Accessibility Guidelines Working Group Teleconference

05 Nov 2019


AWK, alastairc, JakeAbma, Jared_Batterman, Jennie, Fazio, MichaelC, Laura, bruce_bailey, jon_avila, Detlev, david-macdonald, Katie_Haritos-Shea, Rachael, kirkwood, JohnRochford, Raf
JakeAbma, Laura


<JakeAbma> scribe: JakeAbma

Scribe assignments https://www.w3.org/WAI/GL/wiki/Scribe_List

AWK: you've seen my email , about scribing
... got some sign ups, please sign up more, specially in future for all attending the meeting

<laura> awk’s email: https://lists.w3.org/Archives/Public/w3c-wai-gl/2019OctDec/0073.html

AWK: AC and me will start assigning if this doesn't work

<laura> Scribing Commands and Related Info: https://www.w3.org/WAI/GL/wiki/Scribing_Commands_and_Related_Info

WCAG 2.2 Accessible authentication (1st week) https://www.w3.org/2002/09/wbs/35422/accessible-auth/results

<Jennie> John Rochford?

questions/comments on any other WCAG 2.2 criteria proposals

WCAG 2.2 Find help https://www.w3.org/2002/09/wbs/35422/findable-help/results

<bruce_bailey> http://docs.google.com/document/d/1fX4Iw169OGUny5RTd70S8qAneYy5e0hr7zupE21gPBM

JD: back and forth comments going on till 10 minutes before the meeting
... two major issues decide if it's worth talking about everything else
... the chatbot to start with

AWK: if the chatbot is sufficient?

JD: yes
... challenge is data to support challenge of chatbot
... what makes chatbot OK to pass the SC?
... problem is right wording for chatbots and use of chatbots my people with cognitive disabilities

MG: if chatbot reference is deleted from SC, but added to techniques, that might clarify to make it possible
... that is potentially a solution

AWK: in the draft it says now human contact mechanism and the fourth bullet: A fully automated chatbot does not meet the above requirements.

MG: yeh, there is already human support in SC
... validating a chatbot might be part of a technique

Raf: would also not exclude chat bots
... chatbot can also provide info for how to get help

<bruce_bailey> +1 for doc to spell out problems with Chat Bots but not exclude them

<AWK> Jake: Question is what if the information in the FAQ is also present in a chatbot, why is one ok and the other not?

JD: duplication is excellent idea, if both options are present, people can choose

<jon_avila> In my opinion chat bots can be helpful for some things -- but tend to provide specific information or generic information but miss nuanced questions and they just take me in circles and don't work well with progressive discussions

<Zakim> AWK, you wanted to say that it gets very difficult to quantify efficacy. Undoubtedly a support system manned by speakers who have a different native language could have similar

AWK: it's hard to try to quantify how effective a chatbot can be
... we don't have a test or standard for when a chatbot is fine or not

<jon_avila> FAQ pages tend to put all of the questions out there which can be helpful but can also be overwhelming with volume information and organization can be an issue.

<bruce_bailey> +1 to what AWK outlining, human operator on the other end of chat might have same difficulties as chat bot

<Jared_Batterman> True, but could you measure grade level of text and use that as somewhat of a barometer? For example, a chatbot could say "Salutations human, which services may I offer" or "Hi, how can I help you"

<Zakim> alastairc, you wanted to ask whether a chatbot would be like a human (high-ish bar), or whether it would be considered self-help?

<jon_avila> I'm also confused why a FAQ would pass based on the description of the SC

AWK: so you pass with a FAQ of 3 questions and a chatbot with 100000 will fail?!

<AWK> +1 Alastair!

AC: there's different chatbots, some with humans behind, some only rigid answers, big differences
... multiple ways seems a good option

<alastairc> AC - to fill in: If we consider chatbots a self-help option, how about specifying that it should not rely on a typing-interface?

