Accessibility Conformance Testing Teleconference

How to represent manual test steps https://github.com/w3c/wcag-act/issues/104 https://github.com/w3c/wcag-act/issues/112

saz: issue is how to specify tools that support a procedure without hard-coding specific tools into the procedures

anne: tools list?

saz: doesn't currently provide this level of granularity
... years ago we were thinking of mapping tools to WCAG SCs but now we have the rules as well

wilco: is this part of the spec?

saz: not sure

anne: +1

wilco: think maybe not - more process than format

<scribe> ACTION: saz to respond to commenter about "tools support" section in test rules [recorded in http://www.w3.org/2017/10/09-wcag-act-minutes.html#action01]

<Wilco> https://w3c.github.io/wcag-act-rules/rules/ACT-R2.html

wilco: should this be part of the format?

anne: specifies implementation while we agreed the need for more flexibility
... maybe specific more independently

saz: seems like the "user profile" and "context" are the essential attributes for a check
... default assumption is checks are syntax-based and require no context knowledge
... if a check requires semantic verification or context knowledge, then it may require particular input
... not necessarily user input, could also be AI or so

anne: sounds like requirements or properties for the checks

<Wilco> https://w3c.github.io/wcag-act-rules/rules/ACT-R3.html

wilco: if we add properties to checks, would that be a pre-defined list?

saz: rather than "requires sight" and "requires hearing", something more independent like "requires visual verification" or such

anne: could become a long list

wilco: binary?

saz: could be multiple properties

wilco: but each binary or can there be more?

<Wilco> https://w3c.github.io/wcag-act-rules/rules/ACT-R3.html

<Wilco> | User profile | Requires <sight / hearing / fine motor control / HTML Knowledge / Accessibility knowledge / ...> | context | yes | Optional | Interaction | yes | Optional

saz: interaction includes form submissions?

wilco: type of interaction

anne: would be good to be able to distinguish between tests that require scripting vs not

Should we have fewer input types https://github.com/w3c/wcag-act/issues/109

anne: maybe we can write-up something for group discussion
... within next two weeks

wilco: 19th october?

anne: sounds good

Do we need cannot-tell results https://github.com/w3c/wcag-act/issues/69

saz: have not applicable already in EARL
... but not supported by WCAG, so need to be careful
... also think need to keep cannot tell

wilco: yes, comes up fairly frequently
... maybe able to ditch, but do we want to?

anne: to be continued