Silver Task Force & Community Group

19 Nov 2021


jeanne, JF, jenniferS, shadi, Wilco, SuzanneTaylor, Chuck_, Makoto, sajkaj, GreggVan, kirkwood, Rachael, Jemma, Lauriat, Jennifer, ToddLibby, jeanne2


<Lauriat> Thank you for signing up to scribe today, Jennifer! https://www.w3.org/WAI/GL/task-forces/silver/wiki/Scribe_List

Jennifer: scribe

<scribe> scribe: Jennifer


No meeting next week, 26 November for U.S. holiday

Thursday and Friday meetings are cancelled next week

due to the U.S. holidays

Writing Testable WCAG 3.0 Outcomes

Please sign up to scribe in December! :D

<Lauriat> https://docs.google.com/document/d/1sugAtqie_x1XqHDZo1Im7ftDNllWeRV_ty4PULeoTV0/

Sign up to scribe: It's on the main page of the wiki near the bottom with the meeting information. https://www.w3.org/WAI/GL/task-forces/silver/wiki/Scribe_List

Wilco discussing Writing Testable WCAG 3.0 Outcomes

Expect to survey in AG.

Chuck said it is undecided.


Document starts with brief definition or explanation of testable outcome.

<Chuck_> I think Jeanne wanted to chime in.

Four qualities:

address a clear user need

little to no room for interpretation

quick & inexpensive to test (possibly using tools) can't take hours

must be easy to learn

brief section on why testability matters

then defined a process to creating these outcomes, in 5 steps

1. come up with examples & edge cases

2. create a rough outline of the outcome

3. simplify where you can


3. define the terms used in the rough outline

4. simplify where you can

5. complete the how to & methods

6. get feedback and iterate

Hey, that's 6 steps!

In the document, the material is expanded.

<Lauriat> qv?

Testing qualitative aspects

Rather than 'legible' WCAG uses color contrast to evaluate legibility.

To summarize, if you can read more than one meaning into a word, then that's an ambiguous thing.

Series of definitions in the Google doc.

<Fazio> sample sizw?

Next section is on the soundness of outcomes.

David Fazio question: For subjective and qualitative stuff, have you considered sample size requirements?

Wilco: there's no "this many people need to agree". We do it by discussion. There are no shortcuts.

A process we go through over and over until enough people have no objections. It's building consensus.

How Nuanced: a trade-off between accuracy and complexity. No clear way to do that, yet.

Shadi: one of the aspects that addresses subjectivity is to reduce the decision space. i.e., text alternatives.

separate text alts for buttons versus others, then the subjectivity would be reduced.

Wilco: you want to break it down, have a narrower definition.

But we can definitely break it up to do a better job of expanding on that.

<JF> https://www.w3.org/WAI/tutorials/images/decision-tree/

JF: might be out of scope for this conversation, such as the subjectivity of text alts

we have the EO's decision tree

there's subjectivity in determining a text alternative

using that guidance, then that helps contribute to getting consistency

you'll never have 100% agreement, but if we can evaluate that text alt is there and the decision tree was used, it could be worked into conformance somehow

Wilco: the last thing I wanted to highlight was subjectivity vs objectivity. we use those terms somewhat inconsistently.

when ACT uses objective, it basically means programmatic.

"subjective" does not mean vague or poorly defined. things can be subjective and still be tested consistently.

<SuzanneTaylor> +1 in general to this guidance - I think it will be very helpful toward moving forward with writing outcomes

Jeanne: I think you covered it really well.

Chuck: any comments or questions before moving on?

David Fazio: I've been receiving complaints lately that if we need a glossary then it isn't plain language.

Jennifer: we are working cross-culturally so the glossary helps with that.

Wilco: in ACT we try to use only definitions of things that are fairly intuitive. if you can't guess, then it's probably something that should be defined.

No meeting next week, 26 November for U.S. holiday

Resolving Conformance Issues

JF: I disagree with that. I think glossaries are an important part of what we do. Standards are intended to remove ambiguity.

<Fazio> I'm just bringing awareness of the public's comments

Understanding specific ways of using terms, glossaries are likely an important part of what we're doing.

Shawn: we can review later.

<Lauriat> https://lists.w3.org/Archives/Public/public-silver/2021Nov/0025.html

picking up from last Friday, issue 304

<Lauriat> 304 - conformance claims need feedback mechanisms

<Lauriat> https://github.com/w3c/silver/issues/304

Can someone pick up where that left off?

Jeanne: I took an action item to check with maturity model to see if a user feedback was included.

I checked with David, eh said they did.

David joined this call for that.

Would someone please write that answer so we can document?

David: we orig had it in one place
... so people could voice commentary/complaints, so feedback could be actionable

<Fazio> ohttps://docs.google.com/document/d/1Y5EO6zkOMrbyePw5-Crq8ojmhn9OCTRQ6TlgB0cE6YE/edit#

David: so we put it in a different place. The maturity model is public so you can check https://docs.google.com/document/d/1Y5EO6zkOMrbyePw5-Crq8ojmhn9OCTRQ6TlgB0cE6YE/edit#
... we're still editing the Github version.

