Wilco: very lenthy discussion with Kathy at
TPAC around the implementation specific lenthy procedural steps makes it
difficult for her to contribute as hers dont follow steps
... the way we have steps defined implies that one has to follows the
steps, so what other way do we have to document the procedure
Charu: In IBM we have very atomic rules very short context driven
Wilco: what we are describing is fairly consistent with how DeQue or IBM is doing and is there a way we can have less procedural way of doing this?
<Wilco> https://w3c.github.io/wcag-act/act-rules-format.html
Ann: When we looked at the latest draft we
had difficulty understading the manual path, so what is out and what is
in
... we had a question on 6.2, how updated the draft was
Shadi: In section 6 put it something like
this is how the rule works but the implementation can differ
... one thing that may help, is implementation is open to developers,
can be creative as they want as long as they pass and match the results
... Not all test can be automated, but all should be carried out
manually
... Do not talk about automation or procedure and describe one way of
doing it but folks are free to do it any other way as long as they
achieve the test cases
Wilco: Are you not looking at alternative ways
Shadi: what are alternative ways
Wilco: One can say how it should be done or
what should be done
... Describe in pdf where the elements are rendered and how they are
rendered
... if you look at unit test, usually there is some manupulation of the
application in certain state and write some assertions
Shadi: is that different from test procedures
Wilco: they are more like examples
Shadi: can we have an example
Wilco: we can find one, i think we will
need to do some prototyping to figure this out
... it is different then how i am use to writing rules, let me see if i
can work out an example
<Wilco> Precondition:
<Wilco> For each input element
<Wilco> Expectations:
<Wilco> The element has a label element associated with it using the for attribute OR
<Wilco> The element has an aria-label OR
<Wilco> the element has a title
<shadi> +1
Wilco: this does not prescribe, just tells what is expected put it in sligthly different way
Ann: that would help cuz it does not say it has to be automated but can be manully validated
Shadi: that is fine as they need to be translated in some way
Ann: If you make a rule that is more
specific for manual test, you have to prescribe what to do
... less specific can be implimented manually as well as automated
Shadi: so you agree with what Wilco is suggesting
Cpandhi: why would you not automate and do it manually
Wilco: it is possible that folks may not have tools they trust, some may have AI to figure it out
cpandhi +1
Ann: i will have to check with my developers, we rely on applicability
Wilco: applicability is better then
precondition
... Ok so this may work, we need to try it out to see if it will work
... Ann you want to volunteer?
Ann: i can check with my developer and get back
Wilco: Shadi can i sign you up?
Shadi: you can try
Wilco: Write some of the example rules in the new format
Shadi: not sure i am the right person, i do not have developers to delegate
<Wilco> https://w3c.github.io/wcag-act-rules/
cpandhi: i would like one written and then may be volunteer to write the next
Wilco: Ok i will take one complicated one, r3
Ann: there is user input question, are we getting rid of it or keeping it
Wilco: we will need to try it out
Wilco: Question about pointer, we are
calling it test target
... This might be a question for you, Shadi
Shadi: let me check, so this not about
pointers, it is about any kind of reference in the DOM, that can change
... HTTP states and exchanges can be recorded
... we can use content in RDF
<shadi> https://www.w3.org/WAI/intro/earl
Shadi: The schemas dont have to be in RDF, they can be recorded in diff ways
Wilco: is this all part of the pointer?
Shadi: This can be part of the subject
Wilco: Describe the subject in the same
state and so the pointers are valid
... Ok good answer, i will create a response
Wilco: Not really an ACT question, but more to include the remediation in to the format
Shadi: should we update the format to
include this
... so this is repair guidance, no, this is about reporting results not
about how to fix them
Wilco: In deque we include remediation info
<shadi> https://www.w3.org/TR/EARL10-Schema/#TestResult
<shadi> https://www.w3.org/TR/EARL10-Schema/#info
Shadi: yes we could consider
Ann: it could be a technique for each of them in background
Wilco: Yes
... Shadi, would u like to respond