Detlev: if all options would meet the SC, some seems easy, others more difficult, but I don't see requirements for what details are needed
... how do we nail it down?
... quality is not defined, chatbot answer in a second can be much more helpful than email after a day

JD: Alastair's idea might be a good way, workable option to continue

DmD: what about small sites, why should they add a FAQ?

JD: it's an option, only contact would be OK

<alastairc> So if you have a small site with contact info, that's your 'help'.

<Zakim> bruce_bailey, you wanted to ask about how would static document pass

BB: anxious about static documents in this SC

AWK: so this SC would then be about human contact mechanism, human details and self help
... correct Jennie?

JD: yes

<bruce_bailey> +1 to AWK point about validating (human) email is tricky

AWK: difficult to define if the response is human as a tester

<AWK> Set of web pages: collection of web pages that share a common purpose and that are created by the same author, group or organization

JD: next discussion is about definition of "set of web pages"

<Detlev> Seems like a rather rare case for docs?

AC: set of web pages, most documents don't qualify for that

<AWK> From WCAG2ICT: This applies directly as written and described in Intent from Understanding Success Criterion 2.4.5 (also provided below), replacing “Web page(s)” with “non-web document(s)” and “software program(s)”.

BB: the set of web pages thing, before the WCAG2ICT, we said we don't do it for off line documents, I think it's a bad approach for WCAG 2.2...
... difficult to make other than triple A (AAA)

DmD: had big discussion with 100 pages in PDF, is it one or 100, we judged 1 page, because 1 URL

<jon_avila> excluding non-markup langues Could exclude mobile apps

<bruce_bailey> could also have a collection of 100+ pdf documents (with each pdf being 100+ pages)

AWK: maybe we need to add service, or type of web page, like we did with 'legal, etc.'

<bruce_bailey> +1 to pages within a process

<AWK> AWK: offers comment about scoping the SC to web pages that offer services, or sell things, or provide education resources, etc... ?

<Jennie> +1

<jon_avila> blog comments have forms -- but those might not need human help.

<david-macdonald> if it applies to set of pages, then 1 pdf form would not qualify

<bruce_bailey> 2.4.5 might be a good scoping model

<bruce_bailey> 2.4.5 Multiple Ways: More than one way is available to locate a Web page within a set of Web pages except where the Web Page is the result of, or a step in, a process. (Level AA)

<alastairc> you mean sites which have navigation?

Rachael: struggling with if this is so hard, seems very achievable to me

AWK: it seems achievable, support pages or human contact

<Rachael> Also, like scoping around input/interaction but concern if that would exclude any safety/medical information

<Detlev> (maybe someone can convert call-in User_4 to Rachael?)

AWK: it's do-able but the question is if people will do it

<bruce_bailey> Ex: 2.4.x Findable Help: More than one way is available to find help for a Web page within a set of Web pages when the Web Page is the result of, or a step in, a process.

AC: can we make the scope to processes?
... wouldn't want it to scope to processes, but can we add something about where we make it consitent

<alastairc> thinking of scope: would rather not tie it to processes, but to be available when there is navigation available.

<alastairc> Was intending to mean like 3.2.3 Consistent Navigation

AWK: additional question, when help is in a dialog, the dialog is not the page, it's on a page

<alastairc> Suggestion: "On each page in a set of web pages that includes a navigation mechanism, at least one of the following is provided:"

<alastairc> Jake: Based on multiple implementation scenarios, seemed like the intent was that if you are at step 3 of 5, want instant contact on the page.

<alastairc> Jake: Seems possible if your steps are in the page, but if the whole flow is in a dialogue, and with a window over things, wasn't sure if that was matching the need/intent. Was that thought about? Creates extra issues, as Jennie mentioned.

<alastairc> Jake: If this scenario was ok (not having to provide in a dialogue) then need an update in normative text (or possibly understanding)

+1 to Detlev, in page help should also be fine instead of having it at the help page

<Fazio> +1

Detlev: in-page help should also be fine if 'as complete and the right quality'

<alastairc> how the help is best provided is very context dependant, have to leave that to the site.

