Silver Task Force & Community Group

21 Jan 2022


janina, jeanne, Lauriat, shadi, Wilco, jenniferS_, SuzanneTaylor, MichaelC, JF, ToddL, sarahhorton, Rachael, Makoto, kirkwood, Azlan, Chuck_


<scribe> scribe: Chuck

Organization of work

<Lauriat> Newly organized wiki link https://github.com/w3c/silver/wiki

Rachael: Talking last month or two about how we work better as a group. Seeing what we are doing and sub-group decisions are moving along at same pace.
... We have restructured wiki, and I'll walk through it and talk about sub-groups.
... We want to transition more officially next week.
... There's a couple of docs. We have a wcag 3.0 process in 2 forms. Comparable.
... Proposes labelling and how we are going to make editors draft help public review.
... And how things move through, to encourage work.
... We have a diagram that is a visual workflow. Not the place to go to for screen readers, but we have both options.
... We have a schedule that has everything we have to discuss. We have grouped into topics that we need to address before we call wcag 3 done. Milestones.
... A grouping of content that needs to be addressed. Mapped against quarters.
... We have an editors draft link, latest working draft. And github repository.
... I am in github, this content is in github. This wiki is organized same way as everything else is organized.
... Foundational work that we have to get on board before we discuss scoring, content weighting, and actual guidelines themselves as we migrate and add to.
... All content is divided into these 4 items.
... At bottom of wiki is list of same things that are on the schedule.
... We are still finishing, we will have all of these linked. Each links to issues that need to be closed to address topic.
... Using protocols as example, brings me to the page that pulls everything together for protocols.
... Contains basic info, current work, minutes, issues...
... Any referenced research. Please put in your research.
... And a list of decisions. As groups work through, we need them to capture key decisions, so we have a record. Ideally we'll link to those.
... And document history. If the group is informal, there's a participants list. If more formal, like conformance, there will be a link to the sub-group.
... These sub-pages (should be one for every milestone/topic), in addition in github we have organized into projects.
... We are working in foundational work now.
... Every one of those milestones has a card that we can move and track our progress.
... Requirements has moved to group discussions for examples.
... When we get to point of generating content, we can move to later stage in process.
... Requirements will move into it's own sub-topic. This lets us organize and track the work and see where things are over time.

<Zakim> jeanne, you wanted to show the template for the milestone topic wiki page

Jeanne: I want to show people the wiki template page. On Rachael's screen in right hand page, you can look for template for topic pages. If you click on that...
... This is what you can copy to create your new page if you need one.
... It's ok to ask me or Rachael to do it, but if you want to do it yourself, you need to click edit button at top, next to new page. Click and you'll see the code.
... "media wiki" format. You will need to update yourself or it will look odd.

<Lauriat> qv?

Jeanne: Hopefully it's as easy as possible for people to do their own. We could use a mark down version if anybody has expertise.

mc: We can do it easily, would be manual, we can translate.

Jeanne: I don't want to have to learn mark down.

jennifers: Lot's of work here, not a small task. Was there any consideration for our blind participants and how they might use this?
... It's been challenging for my partner and I to collaborate. Any AT testing?

Rachael: Feedback is that wiki was most accessible across the spectrum. We have left it in the structure where people can use their tools of choice.
... Reading portion is good. jeanne and I will help with coding if needed.

jennifers: When my colleague tried to read with a screen reader, it read github page as table, and was confusing.
... Understand the challenge with the tools. Want to get on record that we are similar with rest of world. Technology avialable may have challenges.

Rachael: If you run into issues, we will work with them. We can't change the project portion, that's why it's list based.

ToddL: If we are looking for markdown version, I would be happy to help out and get something up as far as a markdown page.

<Lauriat> Thank you, Todd!

Jeanne: Thank you Todd!

wilco: Github is fairly open to feedback with concerns.
... I'm looking for right place to put that.

<jenniferS_> I appreciate that Wilco, but that has not been my experience.

