– DRAFT –
Accessibility Conformance Testing Teleconference

ACT Standup

<ChrisLoiselle> :)

<Wilco> act-rules/act-rules.github.io#2067

wilco: feedback from AGWG, set up a new survey, help Chris with implementation

<Wilco> So this: https://www.w3.org/WAI/standards-guidelines/act/rules/5f99a7/ now redirects to ./proposed/

daniel: reviewed and approved PR on aria-required ID, proposed tools, work to publish ACT Rules format

kathy: github guidance, label in name Underdstanding PR

catherine: my re-review responses not showing up for SVG element,

suji: re-review of rules

chris: Oracle implementation, re-review completed, Heading has non-empty accname, reviewed label in name PR

helen: table header cell PR 1971 pending work from aaron, PR 2064 for audio element feedback from jean yves

tom: iframe PR has enough reviewers, pass examples for aria and html disagreements in 2 aria rules

Decide if we want to continue using "test cases" as a term, along with example

wilco: rules format today uses term "test cases" but then started using "examples" when writing rules
… rules now have a mix of terms interchangeably. Can we be consistent and pick one?

suji and catherine: like "examples"

chris: either, no preference

helen: depends on context

tom: examples are not how to implement rule. they are examples of what you might test with the rule.
… ok with examples

wilco: any objection to leave as is?

chris: if oracle has a bug, we ask for expected vs actual. it's a use case or test case for problem.
… for educational, I can see using example

wilco: I think we should pick one term and be consistent

tom: in API we use test case, harder to change

wilco: it's quite a bit of work to change in API from test case to example

<Wilco> https://www.w3.org/TR/act-rules-format/#test-cases

daniel: test case is most accurate technical term but can explain or define it more clearly

<ChrisLoiselle> https://www.functionize.com/blog/a-glossary-of-testing-terms , perhaps test case and test script

<ChrisLoiselle> or test plan

helen: example is quite general and can be anything where test case is testing something

wilco: if we use test cases, what do we call pass examples?

helen: passing test case 1

daniel: pass test case is long

helen: heading is test cases so subheadings are examples of test cases

<ChrisLoiselle> test data being what the "example" is? I defer to group :)

wilco: continue as is

<Wilco> draft RESOLUTION: Continue using "test cases" as the main heading, and "examples" as sub-headings

RESOLUTION: Continue using "test cases" as the main heading, and "examples" as sub-headings

wilco: update the rules format to follow modern style of writing rules

Subjective exceptions in the applicability

<Wilco> https://github.com/act-rules/act-rules.github.io/discussions/2061

wilco: had this conversation last week, opened a discussion about subjective applicability
… losing objective applicability allows pretty generic rules
… the generic rule is the SC rewritten in 2 parts
… generic rule does not improve SC and we wanted to avoid that problem

tom: challenges of SC

kathy: agree subjective applicability is not helpful to testing

chris: more strict tests would be better

helen: subjectivity is always going to be present as long as we define anything that can be open to interpretation

wilco: applicability now is an element
… there is a gap. example is pie chart can't be targeted explicitly
… for many elements, WCAG requires semantics but not role=img on an img
… can write a rule for css to have role=img
… objective applicability prevents writing of some rules for wcag

tom: can be objective but not programmatically objective
… a human can make an objective decision with information

wilco: tried to deal with ambiguity by listing out items of interest
… what about splitting out non-text content that has semantics of an image and a separate rule for non-text content without semantics of an image

chris: wcag3 working on an approach

wilco: separate elements with semantics from elements that don't

daniel: may need to try this out

kathy: for non-text content example, there's many types of non-text content so one rule for all of it doesn't work. maybe a rule for each type of non-text content?

wilco draws.

wilco: one circle "semantics", one circle "function". they partially overlap.
… example of heading: overlap is semantic heading functions as a heading
… subjective applicability functional heading, that looks like a heading
… has semantics of a heading

chris: if a heading has aria level 9, is this semantic or functional?

wilco: semantic
… circles drawing works when semantics are required but not for image which doesn't require semantic, but requires accname
… draws a rectangle that overlaps both functional and semantics circles
… rectangle "accessible name"
… rectangle overlaps intersection of functional and semantics circles
… currently don't have a way to test things that function as an image to have an accname

<ChrisLoiselle> Thanks!