<AWK> +AWK
<alastairc> https://www.w3.org/WAI/GL/wiki/Scribe_List#2020_Scribe_History
<alastairc> scribe:kirkwood
<shanew> Present
AC: going through agenda, discussed CSUN last week. March 10 I will be around. will still meet around editing 2.2 criteria but not full decision making
<Jennie> Location at CSUN on March 10th?
AC: JD: call at same time but at same location?
... could be 8-10. hadn’t planned to have a full AG meeting.
AC full reviews infroduction
Wilco: what we have done lately
<maryjom> https://www.w3.org/2002/09/wbs/93339/PublishACTRules/
Wilco: AC task force setup a process for consuming and going through rules specifically ACT Commiunity group.
... have gone through 15ish
... have a second batch ready for publication
... have a survey
... TF seemas ready to go, community group ready to go, small things but ready for publication
... ready for ag to say yay or neah
MM: gone through multiple rounds of review in ADCT community group, any comments
... survey is open now until coming tuesday
... open for responses. goal is to get published before CSUN to show them
AC: that is on agenda for next week
<Zakim> JF, you wanted to ask about "You are allowed to read the questionnaire, but you are NOT authorized to answer with your current credentials!"
JF: when going to survey not able to answer with current credentials
MaryJo: i just updated try again.
<laura> I’m getting: You are NOT allowed to see this questionnaire.
<alastairc> https://act-rules.github.io/rules/59796f
Wilco: there are three rules about language, duplicate id's
AC: any questions?
<laura> I can get to it now. Thanks.
AC: people can get it good
AC: we have results of Silver editors draft update
... can’t get to link to draft
... its basically 6 yes and 6 no
... Sommers want to include updates for visual contrast
... Wilco feedback references previous about pars being testabble
... looking at feedback
<bruce_bailey> Andy shared his edits at Silver meeting this morning
AC: some does have bruce edits proposed discussed with Jeanne to incorporate them
<bruce_bailey> Current version:
<bruce_bailey> https://raw.githack.com/Myndex/silver/patch-1/guidelines/methods/Method-font-characteristic-contrast.html
AC: Jeanne or Shawn notice any categories of feedback?
Jeanne: i tried to capture in introduction of survey
... wanted to address Lauras comments in misunderstanding of discussion of scope want to clarify, in a conformance claim to delclari a scope discussing a subsedtion to acknowledge doing today without gaming system.
... don’t wnat to make a claim beyond scope of what they are claimin
... first paragraph is in scope last paragraph out of scope. wnat to avoid doing a section of product or application. thats what we wanted to address with claiming scope
Shawn: challeng is complex mulitpage web apps, muliple scopes in same url
AC: probably worth discussion normnitive parts
... question wether guideines themselves and testable statements. whether normnitive
... don’t know wether testable parts need to be normnitive or not
... maybe view of document with test methods part of it, was exploring
... as we put for wider review we want to get good feedback
... anyone else have comments about normnitive aspect?
... Jeanne when looking through actual spec, is getting started normnitive or others
<Zakim> jeanne, you wanted to say the research I did about it
Jeanne: my attempt to clearly label normnitive not. part in actula guidlenies normnitive. green box to indicate start of informitive. all tabs normnitive.
... green is all informitive
... asked if w3c requires testable statements
... spoke to editor of process document. asked about testabel statments and he thought did not
AC: that explains
SL: have tech specific documentation available
<Zakim> Wilco, you wanted to explain my position about normative tests
Wilco: the reson I discagree with approach has to do with stabliitiy of conformance model. need to know if you pass some test.
