the information and forward this on to Steve?
... Yes, I will forward to Steve.
... Last time we restricted this to information that would cause legal or financial transactions, because it becomes difficult to say what is a multistep process.
... Otherwise any questionnaire, any single page with multiple fields, it can get pretty complicated.
... at one point we talked about a read only summary...There are a lot of things that need to be defined so it doesn't become harder to interact with the information.

Steve: Yes, I thought about a single page with multiple steps on it.

Mike G: Yes, so we need to nail that down.

Alastair: Yes, talking about pages won't get us any where.
... I don't know if there is a minimum number of questions.

Mike G: It was called Confirm Important Information. Another was timeouts, and was the ability to notify someone before time.

scribe: reads from the old SC.
... The one piece that is new: go back and alter data - this is an important piece to capture.

Mike G: I will send this to Steve.

Alastair: Yes, Steve, please do go through those.
... I think this is a harder topic.

<mbgower> if you put it as a /me comment, it won't be in the minutes

Alastair: The other main comment I have is tightening up the exceptions.
... So maybe we need to move this to the meeting next year.

Steve: Is there a page where it just has all the questionnaires?

<alastairc> https://www.w3.org/WAI/GL/wiki/Upcoming_agendas

Alastair: From any questionnaire it should be on the left.

Findable help https://www.w3.org/2002/09/wbs/35422/findable-help/results

<Detlev> can do it

<Detlev> scribe: Detlev

Jennie: Apologies for late additions - addressed some comments
... draft SC text, reponse to Bruce#s comments
... Bruce was talking about difficulties of PDF docs viewed in browser, single page apps, chat bots

<Jennie> On every page in a set of web pages that are actively supported (same link as in current text, needs a definition for actively supported...) that requires a bypass block under 2.4.1 or use a footer, at least one of the following is provided in a consistent location

Jennie: Its quite wordy but needs to separate out those cases where the SC applies

Steve: Is it skip links?

<alastairc> Bypass blocks: A mechanism is available to bypass blocks of content that are repeated on multiple Web pages.

Jennie: Yes - would ask for a consistent location

Alastair: So that would outscope individual pages without navigation

jennie: dificulties to define pages without / with footers

Alastair: So use Bypass blocks or use a footer

Jennie: Exceptions for smaller sites

Alastair: Difficulty to identify which sites are "actively supported"

Jennie: There is no requirement (yet) for sites to indicate whether they are actively supported

Alastair: Should be worked into the document

Jennie: Should I edit / add to document for further review?
... in the draft SC text there is a new final bullet point

<Jennie> human contact details or link to human contact details

Jennie: bullet 2 (human contact details) should be updated

Alastair: you don't need the phone number straight, you can link to a page that has it
... looking through survey - comment 4 is good, make that change

Jennie: "fully automated chatbot does not meet the requirement"

<Jennie> A fully automated interaction option which understands misspelled words, can identify when a person did not find the answer to their question within x number of interactions and then refers them to a human.

Jennie: defining interaction alternatives
... adressing COGA concerns that it would be able to address misspellings

<laura> it addresses it for me.

Raf: had concern about chatbots, happy with the current solution

Alastair: Big chunk in SC text, maybe put into a definition?

Point about automated chatbot might bette rbe outside SC text?

Jennie: Text in IRC channel should be the replacement, positive wording instead of exclusion
... need to meet all other requirements too

Alastair (putting changes in place?)

<Jennie> https://docs.google.com/document/d/1VSnTxeUK58pYdjjf-wU8HW5AxkctEn5i1_owyZf4Aoo/edit#heading=h.c05wogozf6f3

Jennie: added a couple of things based on suggestions

Removed para out of intent to Techniques section

scribe: in Failure section, added more explanation why some of the things fail
... understanding that tooltips are not discouraged, but that consistent location is required

Jenne: WCAG2ICT content was indicated as weak, therefore I need help on this

Alastair: Would not work on ICT, probable not to be included
... wondering about people who are intentionally anonymous - how would that apply to them? It is an edge case
... no contact point for legitimate reasons

Jennie: blogger who does not what to provide help option
... there should be a way to address these cases - help appreciated

Alastair: we could include in def of "actively supported" an angle on "human who can be contacted", excluding cases where people want to stay anonymous
... people have nit have a chance to consider changes - can you implement changes so we bring thos up again for a final review?

