Silver Conformance Options Subgroup

4 Mar 2021


Azlan, Bryan, jeanne, Jemma, JF, John_Northup, PeterKorn, sajkaj, sarahhorton, ToddLibby, Wilco

Meeting minutes

Agenda Review & Administrative Items

<PeterKorn> sajka: administrative note: we will be out of sync US to the rest of the world on DST.

<PeterKorn> sajka: in two weeks (March 18th, March 25th) we will be out of sync DST. We will retain the US time.

Assigned github issues

Issue 219

<sajkaj> https://github.com/w3c/silver/issues/219

<PeterKorn> Peter reads

<Zakim> JF, you wanted to comment on Sailesh's comments

<PeterKorn> JF: I think that Sailesh is echoing something that Gregg Vanderheiden said as well. When we have "partial", how do we measure that? How good is good enough?

<PeterKorn> JF: On specific requirements, even though we may get to the point where we accept less than perfect, the ability to do that measurement... I see that as being a concern, and one I've been sharing. That's how I read.

<PeterKorn> John: I agree it is a slippery slope, but sliding toward greater conformance. I th ink with our approach, we are encoring that and giving them waypoints.

<PeterKorn> Jeanne: think Sailesh is misunderstanding what we are going for with "Substantial Conformance". So far for WCAG 3, have strict definitions for how much they can stray from perfection, being defined for each outcome of each guideline.

<PeterKorn> Jeanne: We can say what is critical, what isn't. For his example, that could be written as a critical failure if crucial for completing the task. Believe we have started addressing his concerns. [Note too, his comment is from November]

<PeterKorn> Jeanne: Feels it is a misunderstanding of the term "substantially conformant"

<PeterKorn> Wilco: Thinks this is exactly Judy's concern, and the response is in that same line. Intent is not to let people off with sloppy a11y.

<Wilco> peter: because we haven't spoken to substantially conformant, i want to ask Sailesh to review what we have to say about substantially conformant when its's ready.

<Wilco> ... Until we have a rough consensus on how to deal with this is it premature to say anything other than we'll take his concerns into consideration.

<JF> +!, WFM

<JF> s+!/+1

<PeterKorn> [agreement on how we will respond to this issue: close it as we haven't addressed substantially conformant yet; it isn't yet something in FPWD]

Issue 277

<sajkaj> https://github.com/w3c/silver/issues/277

<PeterKorn> Peter reads...

<jeanne> See the Draft Answer Templates

<Wilco> Peter: Thank the submitter, possibly add a link to issue 217 that we just talked about. Common question is if an organisation for some reason lets a significant number of bugs exist for some time, it's an open question, do we recognize that in some way in conformance

<PeterKorn> Wilco: Comment looks a lot like one from Deque. Key point: industry is collecting issue data & prioritizing them. If we use the metrics already in use & widely adopted, we don't have to change entire industry & how they are working. We also don't have to collect a lot of different info

<PeterKorn> Wilco in systems widely adopted. And there are ways around "Spoons" problem.

<Zakim> jeanne, you wanted to say that he is a manager in Trusted Tester

<PeterKorn> Jeanne: Want to clarify - Bruce didn't file this issue. He filed on behalf of Drew Nielson. Came from a very good discussion. Caution is how we handle cumulative issues - "Spoons problem". We must solve spoons problem first.

<PeterKorn> Jeanne: Point system provides a solution, may not be best or final. Wouldn't want us to recommend going forward without that.

<sajkaj> ck saj

<Wilco> +1, that needs a solve, for sure

<PeterKorn> Sajka: candidate answer - people making reference to standard industry practice. We should certainly examine SOP on this (outside of a11y). Are they codified somewhere? A document somewhere?

<Zakim> sajkaj, you wanted to ask aboud industry sop

<PeterKorn> Sajka: thinks a response is great - if we can make it work, but where are standards? Don't think this is in ISO 9001

<PeterKorn> https://www.getzephyr.com/insights/agile-strategies-managing-bug-fixes

<PeterKorn> https://michaellant.com/2010/05/25/a-simple-agile-defect-management-process/

<PeterKorn> https://www.linkedin.com/pulse/testing-agile-quick-tutorial-defects-bugs-everything-between-sherman/

<PeterKorn> Bryan: Likes what Peter said about Agile. Strong correlation in way we approach a11y could benefit from Agile methodologies. Another approach: Underwriters Lab on safety: How do you determine something is safe? Meets a certain level of support? Anything we can pull from there?