<Wilco> https://support.github.com/contact/feedback

Rachael: We are finishing up, next week we would love it if groups would move over. Big change is for automatic minutes. Groups working on milestones should use irc channel that links to milestone.
... It will happen automatically. Format for that is pound 'wcag3-' followed by the topic. such as #wcag3-protocols.

mc: I think that we can have minute script pull topic headings. We can use internet search from there. No solution for coolating topics. HOpefully topic headings will help.

Jeanne: MC and I discussed having groups that handle multiple topics keep current minutes in irc channel and we'll generate a minutes page that will link from milestone wiki page.
... And people can see topics.

mc: Rachael gave guidance on keeping names. Doesn't matter, will happen easily enough.

Jeanne: I think we need to keep separate so we can link by topic.

<Wilco> This is another place to go: https://github.com/github/feedback/discussions?discussions_q=accessibility

mc: Protocols from changing from silver to wcag3, we'll use both.

Rachael: We'll move everyone over to 'wcag3-' topic.
... We are still working this out, so no guidance yet. But start thinking about it.
... Any q?

Shawn: Any other wrap up comments?

Integrating ATAG and UAAG into WCAG 3 (continuing discussion)

<Lauriat> https://docs.google.com/document/d/1MCQG_MIaEyC3nfwkWGoDWS6XQzMykXeC7qOOjj8AG8M/edit

Shawn: Continuation of discussion over last few weeks.
... This is a doc where we are exploring topic, how we integrate atag and uag into wcag3.
... Today we can explore core questions in more detail.
... For example, first question... <pasting>

<Lauriat> How to make it clear (Issue 503) that when a barrier needs to be addressed to clearly identify if it is user agent or content author responsibility?

Shawn: I add authoring tool as part.
... My thoughts were that rather than wcag3 being the one to say "this is who is responsible", maybe wcag3 could offer guidance to applicable levels to where barrier could be addressed.
... We would offer guidance to all places for what they can do to address barriers, rather than assigning a single owner.
... An example, language of page. We have guidance for wcag 2.x on how content author can do this, an authoring tool can prompt the author...
... And user agent could detect language and adjust expose language to AT.
... We could give this guidance without assigning responsibility to a specific level. Each level could then address.

<MichaelC> +1, this is exactly the direction I´ve been thinking

Shawn: What are the thoughts to that approach?
... Designating responsibility has been a theme, and this is my idea on how we can handle it. By not handling it.

ac mich

mc: I agree. For giving pieces of guidance... if you are a developer... if you are a designer, ... if you are developing authoring tool... if you are developing at...
... Create a stack where we try to get things addressed in multiple starts of stack.

Shawn: big +1

Janina: strongly in favor. and want it to be blameless.
... We will realize that sometimes the source could be an opensource library that stopped being developed.
... Could bring in issues (been consulting with interested parties).
... May be good decisions at the time that have become poor decisions as time progressed.

Shawn: Indeed.

Wilco: I'm cautiously positive. There may be some drawbacks.
... It's going to lead to more convoluted requirements. Unlike current scenario.
... if "at needs to do this", may remove some requirements from developers.
... May lead to further confusion and oddities.
... May get worse by doing this.
... By not clearly identifying responsibility, maybe nobody feels responsible.

Shawn: I +1 those two potential draw backs, we should keep an eye on and explore.
... You mentioned wcag 2.x about adding attribute to element, requirement is each element's language can be determined.
... Guidance is the same, there will be different methods for the layers.
... I agree with don't make more convoluted and abstract. may need to work through some examples.

<Zakim> Lauriat, you wanted to respond to the example with https://www.w3.org/TR/WCAG21/#language-of-page

mc: To claim wcag 3 conformance you need to test outcomes. How you achieve is down to methods. We can hopefully provide guidance on state of stack. Relying on os feature widely deployed may make conformance claims easier.

<Zakim> MichaelC, you wanted to say outcomes still have to be tested

mc: Regardless of where in stack it is met, you still need to meet.

