Silver Task Force & Community Group -- 06 Mar 2020

<Chuck> Hi Jeanne, I'm camped on what I believe is the Friday instance of our Silver call, but nobody else is there. Are you there already? Or just joining now?

<jeanne> goes

<jeanne> virtual

<jeanne> planning)

<Chuck> scribe: Chuck

Silver Face to Face meeting at CSUN

Shawn: Silver F2F at CSUN
... F2F is going virtual. Companies and individuals cancelled, SoC called emergency.
... We are still working on schedule and agenda.
... Jeanne and I worked on some things with Judy and others. What schedule makes sense? We have something that we think works.

<jeanne> https://www.w3.org/WAI/GL/task-forces/silver/wiki/2020_March_F2F_Meeting_at_CSUN#Agenda

Shawn: 2 2-hour sessions each day. Spread out. Participants from around the planet.
... Daylight savings hits US in 48 hours. We spread the pain. We are looking at Monday from 9am - 11am ET.

Jeanne: We wanted to switch that to avoid AGWG.

Shawn: AGWG is cancelled, we don't need to switch.
... Alastair let us know.

<jeanne> https://www.w3.org/WAI/GL/task-forces/silver/wiki/2020_March_F2F_Meeting_at_CSUN#Agenda

Shawn: We are currently looking at 9am-11am ET, then 4pm-6pm ET, the 11am-1pm ET, then (one more late session)
... Topics and times are still tentative. We want to give people opportunity to chime in on suggestions for topics and times that work for them and their interests.

Jeanne: Big break between sessions.
... Big breaks for personal time is helpful.
... Not set in stone. Please provide feedback.

Shawn: We'll also send in email.

<jeanne> scribe: jeanne

Chuck: Those of us who are still going to be there in person, is there a room or gathering place?

<Chuck> Shawn: There is no room for the people who may be present.

SHawn: We don't. The room for 30 people was cancelled.
... I suspect there will be free rooms available that you can use and shut the door

Chuck: I'll look and see if there is anything promising. I'll look.

<Chuck> Chuck: I'll look around on Sunday to see if there's any possible suitable room for the individuals who are physically present.

<Chuck> PK: Are the dial in details set up yet?

<Chuck> Jeanne: We are setting up something with zoom features, that should be accessible. Will be bumpy for the first call.

<Chuck> Jeanne: I'll send email to Ralph and have him set up the times, and provide host keys. It's a new interface, but we know people appreciate zoom and accessibility.

<Chuck> Jeanne: Curious to try new technologies.

<Chuck> PK: CSUN opening keynote openings 9pm pacific, we've an overlap that we don't want to have.

<Chuck> Shawn: We can include... we'll keep schedule as is at the moment, we'll send out notification and highlight that fact.

<Chuck> Shawn: 7-9 is compromise time.

<Chuck> Shawn: We'll send out the times, highlight the known issue, and if we get large number of folk who go to keynote, we'll adjust.

<Chuck> MC: We should nail down times. We want that to happen before he leaves work today. Otherwise won't be set up until Monday, which may be too late.

<Chuck> Shawn: Does that mean that we will have some flexibility in the times? Or should we expect that the times stick but topics can shift?

<Chuck> MC: We could move other time slots if needed, but easier to just shift topics.

<Chuck> PK: Monday morning may be challenging for many.

<Chuck> Shawn: That's the thing we were working out yesterday. How do we set up times where at least a couple of times work for everyone, and nobody is excluded?

<Chuck> Shawn: No individual will look at all the times and say everything works.

<Chuck> Shawn: Should we highlight the topics, or go into normative vs. informative?

<Chuck> MC: Quorum is a bit low.

<Chuck> yes

Normative/informative lines, pros and cons for each (session

<Chuck> Shawn: One of the other challenges is what can we work on, given we'll have a different format than we planned.

<Chuck> Shawn: Most will be remote. One of the things we want to work on is lines between normative and informative.

<Chuck> Shawn: There's open questions from rest of wg we need to address.

<Chuck> Shawn: Our idea was essentially to look at the information architecture of Silver, look at each of the places where we could draw the line (maybe 4)...

<Chuck> Shawn: Define each of those places and then next week when we do the exercise we can walk-through with everybody, show where we put the lines, and state pro's and con's of each.

https://docs.google.com/document/d/1QF5Olq8880_OmP0Eyr7CmFti5O1VIFUXPhguA6DFzlE/

<Chuck> Shawn: That may help with the varying perspectives and opinions. It will be structured, and will give us more documentation for our reasoning.