... requirments need to be stable. normnitive addresses. informative you lose all of that
... ACT works well as project could be built on top of these. but take out testable then can’t really build on top of that
... takes whole conformance model to be meaingful
SL: that is by design in order to say this is reocmmend. not getting rid of testable statements. can build on work ACT to provide framework. and be valid. requirmihng transparency on how meeting these things. transparency on how to meet guidleine, even through strange
... still cover tests can update with browser support changes
JF: i have to push back on that
JF: measurable, testable, repeatable for
... US gov’t for example
... don’t know how to take some of these as norminitve statements
... need to take a mechanism to create somehting like customers want like a dashboard to actual meashurements
... when comes time to measure like the words sufficient don’t know how that can work
<Fazio> +1 good job John
SL: good exmaple of speed limit. The speed limit changes according to context. because of that gives you top level guidance depends on context of text itself
JF: anolgy to speed limit in towns/context continures
... to give percentile most of time its good, how do we roll up into a dashboard score. granuallar but devil is in details
SL: haven’t built up rubric of guidance for this one yet
<alastairc> 1+ MichaelG
<Zakim> AWK, you wanted to say that the WG may still require testable statements, even if process document doesn't. What I'm hearing here calls for clarity in the conformance model.
AK: wanted to respond related to process
... the Working group may. need for clarity in conformance model
... concern that i have amounts to where we start having deharmonization
... others may be looking for yes or no answer
<JF> +1 to AWK
<Wilco> +1
AK: I worry if we are not specific enough we will encourage development of additional requirments
<Zakim> jeanne, you wanted to add that the needs of some disabilities require more flexible testing
Shawn: plus one to theat
<Chuck> +1
Jeanne: what is in the update we got this morning is the failure levels for eadh one of the contrast in the table this morning
... working on it
... when tests are not normnitive flexibility to include more needs of peple with diabilites, including COG and low vision.
<Lauriat> +1, thank you for bringing that up, Jeanne!
Jeanne: specifice needs to open up to more disabilites COGA and low vision
<GN015> when not having testable statements, how is compliance determined?
Jeanne: what we do in WCAG today very strict, other areas not so strict in us government.
<CharlesHall> i think the goal is still testable. the flexibility is that there are many kinds of tests.
<Zakim> alastairc, you wanted to say we will always be drawing a line somewhere between top-level and detail for what is normative, but does it matter whether the detail is normative?
<Lauriat> +1 to Charles
Jeanne: i think we can do this if we move qawy very strict requirments for testable
AC: always drawing a line somewhere.
... creating testable statements of SC all those requirments. put testing statemnt under eveluate tab. could we make one of those tabs?
... possible to create a spec high lievel guideline with normnitive included underneath
<Zakim> Wilco, you wanted to bring up process of normative
AC: we need to bridge the gap somehow
Wilco: concerned about is that it would lower bar of equality standard
... the direction this would take almost becomes a living standard
... without the effort and quality assurance iwould give us. seems like bypassing w3c altogeither and creating a moving target
Wico: if you look at techniques they get outdated and worried qabout keeping up with technolgy
... I’d like to know how we would do that with methodws
... if you build test procedures on top of technology agnostic testable statement
KH: think we do have an answer here
... mke the how’s the testable statement, those could be added to over time
<JF> how to do, or how to test?
KH: makes for complication, but allows for great level of points ot come up with score
... ho back to each of the how’s and make those the testable statement seems possibile to scale over time
MG: I want to go back to speeding of John. in one minute you fail.
... need to be able to measure relative level, as opposed to meeting speed limit at all times
... we have one technical way, but other ways are what is happening with Silver
... we are removing all guidance while removing dev and tech techniques, except for design, very difficult to capture methods for design. i think doing a good job at merging
AC: so not having testable statement doesn’t bother you
CA: JF opened up iwith visual contrast. it is fairly binary but font size and weight that mkes it now more difficult
<Ryladog> There is also ambient light though
CA: but still boils down to a true false testable approach
<Lauriat> and also "readable, but still difficult"
AC: even though more of a formula now, guideline is to fasst
<JF> How do you measure "too fast"?
AC: i can see if there are problems there are areas becomes too specific
MM: one other complication, when you consider eventually becomes a stndard it goes into other languages. how much of Silver strandard is planned to be translated?
AC: where do we draw line between normnitive an not
JF: back to seed limit
... can push it in certain areas… but when go past school show relativity should come in
... if goal is a score not undersanding how i get a percentile
... some places things are critical than others, how do we account for variables in scoring?
AC: to topic, things to discuss with score
JF: we need a unit of measurement
<Zakim> AWK, you wanted to talk about speed limits and malleable evaluation
AC: agreed but where we draw line normnitive and supportive is a differnt question
<jeanne2> Jeanne wants to respond to John that it is becoming clear that the unit will be by guideline, not by an overall conformance.