<Zakim> jeanne, you wanted to say over time, versus snapshot

<Jemma> +1 to bryan

<PeterKorn> Jeanne: 2 things developing. A snapshot in time, and a process by which issues are fixed. In many ways, our approach has been a snapshot in time. Meets at this moment, and so conforms. But what we've been talking about @ Silver & Gold levels is looking at a system over time.

<PeterKorn> Jeanne: What do we want to encourage people to do with a system over time. Is this something that is measurable that could go to Silver level (how does a company address problems as found)

<Jemma> +1 to jeanne

<Jemma> good summary of challenges we have

<JF> +1 to Peter

<Jemma> I like the subtle distinction between badge and medal

<PeterKorn> JF: want to riff on Peter's comments. "Mechanical a11y" and "editorial a11y". Codebase may be great. Then someone writes "Click the red button on the left". Concerned about snapshots for recording these issues.

<jeanne> I think Peter has an interesting idea of separating out a mechanical level

<PeterKorn> JF: Struggling to figure out how we record / measure over time.

<PeterKorn> JF: maybe we should revisit badges/multiple currencies idea from before

<jeanne> I think the "multiple currencies" is Silver and Gold level

<PeterKorn> Wilco: definitely automation part... Things aren't black and white: "This can be automated, this cannot". Tools make quite a few assumptions about what they are doing. Tools that assume more can automate more.

<PeterKorn> Wilco: also, the more you know about the site/tool, the more you can automate.

<PeterKorn> Wilco: making decisions on what can be automated today or not isn't the way to go

<PeterKorn> Wilco: Also want to revisit the idea of making WCAG 3 multiple standards. One is snapshot "this is a requirement". Another "as the whole - website or app - how do you test the whole thing"? As Peter said, can't have 1m humans testing every page update.

<PeterKorn> Wilco: also, did you organize yourself to address a11y in your development processes.

<Zakim> jeanne, you wanted to talk about WAI-CooP and the plan for workshops for Sampling and Scoring

<PeterKorn> Jeanne: New WAI group funded by EU "WAI-COOP" [Wilco is in it]. In meeting with Shadi, who was suggesting 2 workshops. One on sampling, one on scoring.

<PeterKorn> Jeanne: Could present some of our / WCAG 3 ideas on scoring, get expert feedback from such a workshop.

<PeterKorn> Jemma: returning to how to respond to issue... will hold her thoughts.

<PeterKorn> Peter [attempting to scribe himself a few moments after the comments were made]: Shall we return to the GitHub issue? Respond thanking them for this idea, sharing our appreciate for it & that we will think about it. Then linking to the previous issues (#217)

<PeterKorn> Peter: (continuing), and without suggesting a solution, noting that there are organizations that allow some number of Critical bugs for some short amount of time. If there are critical a11y bugs, what then?

<Zakim> jeanne, you wanted to say that I do

<PeterKorn> Jeanne: don't think we should link to 217. They are different enough, they might confuse things.

<PeterKorn> Jeanne: suggest we will get back to him when we have a specific proposal for him to look at (speaking of issue #277).

<PeterKorn> Sajka: we might also ask him to point us to documentation on the severity levels we can cite?

<Wilco> peter: Suggest we add label.

<PeterKorn> Jeanne: cannot close the issue. When ready to close, needs to be in a survey that goes to AGWG, is closed by AGWG.

<Wilco> peter: good enough for in the subgroup, so we can track it

<Wilco> Peter: keep it in the group, looking at use cases.

<jeanne> https://www.w3.org/WAI/GL/task-forces/silver/wiki/Procedure_for_Processing_Comments

<Wilco> Jeanne: This has the draft answer templates.

<Wilco> Janina: looking for more information before we want to finalise

<Wilco> Jeanne: Could divide into separate issues.

<Wilco> Janina: Not ready for a response. Will ask for more information on industry practices.

Minutes manually created (not a transcript), formatted by scribe.perl version 127 (Wed Dec 30 17:39:58 2020 UTC).