Accessibility Conformance Testing Teleconference

01 Apr 2021


Wilco, trevor, sajkaj, bruce_bailey, JakeAbma, MelanieP, SuzanneTaylor, shadi, anne_thyme, CarlosD, Makoto, Francis_Storr, kathyeng, Lionel_Wolberger, Jemma, jeanne



Wilco: Welcomes all, noting it's a first joint meeting for Silver and ACT
... Reviews goals for the mtg
... Find areas of collaboration
... Identify common problems and attain a shared understanding
... Define how to collaborate on further work
... Would like not to talk about solutions today

Wilco: Asks we use the queue today, as we have significant participation

wilco: Starting with a short ACT intro ...
... ACT started some 6 years ago with the goal of promoting consistent testing
... Noting that various orgs had inconsistent ways of testing--one org compared to the next

<Wilco> https://act-rules.github.io/rules/23a2a8

Wilco: ACT defined rules of what needs testing and what the reqs are for that
... The rules may be automated or manual -- or inbetween at some level
... Have some 90 rules with 7 orgs implementing
... We track implementations to document who has implemented which rule and whether they're consistent
... Strength of ACT is to make softer reqs more substantial
... Asks Jeanne to intro WCAG3 and where ACT might help

jeanne: WCAG3 should be the next major rev of the WCAG series

<ChrisLoiselle> need to step away for a minute, apologies

jeanne: Notes now stands for "W#C Accessibility Guidelines" because encompasses more than just web; apps, etc
... Incl authoring browser and AT ...
... FPWD in January more illustrating concepts and approach
... Always wanted to work with ACT
... But we had to figure out how WCAG3 conformance was to work
... Reducing ambiguity and covering more a11y needs was a major goal
... This pushed us away from the binary pass/fail approach
... Took longer than expected to get our approach
... Have good response on FPWD on approach, but want to see less ambiguity on methods
... Most participants are not testing specialists
... Believe we've consolidated on the general direction, point system, incl more pwd, allowing rating scales, -- generally positive feedback
... Now makes help from testing experts most welcome

<Jemma> good summary!

Lionel_Wolberger: Asks jeanne to elaborate on testing ask? eval, vs testing ...

jeanne: One of the great questions for guidance from ACT
... We want repeatability, clearly
... Notes two subgroups, Conformance Options, and also Conformance Architecture Testing
... Former doing use cases, particularly edge
... Later looking at validity type concerns

Wilco: ACT looks at the narrowest of scopes; whereas eval is a much bigger thing
... So building up from small pieces may be locus of overlap
... Asks for an example to consider ...
... Clear words??

jeanne: Color contrast ...

<jeanne> https://www.w3.org/TR/wcag-3.0/#visual-contrast-of-text

jeanne: notes guideline, like to a howto which is tech neutral, section on outcomes, ...
... Reads outcome -- and notes expanding icons with categories where this applies, then also critical errors category
... "Critical Errors" is Silver's way of making WCAG3 more consistently fair across pwd groups
... Notes existing concern that categories of pwd aren't being treated equitably in current WCAG
... Trying to be more equitable as a equal rights issue
... Then notes the rating and suggests this is the eval where ACT could help
... Notes the testing tool which is an algo
... Testing provides a score based on the algo;
... One challenge is many tests need still to come back to a consistent way of evaluation
... different guidelines may have different evaluation mechanisms
... in Color Cointrast more than 15% below yields a 0 rating -- considered a fail

<Francis_Storr> Chrome bug tracker for implementing Silver visual contrast in Dev Tools: https://bugs.chromium.org/p/chromium/issues/detail?id=1121900 You can find this working in builds of Chrome Canary.

jeanne: rating 4 all tests are at or above expected threshold values

<jeanne> https://www.w3.org/2021/03/26-silver-minutes.html

Wilco: Suggests this one particularly well defined by Silver

jeanne: Our notes suggest it highlights inconsistency of how we apply across guidelines
... Not all the 0-4 for each guideline are consistent
... We have guidelines where 4 exceeds a la AAA types
... We certainly could discuss clear words

Wilco: Asks any comments from ACT?

trevor: Some thoughts, but a different discussion ...

kathyeng: Question: How it determined the break points in the visual contrast scale?

jeanne: Comes from the methods ...
... Expect this will be automated

<jeanne> https://www.w3.org/WAI/GL/WCAG3/2020/methods/font-characteristics-contrast/

bruce_bailey: The numerical rating has lots of math and science behind it, but the break points are in flux as I understand it
... We were figuring out how to assign the adjectival ratings

jeanne: Different subgroups picked different ways of slicing 0-4 ratings

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

jeanne: On visual contrast there is no 7:1 equivalent; rough approximation to 4:1

ChrisLoiselle: We worked on a table which needs updating, but goes into detailed based on the table

Wilco: Speaking for myself--have built out what needs to be a req in a rule to get high quality consistent rules
... suggest formalizing these areas

jeanne: notes alt text referenced ACT rules;
... Are ACT rules at testing or eval level?