<Chuck> SAPC
Andrew: i’m saying opposit of Jeanne, the law does define speed limit. conformance is evaluatin is wether have mechanism to see if contrast enough
... alorythm turns into a testable thing. if different way to make work we need to eveluate that
<alastairc> Looking at the detail, surely there is something above this level that is normative? https://raw.githack.com/w3c/silver/ED-draft=comments-changes-js/guidelines/methods/Method-font-characteristic-contrast.html
Andrew: there is a measure to be set. alternative text for all eimages.
PK: like speed limit anology
... i like th model is goal not endangering other drivers. to have actual guidance rather than leaing up ot predjudice enforcment officer
KH: we are ging to have multiple points of testing. depends on weather. need to take common scenerios. taking different context create alogorythm complicated but think we can do
AC: needs to be normnitive
<PeterKorn> +1 Ryladog. Actual accessibility is different from a fixed requirement that doesn't involve context.
KJ: needs to be normnitive
<sajkaj> Wondering whether 5G technology deployment will eventually give us variable speed limits to account for current road conditions
KH: speeding on snow day, not going to be impossible but need to add on to them
... changes alogorythm but getting covered, context is almost always relative
AC: not convinced everything like that is normnitive
<Zakim> Wilco, you wanted to get on the "speed limit" bandwagon
Wilco: not really about more or less testable, process to get there, for informitive its AG can change notes
... that means speed limit today can change in the future, which seems unfair
<alastairc> Um, that's a very cautious process, as you pointed out about the slowness.
Wilco: that is why normnitive is very important won’t pull rug out from under you
Wilcom: a standard to sontinue to rely on
AC: do have proces for informitive materials,
Wilco: thats just us
<Zakim> jeanne, you wanted to say that the law is to obey the speed limit, but the actual speed limit isn't set by law, it's flexible.
Jeanne: i agree with Katy context is very important which is what we are trying to capture. interested in trying to get into the how. the flexibility of not being locked into testable is very importnat, in partiucular addressing under servied goups of people with disabilities
<mbgower> https://www.w3.org/WAI/WCAG21/Understanding/headings-and-labels.html
Jeanne: one of those meeting in AC meeting Boston, if we tie down too tightly to accessibile testing stqatements, make sless useful
... examples in clear language
... working for testing company have to have the flexibility, if we have very strict rules in professional testing company, there are kids left out
... important to serve those who needed if strict normnitive test leave people behind
... want to do it but don’t want to lock people out
JD: work for a gov’t state agency,
... wilco about officially recognized.
... we have built a lot of work on evaluting according to conformance
... how do we show compliance and eveluate how applies
<JF> +1!!!
JD: challenge when not normnitive when 30,000 state employees meet requirments
<GN015> +1!
<jeanne2> Jeanne: Jennie, we are including tests. We are just not requiring that the tests are normative.
<Zakim> mbgower, you wanted to say that several existing SCs lack normative language in whole or in part
JD: struggle with contextual piece but applaud the ieffort
<mbgower> 1.1.1 "text alternative that serves the equivalent purpose"; 2.4.2 "titles that describe topic or purpose"; 2.4.3 "preserves meaning"; 2.4.4 "the purpose of each link can be determined"; 2.4.6 "headings and labels describe topic or purpose";
MG: some of existing SC have value yet you cant really creat e test
<alastairc> scribe:Wilco
<kirkwood> when!
<kirkwood> ;
<Ryladog> That is true Mike....not measurable with out a human
<AWK> WCAG never said that SC had to be testable without a human
david: One of the test for 2.0 was "does it have a high inter-rater reliability". Would most experts agree
... at one point it was 8 out of 10, but we went with "high" later.
Detlev: I struggle that the guidelines are normative. I don't see how that's possible. Calling it normative implies that it can be measured in some way.
... I think normative is at a lower level, and it has to be there
... I think there are many measurable things that should be kept
... The problem is that for many SCs it's not the value you need to make the assessment. For example bypass block, there are different way to meet that.