<Chuck> Shawn: Today in Silver we are building up... we have the user needs, from there we build up the tests for whether or not we meet the user needs, from the tests we build methods,

<Chuck> Shawn: from the methods we build top level guidance.

<Chuck> Shawn: 4 levels of information. We have other bits that are tangental, like tags.

<Chuck> Shawn: We have essentially 5 different places to draw the line for informative vs normative.

<Chuck> Shawn: Going to jot some things down.

<Chuck> Shawn: User needs, tests, methods, guidance

<Chuck> Jeanne: How to's.

<Chuck> Shawn: The tests and methods are wrapped up in the how to.

<Chuck> Shawn: The how to is a categorization of content.

<Chuck> Jeanne: The understanding document equivalents. I would be surprised if we said "how to" was normative.

<Chuck> Shawn: My understanding is it kind of looks like "this...". "how to" and top level guidance.

<Chuck> Jeanne: That's how it looks, but how it looks can change.

<Chuck> Shawn: What other info does how to include that isn't user needs or methods?

<Chuck> Jeanne: Activities, plans, develop, design, evaluate.

<Chuck> Shawn: Those are chunks of tests or methods, right?

<Chuck> Jeanne: No.

<KimD> Information by role

<Chuck> Shawn: It would be good to review an example. Got one handy?

<Chuck> Jeanne: <pulling an example>

https://raw.githack.com/w3c/silver/ED-changes-25Feb-js/guidelines/explainers/visualContrast.html

<Chuck> Jeanne: We can look at "visual contrast".

<Chuck> Jeanne: <reads from docs> If you click on methods it has 2 methods, you can see that what is in the method is the test.

<Chuck> Jeanne: The tests are inside the methods, the methods are inside the how to.

<Chuck> Shawn: So then the rest... it makes more sense to split out... so the "how to" is a categorization, it may not need to be it's own thing.

<Chuck> Shawn: The activity by roll seems like it is kind of suplimentary to the user needs. In between tests, methods, and user needs. Good way to think about it?

<Chuck> Jeanne: yes, if you do it in the order we wrote it, the activity by role goes after methods.

<Chuck> Shawn: It doesn't have to be the order you wrote it. We have this spectrum of information. Which goes from decidedly not normative to near 100% certain it should be normative.

<Chuck> Shawn: I'd like to have it ordered by that. Then we can draw up pros and cons, where everything is normative, vs nothing is normative. These are extremes that will be easy to draw the cons.

<Chuck> Shawn: The lines between will be interesting.

<Chuck> Jeanne: Right now we have it between methods and guidance.

<Chuck> Shawn: For those who haven't participated for an extended period of time, that is not clear why we think that. I think the activity of pros and cons for each of these lines is helpful.

<Chuck> Shawn: We could share why we did it, but it leaves open questions about "why not over here"?

<Chuck> Janina: And that WILL be asked a lot. If we can keep it succinct will be better.

<Chuck> Shawn: My thinking is we have these virtual lines drawn for each possibility. We go into the group activity starting at one extreme and work towards the other... or maybe both. Most productive...

<Chuck> Shawn: Would be to start with "everything is normative" and work our way towards the other side.

<Chuck> Shawn: ...then we dial it back from there.

<Chuck> Shawn: One of the things I anticipate is we don't get to work through each line in a span of 2 hours, particularly if it's a large group. We can probably reasonably say that the top 2 lines we have at the moment...

<Chuck> Shawn: for everything is normative or user needs, probably not super useful. We can get through them, would be helpful to document out, but reverse order is probably more productive.

<Chuck> Jeanne: If you are ok, you have between methods and guidance, I think we should say "guidelines".

<Chuck> Shawn: Agree.

<Chuck> Jeanne: Already more than one page :-)

<Chuck> Jeanne: Let's start with easy ones. Between user needs and activities.

<Chuck> Shawn: This is skipping a couple of levels?

<Chuck> Jeanne: What's being skipped?

<Chuck> Shawn will copy and paste.

<Chuck> Jeanne: First con is that everything is normative is that it's too difficult to maintain.

<Chuck> Shawn: Agreed.

<Chuck> Shawn: It's worth rephrasing that... "why is it difficult to maintain"?

<Chuck> Jeanne: It requires an update for any changes.

<Chuck> Janina: One level up, I am having difficulty determining how you make a user need normative? How do you "make" them have user needs?

<Chuck> Janina: That seems undefinable to me.

<Chuck> Shawn: Also means that the act of us making a normative list of user needs, we would be getting consensus on excluding others.

<Chuck> Janina: Everything normative is impossible.