<laura> Scribe: Laura

AC: inclined to take out consistent location.
... then provide good examples.

<AWK> Jake: more I think about this SC the larger gap I see between what the intent and the solutions we will get from the SC text

<AWK> ... help is about the quality

<AWK> ... if you have an email forum or telephone number it doesn't say anything about the actual help you get

jake: big gap between what we want and sc text.

awk: we want ot increase the chance of success. People can still game the system.

JD: subgroup wanted to meet the intent and testability.
... group worked though some the scenarios.
... locations was one.

jake: if it is about location we have another SC will cover it,

Rachael: we had looked at that SC but the needs are different.

<Fazio> If its always in the navigation bar, wouldn't that solve it?

Rachael: location on the screen i is what we were conserned about.

Jennie: yes. In the same place.

awk: are we getting closer?
... alot of questions.

detev: include inline help.
... email would be easy to meet.

Jake: indicate maybe 2 or 3 kinds of help.

<alastairc> Quick suggestions to tackle the scope & chatbot bits: https://docs.google.com/document/d/1NcfHfwH_HcuU42juaEWQj3FrE6sqBC3nxIjqJDkycn4/edit?usp=sharing

Jake: help is a broad topic.

<Fazio> I agree "help" needs to be clearly defined

David: director asked us at TPAC to incorporate more coga.

<Ryladog> +1 to David

David: think we should try to get it in.

Laura: +1 to david.

<Rachael> +1 david

<bruce_bailey> +1 David M

<alastairc> that's fine, we can slot things in later.

Jennie: could come back Nov 26.

awk: not a problem.

Jennie: the following week coiuld work.

WCAG 2.2 Accessible authentication (1st week)


awk: reminder look at the scribe list.

ac: separate understanding doc.
... had a fair bit or work in 2.1

<bruce_bailey> link from survey: http://docs.google.com/document/d/1J3NFw6NPyj7QGddBtRmagrtS-x4t9BWan8PYEdSpMZM

ac: taclkes: passwords, memeorizing, transcriving
... don’t rely on one method.
... simple way is to use an email option.
... web auth is another way.
... worked thorugh a lot of examples.

Detlev: 2 biometic methods would not be sufficient?

AC: yes.

<bruce_bailey> Here is how we phrase the biometric requirement in 508:

<bruce_bailey> 403 Biometrics

<bruce_bailey> 403.1 General. Where provided, biometrics shall not be the only means for user identification or control.

<bruce_bailey> EXCEPTION: Where at least two biometric options that use different biological characteristics are provided, ICT shall be permitted to use biometrics as the only means for user identification or control.

jake: only speific email solution. still need to copy and paste.
... email could be a technique but not a solution.

ac: did discuss in small groupls email providers.

<jon_avila> I believe Slack is an example of where you can login with an email confirmation without entering credentials.

ac: second factor. Thre are solutions.

jake: but they use the same modalities.

awk: from what perspective?

Jake: coga
... you have to remember things.
... need to provide different modalities

<JustineP> +1 to defining cognitive function

ac: we allowed for certain types of rememberting.
... if offloading to a different provider? yes.
... we are not banning passwords.
... putting in a password is not a cognative function.

David: coga function is a wide concept.
... didn’t we used to say use of memory.

ac: it is explained in the document.

david: we could make our own definition.
... trying to avoid 1 biometic. 2 or more may be ok.

need to define cognative fuction in more detail

<bruce_bailey> +1 to have multiple cognitive options as okay (memorize vs calculation)

AC: we can make it more specific
... avoid modality.

<alastairc> Cognitive function was previously defined as: memorizing character strings (except for your own name, email address, or identification number);

<alastairc> using correct spelling;

<alastairc> performing calculations or puzzles;

<alastairc> copying characters into an input;

ac: on the biometics not sure we want to use the same apoach.