... Where do you draw the line? At some point you have to make a call. This won't go away with atomic statement. You have to find judgements.
<Ryladog> +! to David - the interator testable for those things that are not machine testable
Detlev: You can get half-way values between fail and pass, that you can add up and is somewhat closer to the truth
JF: I agree, there is subjectivity to this. An expert has deemed that at some point the limit is interpreted by the expert
... For the speed limit, if I go 3 miles or 12 miles, there is a measurable difference. We should be able to report on that difference.
<Fazio> Usability testing solves this
<Chuck> something actionable.
<Fazio> create a set requirement
<Fazio> it's also referred to in "understanding WCAG"
Alastair: I think it would be really useful. What we don't have is the equivalent to a testable statement in Silver. If you cary on, the question is if that detailed level has to be normative or not
<Fazio> 3rd party watchdogs do this sort of thing all the time
Alastair: But where do we pull that next level from. Katie suggested the "how to" part in the tabs. That might be a useful direction to look at
... Is this something people consider critical?
... What we are talking about is recognising that this is needed, but we're not sure where it is going to go yet
... I think we are agreed there is something needed, but we have to draw a line somewhere
... Can we put an editorial note to say this is something we are working on.
Fazio: Usability testing is very different from accessibility testing. You get average users with no particular expertise, have them go through a workflow and see how usable this is.
... This is mentioned in the understanding documents. Can we set some sort of standard for getting some amount of PoW for testing.
... Third party watchdogs do this all the time; anonymous checking of types of "climates" of different kinds of industries.
<CharlesHall> it has always been the goal of silver to include and advocate for additional forms of tests, including usability methods – particularly on critical flows
Fazio: I think it would address the issues of normative statements.
Alastair: We've had long discussions on this before. It is a very difficult requirement to put in, because you are immediately adding fairly substantial costs.
... I see a place for that, but it would be in a process rather than a specific guideline
Fazio: Isn't that the point of Silver?
Alastair: yes, but we have a lot of work to do before we can work out what that might look like at that higher level
... We've not got agreement to publish. I can see immediate next steps for incorporating feedback.
<AWK> I think that the amount of discussion here underscores the need for more of the main WG's time on this.
<jeanne2> Scoring Example https://docs.google.com/document/d/1LfzTd_8WgTi0IUOOjUCRfRQ7e7__FRcnZow4w7zLlkY/edit#heading=h.qenaldiie00y
Alastair: I think it is worthwhile to show the examples, say it is being worked on.
<jeanne2> https://docs.google.com/document/d/1LfzTd_8WgTi0IUOOjUCRfRQ7e7__FRcnZow4w7zLlkY/
Alastair: When that has been worked on I think this will show how the process works
Jeanne: I don't think we have consensus on how much needs to be normative yet.
<JF> Jeanne and I agree!
Jeanne: This is why I think it is so important to get FPWD feedback. People have different perspective, I respect that. I would like to see us go forward.
... I am very willing to make changes that people thing we need to go forward. Questions are good at this point, help us go in the right direction.
<CharlesHall> apologies / regrets, I have to drop off the call.
Alastair: We need to make some sort of change to normative before this goes out. We can include something that says what the new approach is to what is normative and what the plan is, or we can pull something in to flesh out what is the normative aspect of the spec
... The other thing is to have the example, possibly in the explainer. That could be another requirement.
... There is some overlap between how-to and understanding. It was difficult going through that to read lots of things in the guideline, and similar thing in the methods
Jeanne: that was on my list to do. We want to bring that in alignment
Alastair: My editorial comments are in the document
AWK: I have some questions about how this document is positioned. We need to clarify what we are doing there
... this is a rec track document.
Jeanne: Waiting for Michael before making those changes
<Ryladog> +1 to AWK
Jeanne: Given conversations on other documents I thought it was better to leave it in the community format until everything is worked out
<alastairc> Wilco: Seems that we're so divided on a number of things, is it ready for a wider review?
<alastairc> ... utilise the group feedback
AWK: I'm inclined to agree. The reason we don't have consensus is because much of the WG has not been focused on this.
... that is unfortunate, but it is the reality we have with doing 2.2 as well.