JF: To claim wcag 3 you need to test outcomes, that's tbd. We haven't figured out conformance model.

mc: Red herring for this discussion.

jf: yes and no.
... We just finished protocols call, one of things that we sort of discussed but differed is how to integrate into conformance and scoring.

mc: Referring to outcomes in larger group. General principal is that you know where you are in stack for meeting the outcome.

jf: ok

shadi: confused by discussion. wondering maybe we split it up along the lines of the requirement... page needs to have language attribute. additional requirement... if page allows more pages, need to be able to set pages.
... and need to pronounce... splitting up and so if it applies that particular requirement kicks in.

wilco: The way I see it, what's being proposed is you no longer make the content author responsible for everything. Here's the problem, somebody needs to do it, somebody is responsible. Whoever does it, we need accessible outcome.
... The outcome is it needs to be pronounced correctly. Whoever addresses it at whatever level.

<janina> Pronounce the name of this Austrian Christmas carol correctly: "Still, still, still"

Wilco: Is this then... seems wcag 3 would include all uag and atag. I think that would lead to pushback in chartering conversations.

<JF> +1 Wilco

Wilco: What are the thoughts?

<Zakim> Lauriat, you wanted to add what we have today as an implied user agent requirement for language of page and to speak to the "all of ATAG & UAAG" bit

Shawn: One last thought to language of page... we currently have an implied user agent component.
... the user agent will pass on to AT.
... For language of page common, but newer parts of aria, common for user agent to not pass on.
... And regarding uag and atag, I don't think we are proposing to bring in everything. We will review at each guideline for different methods and info that can support he guidance.
... Chartering is a different discussion we will begin shortly.

<Zakim> Chuck_, you wanted to say that maybe all 3 participate.

Chuck: All 3 levels could participate.

Makoto: Language of page. For jp content, even if we don't set lang attribute, browsers have no challenge in rendering. Screen readers have no issues as well.
... In such case, content providers may think that they don't have any responsibility, already works well.
... My question - does it mean that others don't have to set lang attribute and web page conforms to guideline in wcag 3 automatically?

Shawn: That's a good example to use. Technically that is programatically determinable by examining character set.
... In some instances character set may not be useful, but in this case it does work.

Janina: Saying exactly the same. Could be difficult, you would want a declaration on html element.
... English/german no system may get that unless the agent has cultural knowledge.
... We are working on a spec for addressing in greater detail pronunciation.
... So it can be heard correctly. Hopefully that spec will be ready soon.

Shawn: I have asked for related research.

<Zakim> MichaelC, you wanted to address charter scope

mc: Address q of charter/scope. years ago authoring tools and other groups were not renewed. Didn't think there was enough work to get done.
... If we were to propose a whole new guideline, ac would question.
... If we work on a plan to work on guidance, and come with a plan, we can sell the idea. Charter scope should not be reason to not consider something.

Shawn: Thanks for mentioning.
... This seems worth exploring. We have some complications that we will need to monitor, and actively test.
... Any other thoughts?
... That was first core question. We have addressed 2nd and 3rd questions already!

wilco: I mentioned that I was in favor, but I think that this is critical to wcag 3. machine learning is developing rapidly. We may see things that we have done by hand being done automatically in the future.

<Lauriat> +1 with bells on

<shadi> +1 to Wilco!

wilco: We need a standard that can address and adapt to such circumstances. It's not optional.

suzanne: I agree with wilco. has the benefit of reflecting reality. Will make accessibility more welcome in engineering teams.

<Lauriat> +1 to Suzanne as well, definitely a good benefit

suzanne: The more it can be realistic the more adaptable it will be, even for those w/o subject matter expertise.

<Lauriat> Are we willing to commit to writing each method to target just one of Content, UA or AT?

Shawn: I have a question on next question, pasting...
... Ask clarifying question. For each method, we specifically say that this methos applies to use case of writing content, or applies to user agent?

janina: Suggesting that this makes things a bigger job. end benefit will be worth it though.
... Will enable consumers to focus on engineering focus. Lang and embedded content in different language. A user agent could allow for pronunciation to occur more slowly.

mc: strongly advocate for splitting methods up by what they apply to. Not necessarily comfortable in all cases. May result in more methods. Important because it makes sure we have a clear idea of where the problem is being solved.
... If we allow ourselves methods that apply to multiple audiences, may need to review and split up. I also think it will help with clarity.
... existing techniques are not clear because they focus on more than one audience.
... will help in long term.