<Chuck> Shawn: That's why we want to call out the extremes, because we'll get consensus among everyone. That paves the path towards consensus for something in between.

<Chuck> Jeanne: I'm going to put "non" for pros. Let's move on.

<Chuck> Shawn: Let's skip to everything is informative. The act of walking through this list for the between boundaries are the important ones I want to discuss with the larger group.

<Chuck> Shawn: If we can start off with the structure, and if we show the pros and cons we have already filled out, for the everything is informative, means it's not a guideline you can rely on.

<Chuck> Jeanne: One thing that is troubling, one of the proposals from agwg is that the tests are normative.

<Chuck> Shawn: Test and methods and guidelines are all normative.

<Chuck> Janina: We could say everything is informative, but then we require nothing from anyone, and there's nothing to conform to.

<Chuck> Shawn: It's essentially it's saying "here's how you can conform, can change at any time, good luck".

<Chuck> Jeanne: It is flexible and easy to update, we wouldn't even need a charter!

<Chuck> Janina: Even a w3c note needs a charter.

<Chuck> Shawn: We could say everything is in github. :-)

<Chuck> Shawn: Are there other cons?

<Chuck> Shawn: for the "everything is informative"?

<Chuck> Jeanne: Not suitable for regulations. But is that because of the first item, or other things?

<Chuck> Janina: There would be no international consensus. Policy on the planet means some combined entity like EU, or US (50 states).

<Chuck> Janina: There's nothing for them to hang on to, they would have to invent it. Lots of disagreements.

<Chuck> Jeanne: I want to point out that there are places in regulation around the world that do mention the bbc guidelines for mobile.

<Chuck> Jeanne: It's not that it's required for regulation, just better.

<Chuck> Janina: When it didn't exist, they found one that did exist, or the governments would need to get together and invent something.

<Chuck> Jeanen: For a con it would encourage divergence of standards, not harmony.

<Chuck> Janina: Nobody in the industry would support it.

<Chuck> Shawn: We have similar things like that today. To your point it would be worse.

<Chuck> Janina: We are trying to minimize that. We have divergence of blocking internet at state boundary. Don't know we could do it that granular.

<Chuck> Janina: Aren't those full stop... rather than actual analysis?

<Chuck> Shawn: I have a q... around the context of introducing this exercise for the group activity.

<Chuck> Shawn: Where are the definitions of normative and informative for W3C?

<Chuck> Janina: IETF's RFC 2119.

RFC2119

<Chuck> Janina: I'm happy to go find it. If Michael doesn't know where it is.

<KimD> within WCAG 2.1: https://www.w3.org/TR/WCAG21/#dfn-normative

<Chuck> Janina: We are looking for w3c definition of normative and informative. Do we specify it?

<Chuck> MC: I've not found a w3c definiton. There are some common definitions used in specs. No central resource that defines it.

<KimD> within 2.1: https://www.w3.org/TR/WCAG21/#dfn-informative

<Chuck> Shawn: Very helpful, especially Kim's comments.

<Chuck> Shawn: Can we define for ourselves? Do we need?

<Chuck> MC: I think we need them. They are widely expected. There's a doc which says there should be a conformance section. We'll be asked about that.

<Chuck> MC: We could in principal have a different definition than other groups use, but that could introduce confusion, wg may not accept that.

<Chuck> Shawn: I bring it up because of the desire to use plain language. The definition of informative includes definition of jargon. Not very plain.

<Chuck> MC: We could adopt some plain language for the definition.

<KimD> Also this: 5.1 Interpreting Normative Requirements§ The main content of WCAG 2.1 is normative and defines requirements that impact conformance claims. Introductory material, appendices, sections marked as "non-normative", diagrams, examples, and notes are informative (non-normative). Non-normative material provides advisory information to help interpret the guidelines but does not create requirements that impact a conformance claim. ...

<KimD> ...The key words MAY, MUST, MUST NOT, NOT RECOMMENDED, RECOMMENDED, SHOULD, and SHOULD NOT are to be interpreted as described in [RFC2119].

<Chuck> Shawn: It seems like the important thing to have within anything that comes from w3c is that there's some kind of line that says what the line is for conformance. Everything else is to help understand.

<Chuck> MC: The conformance spec.

<KimD> https://www.w3.org/TR/WCAG21/#interpreting-normative-requirements

<Chuck> Janina: Does plain language require that we use the simplest possible terms? jargon can be useful.

<Chuck> Janina: Without jargon may take education to understand. Could be difficult to express in plain language. We would want to go the other way, and define jargon term, when standard practice for setting standards.