Wilco: Rules take a page and figure out are there things to test; then answer a set of expectations for each item there is to test

<bruce_bailey> Just a point of clarity on the 15% figure Jeanne mentioned several minutes ago, and Kathy asked about...

jeanne: So we want ACT rules included in testing; in our methods which is WCAG3 version of techniques
... Anywhere we could find a rule we've included it

shadi: Is not the guideline what I need to meet? the normative part?

jeanne: guideline and outcome

<bruce_bailey> That 15% was what percentage of body text content was allowed to slip below the minumum allowed for a targeted SAPC rating.

jeanne: guideline is normative but more of a classification

shadi: looking for the test I have to meet normatively? what am I required to do as a web author

jeanne: luminence contrast between background and text
... that makes text easy to read

shadi: Repeatability would have us dive into this to make "easy to read" as clear as possible
... arguable it could stop at adequate luminence/contrast
... Brings up also question of situations like transparencies
... need to work on how independent people would interpret and understand
... what is foreground or background, what text, ...
... ACT not necessarily the testing or eval level
... perhaps some defs are missing or insufficiently clear
... eg. images of text

ChrisLoiselle: we've had an thread reviewing ACT to see how we can embed
... it's all fluid right now
... notes several subissues we're looking at to break the overall task into subparts
... definitely interest in that

jeanne: one thing that might be helpful is to work with us on these early guidelines and point weaknesses, guideline by guideline
... perhaps starting with the first 5 published guidelines
... ould help us learn of what to look for; training on how to approach this
... would help quality of our writing improve ... might that work?

<SuzanneTaylor> +1 to idea of that kind of training/feedback meeting

Wilco: good idea, will need to consider
... perhaps a joint Silver/ACT subgroup

jeanne: could have someone from each Silver subgroup writing content -- very interested in training as many Silver possible as reasonable

shadi: suggest ideal for F2F
... Little harder when virtual
... seems to need a focused work session -- not just hour telecons

jeanne: Notes plan to work on comments in a daylong with AGWG; could try a virtual day for ACT?

Wilco: Might be easier to commit to one or two block meetings rather than ongoing subgroup

<MelanieP> Love the idea!

Wilco: Asks any objections to the day block approach


<Jennifer_C> +1!

trevor: How much are we considering a new ACT format for WCAGT3?

trevor: Any recommendations on rules format revision?

Jemma: applicability for atomic and composite rules -- don't see problem for atomic
... but do for composite
... phps good first step what kind of rules best for silver? and how to deal with composite rules for wcag3
... we need to consider how to make composite work

<jeanne> I think that working through some of the existing guideline examples will help us identify whether a new ACT structure is needed.

Wilco: very much in favor of updating rules
... blend fules with WCAG3

<jeanne> +1 to blending

<shadi> +1 too

Wilco: figuring out how to blend strength of ACT with silver

Jemma: define blending?

Wilco: both what we need from ACT rule and Silver methods

<Zakim> jeanne, you wanted to say that approaching the details will also require WCAG3 changes

Azlan: our composite is not necessarily in line with silver

jeanne: silver is willing to make changes to better accommodate what already exists in act
... it's not just ACT we're looking from adaptation from; silver also willing
... we're not deliberately working against act approach
... virtual workshop should help us learn what needs to adapt

anne_thyme: act rules could be better adopted, hopeful wcag3 can do more

<ChrisLoiselle> I think wording needs to align, An ACT Rule must contain one or more expectations. The expectations describe what the requirements are for the test targets derived from the applicability. An expectation is an assertion about a test target. When a test target meets all expectations, the test target passed the rule. If the test target does not meet all expectations, the test target failed the rule. If there are no test targets, the [CUT]

anne_thyme: getting consistent results from rules testing would be very helpful outcome

<ChrisLoiselle> https://www.w3.org/TR/act-rules-format-1.0/#outcome

ChrisLoiselle: notes expectation URI above

<Jemma> +1 to Chris

<jeanne> +1 to aligning glossary and terms

<Wilco> https://www.w3.org/TR/act-rules-format-1.0/#accessibility-requirements-mapping

Wilco: challenge may be what satisfied/not-satisfied decision might look like under wcag3
... probably the biggest challenge

jeanne: Thanks ACT for time and interest, looking forward to more

Wilco: Thanks all for joining

RESOLUTION: ACT and Silver agree to hold a day-long set of virtual meetings to begin joint effort sometime in the next couple months

<jeanne> ACTION: jeanne will work with ACT leadership to schedule and plan meeting. Also include AGWG chairs

Summary of Action Items

[NEW] ACTION: jeanne will work with ACT leadership to schedule and plan meeting. Also include AGWG chairs

Summary of Resolutions

  1. ACT and Silver agree to hold a day-long set of virtual meetings to begin joint effort sometime in the next couple months
[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.200 (CVS log)
$Date: 2021/04/01 14:44:24 $