<Lauriat> https://github.com/w3c/silver/issues/352 - multiple conformance models vs simplicity

Jeanne will write the answer to that for next week./

Chuck: Encourage others to take ownership, so Jeanne isn't doing 99% of the work.

Suzanne: I can write the response, I'll write it now.

Jeanne: Awesome! Thank you, Suzanne.

<Lauriat> https://github.com/w3c/silver/issues/352 - multiple conformance models vs simplicity

Lauriat: have we been reading through and discussing?

<Lauriat> "What do you see as the possible tradeoffs in using a flexible conformance approach? Would you want to see multiple conformance models? The tradeoffs are that you’ll have some loss of quality in areas. I wouldn’t want multiple conformance models. I would encourage simplicity."

Jeanne: yes, that's what we've been doing, unless really complex. opening up for discussion.

<Chuck_> +1

Lauriat: We want to compare & contrast, and get down to one conformance model.

Jeanne: I think the person who contributed that interpreted that as multiple, so I think it's just a clarification.
... the flexibility is in the types of tests that we are writing and in the moving away from 100% or failure.
... so that's more the principle of what we meant by flexible, rather than multiple conformance models.

<Zakim> JF, you wanted to note that we will still have 100% or failure

JF: note that we, even with flexible conformance or scoring model, we will still have 100% or failure. In my mind, we do have some tolerance for some failure (silver vs bronze)…
... this is something that has concerned me going forward, as we move to bronze, silver, gold… bronze will be less than perfect.

Jeanne: that isn't what went in the FPWD, I think we should stick to that, this is the direction we're going, and that may change.

<Wilco> +1 to Jeanne

Jeanne: plz put that in, that we do have disagreement, and that may change.

Chuck: we will have one conformance model, rather than this will be the conformance model.

JF: the question in my mind is still if gold is 100%, then silver & bronze are less than 100%.

Jeanne: bronze was regulatory, silver was usability, gold was [something]

Chuck: I don't think the answer to this needs to address that to provide a satisfactory answer at this time.
... I'll write something, and we can tweak if as necessary.

<Lauriat> https://github.com/w3c/silver/issues/386 - if we allow 95% to pass, people will stop at 95%. Proposal for addressing minor errors

Lauriat: this comment: I do think that organizations will stop at 95%. call me cynical, but I think that's true. I'm not a technical person in this area. is there a way to weight this that things have to be 100%. is there a way things can be missed without failing… details in link.

<JF> I won't comment further here, but again, issue #386 is asking for less than 100% - but 100% at Bronze, or Silver, or Gold?

Jeanne: it's an interesting idea. Not exactly what he said… trying to identify what the images… what I think is an interesting idea is the idea of looking… instead of defining what's the task… which we've struggled with… it might be easier to say,
... something that's used only once or not common, it might be easier to… it's much easier to count the negatives instead of the positives. might be interesting idea to explore for the group working on tasks, defining processes.
... I know Rachael has worked on that.

Rachael: we've only just started to discuss.
... we've brought up the difference between critical error and the concept of priority against any given item within a set.

<JF> who decides the priority?

Rachael: instead of top down, think you have an image, in the order of things it may have a different priority… I have it noted as something to discuss with group sometime after Thanksgiving.
... I'll take it to respond.

<Rachael> @JF, priority would be set by working group.

JF: who decides priority of…
... it's dependent on how it's used, so subjective/contextual discussion. Industry wants less subjectivity.

Rachael: I understand, John.

Jeanne: I don't think we can close it, I think we can transfer it to the group working on it.
... transfer to the process group, but not close it.
... you can take this off your to do list

<Lauriat> https://github.com/w3c/silver/issues/394 - the effort of proving conformance can overshadow development.

Rachael: I run that subgroup, it's still on my to do list.

<Lauriat> We are concerned that the effort of proving conformance would overshadow the effort of conformant development. Will you provide guidance for large organizations for testing and showing results?

Lauriat: my sense is the answer to this is yes, we will have some amount of this, here's how you can put together a statement of conformance.
... this is something we haven't completely solidified, yet.
... so a response to this, so far, the way we've approached this convern
... it's certainly not the first time we've heard that this might be a lot of work
... as we start, things will start off complicated — what can work — then we can refine it from there.

<jeanne2> +1

Lauriat: volunteer to draft that response?

<Fazio> we are conscientious of scalability

Lauriat: 30 seconds to find a volunteer

<Fazio> in our Maturity Model at least

Lauriat: I will draft it.

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: 2021/11/19 16:02:13 $

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)

Succeeded: s/ @stevelee - I think you may be in another call.//
Default Present: jeanne, JF, jenniferS, shadi, Wilco, SuzanneTaylor, Chuck_, Makoto, sajkaj, GreggVan, kirkwood, Rachael, Jemma, Lauriat, Jennifer, ToddLibby
Present: jeanne, JF, jenniferS, shadi, Wilco, SuzanneTaylor, Chuck_, Makoto, sajkaj, GreggVan, kirkwood, Rachael, Jemma, Lauriat, Jennifer, ToddLibby, jeanne2
Found Scribe: Jennifer
Inferring ScribeNick: Jennifer

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]