Jennie: When should it come up again for review?

Alastair: 1st or 2nd week of January

Jennie: will implement changes till the end of this week, givng people time to review

<Jennie> Scribe: Jennie

WCAG 2.2 timeline

Alastair: We are getting towards end of useful drafting stage of 2.2.
... Where our current agenda stops towards end of January, but by end of February we wouldn't add new success criteria.
... Working backwards from the publication deadline, we want to be using February to finalize the pull requests for success criteria.
... By end of January we need to agree the SC we are including in 2.2. are ready to go.
... So in the Google doc or whatever other format we are looking at the end of January, so it can be made into a pull request we think people will approve.

Non-text contrast update https://www.w3.org/2002/09/wbs/35422/ntc-updates-dec-2019/results

Alastair: There are a number of github issues.
... This is something I did a few weeks ago because I noticed a number of issues on this that could be tackled together.
... The survey result shows 3 people happy with it. 1 person who wants some changes.
... I have a lot of information in the pull request regarding changes.
... E.g. removing the explanation section, adding more examples into the table...
... and the 1 Andrew was unhappy with was the restructuring of the headings.
... I think Andrew thought they should align more closely with the SC types.
... But because this is an Understanding Document, if people are happy with the changes as presented...
... I will update the headings for the Techniques.
... 1 will be for graphic objects, the other for status indicators.
... Any questions, comments, or objections to the Understanding Document?

Detlev: The pull request makes it difficult to see the changes.

Alastair: Look at the bottom one, and click on it. Once it opens it should be clear.
... if you load the dif

Detlev: Will there be time for comment and review?

Alastair: Yes
... it has been on the agenda for a few weeks but we can put it to next week.

Detlev: Please go ahead.

Alastair: Any objections?

<Detlev> souns good to me!

RESOLUTION: Publish PR 931 as amended

<alastairc> rsaagent make minutes

<alastairc> rrsagent make minutes

<alastairc> RRSAgent make minutes

<alastairc> RRSAgent create minutes

Summary of Action Items

Summary of Resolutions

  1. Publish PR 931 as amended
[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.154 (CVS log)
$Date: 2019/12/10 18:01:57 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.154  of Date: 2018/09/25 16:35:56  
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: AlastairC, Jennie, jon_avila, JakeAbma, JustineP, MarcJohlic, janina, Raf, stevelee, Detlev, JohnRochford, Chuck, Laura, mbgower, johnkirkwood

WARNING: Replacing previous Present list. (Old list: AWK, Chuck, Raf, Fazio, janina, Rachael, Brooks, Katie_Haritos-Shea, JakeAbma, JF, johnkirkwood, MarcJohlic, mbgower, bruce_bailey, Avneesh, PeterKorn, david-macdonald-)
Use 'Present+ ... ' if you meant to add people without replacing the list,
such as: <dbooth> Present+ AlastairC, Jennie

Present: AlastairC Jennie jon_avila JakeAbma JustineP MarcJohlic janina Raf stevelee Detlev JohnRochford Chuck Laura mbgower johnkirkwood
Regrets: Bruce John DavidF
Found Scribe: JakeAbma
Found Scribe: Jennie
Found Scribe: JakeAbma
Inferring ScribeNick: JakeAbma
Found Scribe: Jennie
Inferring ScribeNick: Jennie
Found Scribe: Detlev
Inferring ScribeNick: Detlev
Found Scribe: Jennie
Inferring ScribeNick: Jennie
Scribes: JakeAbma, Jennie, Detlev
ScribeNicks: JakeAbma, Jennie, Detlev

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

Found Date: 10 Dec 2019
People with action items: 

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]