Shawn: I think I agree in principal.

Wilco: I suspect that won't be the case. For a quite a few I suspect we will find a single method will work regardless of content author, user agent or at.

<Lauriat> +1 to Wilco

wilco: We will have to try and see how it works out. Let's not restrict ourselves early.

+1 to lets try it out and see how it goes.

Shawn: I think another aspect that's much more blurred, authoring tool and content author, when you have a website that has authors, they are responsible for marking up language.
... If that site has ability to accept user generated content, it is a mix, and resulting content is a mix. Not necessarily a distinction between the 2.
... May be same method that applies at different points.

<Zakim> MichaelC, you wanted to consider dependencies and to explore completeness

<Zakim> Lauriat, you wanted to ask about the less-than-clear line between a web app and an authoring tool for the case of things like user-generated content

mc: To address question... we could address dependencies between methods.
... I don't want to create artificial rules for ourselves. I think that crispness is important for our clarity of thought.
... one more thing, the question of how do you know what is sufficient, I think that any one of it's methods is sufficient, but that may not be true anymore.

Shawn: We have lots to review. For this aspect...
... Clear distinction between these vs one method that can cross... concrete examples will help us to review. I would not be suprised that it comes down to guideline.
... And I expect over time it will change, reality of modern technology is evolving rapidly.
... More we can make sure things are maintainable the better off we will be.

<Zakim> jeanne, you wanted to ask for volunteers to write up concrete examples

<Lauriat> We have a starting point of language of page, I'll dig it up.

Jeanne: Suggest why don't we ask for volunteers to write up concrete examples? If you have an example in mind, we should draw up an outline of what the issues are and how it could work.
... Don't need to write up entire method and outcome. We had some examples already mentioned.
... Can someone write things up?

Wilco: Programatic language group will pick up.

Jeanne: Anyone else?

<Lauriat> Language of page example with authoring tool & VR both noted: https://docs.google.com/document/d/18JyGF-AK8Qgq7DPyVlDYmxoj6814rORxuCf0l0oSb7U/edit

<jeanne> Test Reliability subgroup is writing up prgtrammatic language

Shawn: I dropped an example. Language of page we worked through 3 years ago.
... Doesn't include the user agent aspect. We talked through the different methods we discussed.
... <reads from doc>
... We have vr town... and last one is content management system and authoring tools.
... This is old, but its relevant as a starting point. We could add in user agent aspects.
... And the AT level as well.
... And Makoto's example of looking at character set!
... Only a couple of minutes. Jeanne's question is a good place to end. Other volunteers for creating examples? Can be just as rough as my example.

suzanne: Thinking about writing up an example, COGA examples? Does the AT actually need to exist, or could there be methods for an AT that doesn't exist yet?

Shawn: Really good question. We've not come up with a concrete way of how it works. But that kind of method about "how it can happen" would be good to highlight gaps in tech stack.
... Depending on vr platform being used, there may not be a way to specify language attribute. But could highlight the need.

suzanne: I volunteer.

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.200 (CVS log)
$Date: 2022/01/21 16:01:06 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision VERSION of 2020-12-31
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: janina, jeanne, Lauriat, shadi, Wilco, jenniferS_, SuzanneTaylor, MichaelC, JF, ToddL, sarahhorton, Rachael, Makoto, kirkwood, Azlan, Chuck_
Present: janina, jeanne, Lauriat, shadi, Wilco, jenniferS_, SuzanneTaylor, MichaelC, JF, ToddL, sarahhorton, Rachael, Makoto, kirkwood, Azlan, Chuck_
No ScribeNick specified.  Guessing ScribeNick: Chuck_
Found Scribe: Chuck

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

WARNING: No date found!  Assuming today.  (Hint: Specify
the W3C IRC log URL, and the date will be determined from that.)
Or specify the date like this:
<dbooth> Date: 12 Sep 2002

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]