Bruce: provided 508 text.

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

JR: Understanding doc addresses issues.
... People with physical disabilities (i.e., those with missing or occluded body parts such as eyes or fingertips) may not be able to authenticate via biometrics. Providing an authentication method that does not rely exclusively on biometrics is essential.

<bruce_bailey> for 508, we thought it was too much to say that an agency could never use biometrics

ac: define cognative bits.

JR: I removed them but I could put them back.

<Jared_Batterman> The prevalence of dual identification also complicates this a bit.

ac: perhaps have cognative function definition

<bruce_bailey> basic security is (1) what you know, (2) what you have, (3) what you are: pick two

awk: for biometics is it different?

<bruce_bailey> if we don't permit biometrics, what else is there for (3) what you are?

ac: need to use something different anyway.
... offload to the users device.

DF: important to define cognative fuction.
... 2 types of memory: recall and recognition

JR: added a definition.

<Zakim> bruce_bailey, you wanted to say i think switching modalities is pretty hight bar

bruce: reading his IRC comment.
... thinks it is a no go with the security community.

ac: they are off loading it.

<alastairc> https://webauthn.io/

ac: WebAuthn

<alastairc> Windows hello implementation: https://support.microsoft.com/en-gb/help/4028017/windows-learn-about-windows-hello-and-set-it-up

awk: are we okay where this is generally going?
... any issues with the SC text?

jake: 1 thing to clarify: multistep authication.
... using more than 2 modalities
... and thier reuse.
... not sure we can exclude.

JR: how yould you phrase that.

jake: not sure.

JR: I will wotk on it.

AC: having 2 steps.
... might work.
... need to insure we have examples of 2 step authentication.

<Detlev> Jake's at a bank, right?

<Detlev> I have to drop off

JR: include in the techiques?

AC: need. examples

JR: we do have some in the doc in the deprecated section.

awk: this one is viable.

AC: will go through what needs to be done.

<AWK> trackbot, end meeting

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.154 (CVS log)
$Date: 2019/11/05 18:04:00 $

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)

Succeeded: s/human operations/human operator/
Succeeded: s/100+ pdf document/100+ pdf documents/
Succeeded: s/Iinclined/inclined/
Succeeded: s/thoughsome/though some/
Succeeded: s/sinearios/scenarios/
Succeeded: s/lalot of cquestions./alot of questions./
Succeeded: s/bak Nov / back Nov /
Succeeded: s/suficieant/sufficient/
Succeeded: s/perpective/perspective/
Succeeded: s/membory/memory/
Succeeded: s/1 or more may/2 or more may/
Succeeded: s/ntte/need/
Succeeded: s/exampeles/examples/
Default Present: AWK, alastairc, JakeAbma, Jared_Batterman, Jennie, Fazio, MichaelC, Laura, bruce_bailey, jon_avila, Detlev, david-macdonald, Katie_Haritos-Shea, Rachael, kirkwood, JohnRochford, Raf

WARNING: Replacing previous Present list. (Old list: AWK, Jared_Batterman, MichaelC, Raf, Jennie, bruce_bailey, Wilco, Rachael, jeanne, mbgower, alastairc, maryjom, david-macdonald, JF, Katie_Haritos-Shea, Laura)
Use 'Present+ ... ' if you meant to add people without replacing the list,
such as: <dbooth> Present+ AWK

Present: AWK alastairc JakeAbma Jared_Batterman Jennie Fazio MichaelC Laura bruce_bailey jon_avila Detlev david-macdonald Katie_Haritos-Shea Rachael kirkwood JohnRochford Raf
Found Scribe: JakeAbma
Inferring ScribeNick: JakeAbma
Found Scribe: Laura
Inferring ScribeNick: laura
Scribes: JakeAbma, Laura
ScribeNicks: JakeAbma, laura
Found Date: 05 Nov 2019
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]