... I think if we're honest with ourselves we could make a greater effort to have greater consensus within the WG first.
... But this is one of several that probably needs to happen.
<Ryladog> Can we make the HOWs into testable statements?
Alastair: We've got the normative aspect to work on. An example would help on how to score things. And there are quite a few editorial things as well.
... We'll have to leave it there for today.
<PeterKorn> I have to drop.
<alastairc> Doc: https://docs.google.com/document/d/1pcg6ixAfuwlo6jb2tkZBGTDhF0fAiO49h21E6HCbQ6I/edit
Steve: One question is, is the submission bit important, given this is a multi-step process
... that was felt to be a line in the sand, and being able to go back is a bigger process. But a multi-step process doesn't cover submission.
... I was hoping we could drop that.
Alastair: The step is the focus of that. That sort of thing comes down to what is the actual focus of it.
<GN015> Isn't it covered by error prevention?
Steve: The original idea came from allowing people to create errors, as opposed to something about submitting
... What the users are supposed to do is correct errors before they submit. I don't think we need to mention that final submit
Rachael: I agree the submit portion is a little confusing. One of the questions is are you going to look at the data or allow edit of it
... I think a note rather than an exception
Steve: related to that, we have currently two bullets. I wonder if we could change it to "each previous step can be accessed unless it becomes unavailble for ... reasons."
... Would that cover also if you are submitting each step to the backend?
... I suggested changing completed to visited
Alastair: The previously visited step may be better in that it is more flexible that way. Not sure if we'll get comments around if previously visited allows you to go forward.
Steve: I was thinking going backward and forwards, accessing any step
JK: I think if the language was "visited" rather than completed, that's worth debating. Not sure how to address that
Alastair: Completed is almost like a get-out. There are some applications where you don't have to complete a step before going to the next one
Steve: should we say "each reviewed step"
JK: no I don't think that's right
Alastair: This SC builds up on error prevention. This SC is about letting people navigate through a process before a confirmation
... The bit we're struggling with is how to define steps in a process without triggering logic issues
... which is what we mean by the exceptions.
Gundula: I compare it with error prevention and feel it has overlap. Multi-step process is just another way. Error prevention includes multi-step process.
Steve: The issue for people with cognative disabilities is if on step 2 they realise they made an error in step 1, they need to correct it immediately.
<Fazio> No
Gundula: Could this be covered by extending the error prevention SC
no
Alastair: the decision for backward compatibility is to add success criteria.
JK: I think this can now be called "reviewing steps"
... If content has become invalidated, if possible the user should be warned.
<Fazio> What do you mean by revisiting steps?
<alastairc> Each previously visited step can be accessed unless it becomes unavailable for logic, security or privacy reasons;
Steve: I was thinking to drop "privacy" as it is part of security
Alastair: I'm wondering if the second bullet is needed with the new first bullet
... I think it is covering something slightly different, covering a warning
Steve: People will still get an error at the end with the existing SC, but it's better to let them know at the point of the change
... It is important to give a warning immediately, rather than waiting to submit.
JK: The accessibility reason would be there is a lack of understanding of a change of information. If you're not giving a warning that is an accessibility issue
... The warning makes it usable for a person with a cognative disability
Alastair: with the change in the first bullet, I don't think the exception is needed
... the logic one is the most important one. We need to explain that carefully
AWK: Why get rid of legal?
Alastair: It was about regressing steps; Anyone could say they have a technical reason. The legal reason, I couldn't think of why there would be a law that stops you changing a previous answer
AWK: me neither, but
... ok
Gundula: When answering my question you stated it needed review before submission. This could go into a note
... You could note it in error prevention as well
... it requires navigation between steps, and reviewing intermediate steps
Alastair: It's trying to require navigation between steps, and asking for a warning in that situation.
Steve: Maybe something about the definition or process?
Alastair: Patrick's comment was about the overlap. I think that has been dealt with
Steve: registration can be multi-step
<alastairc> Wilco: There are situations where it doens't matter if you made a mistake,
<alastairc> ... would be outside the process
Steve: doesn't process define steps?
Alastair: We'll pick this up next week. It is worth people rereviewing