<Chuck> Janina: Those that build this stuff expect jargon. If we change that we will introduce a complication.

<Chuck> Shawn: Back to the q of plain language, do we have to use simple language no matter what? No, if we are defining a technical method, we have to use technical terms for svg.

<Chuck> Jeanne: If you go back to the example of visual contrast, where I wrote normative, I put "required" (normative), the way that we have been advised by coga to handle this...

<Chuck> Jeanne: you say the explanation first and then you put the jargon in paranthesis and then you can use the jargon.

<Chuck> Jeanne: We want to give people the ability to get a one click or hover (or however) means to popup the definition. So that they always have the plain language. We don't stop using technical terms.

<Chuck> MC: In wcag 2.0 our definitions could be dropped in sentences and work, may not work always in practice, but was the effort we made. We want to avoid overcomplicating.

<Chuck> MC: We want to avoid over explaining normative and informative. We want to document what sections are normative and what are informative. Let's not make it harder to read.

https://raw.githack.com/w3c/silver/ED-changes-25Feb-js/guidelines/explainers/visualContrast.html

<Chuck> Shawn: If we start the conversation with folk who haven't been in agwg for a while, starting the conversation may not be welcoming to draw them in unless we explain normative and informative, then we can move on tot he acvtivity.

<Chuck> Jeanne: <reads from visual contrast for an example>

<Chuck> Janina: That sounds good.

<Chuck> MC: That could get repetitive if on every page.

<Chuck> Shawn: That's one of the challenges if we don't know the entry point, but tough figuring out between now and next week.

<Chuck> Jeanne: We do have to figure out where we are drawing the line. This is a good exercise for that.

<Chuck> Shawn: What does everybody think about for how we have the doc organized? Start from bottom and work our way up?

<Chuck> Janina: you could focus on things that aren't controversial, get consensus among the group, then start figuring out in between ones.

<Chuck> Shawn: We have a few pros and cons listed for each of those, but we can start with the extremes.

<KimD> Does "Between Methods and Guidelines" mean "Methods and Guidelines are normative"

<Chuck> PK: That makes sense, but I would add, how we want to deal with conformance, at the page and site levels. Where we are providing guidance. May be tricky to hash out, but I don't want us to forget it.

<Chuck> Shawn: That's an activity Monday morning, working through scoring examples.

<Chuck> Shawn: Talking about scope, how it works for non-web. We need to work out, but in terms of normative vs informative. Even though affected by it, I don't think it's necessary for initial consensus.

<Chuck> PK: Makes sense, but would be useful for Monday morning would be "some of what we do will be informative examples", and some of what we do will be "you must do this..." examples.

<Chuck> Kim: Confused by "between tests and methods". Test's and methods would be normative? Or what's normative goes in between?

<Chuck> Shawn: The way we've been thinking of this has been the line itself between normative and informative, the line between tests and methods would be that tests are not normative, but methods would be.

<Chuck> Shawn: Something else to consider (but may be overly complex) if we had test but not methods. In order for us to say "the things that needs to pass" but we don't say how you get those things to pass.

<Chuck> Janina: I think we have something like that. The mathematics are a test. Very quantifiable. I suspect we are going to make it more specific than a test and informative. Aren't we going to put the math at the guideline?

<Chuck> Shawn: Yes-ish. the math itself is not necessarily the test. In wcag the math is the test. But for visual contrast it depends on contrast. The math in context is to assess, but not the sole way of determining if you pass.

<Chuck> Shawn: Like the share button on google docs. Visual Contrast for that is good for a button, but poor for some text.

<Chuck> Shawn: The mathematics are concrete, but the line for passing and failing may vary.

<KimD> I think a one-by-one is better

<Chuck> Shawn: We only have a minute. With that... should we restructure this to a concrete strictly "normative/informative" discussion? Should we tackle one topic at a time?

<Chuck> Janina: Yes.

<Chuck> Kim: one by one is better.

<Chuck> Janina: Your approach is probably more correct, but this will be easier.

<Chuck> Kim: If the pages re-arrange, I think it's better to understand conceptually, "yes methods should be normative" for example. Doesn't matter if they move or if they are inbetween.

<Chuck> Shawn: I think we are prepared for next week. We'll follow up in email.

- DRAFT -

Silver Task Force & Community Group

06 Mar 2020

Attendees

Contents

Silver Face to Face meeting at CSUN

Normative/informative lines, pros and cons for each (session

Summary of Action Items

Summary of Resolutions

Scribe.perl diagnostic output