Meeting minutes
<jesdaigle> +present
App development workstream
Seth: For the aria-at app, we have been working on requirements. We are starting to have conversations with James and Sina. We will start our bi-weekly aria-at app sub-meeting starting next week. Thursday 4/1 at 2pm ET.
Matt: what zoom should we use
Seth: can we use this zoom?
Michael: yes
Matt: we recently had some discussion on the working mode. If this will affect our discussions next week, should we spend a little more time on it this week?
James: we decided on 'draft', 'wide review', and 'finalized'.
James:
James: 'Wide review' is when we feel that the test is ready for review outside of the CG, for example by AT developers. Once it gets to this status, there would be 60 days for feedback, and another 60 days for us to fix any issues.
Matt: thank you. We will want to add the ability to track the status of the test into the data model. Seth, should we talk more about this today?
Seth: I'd love to talk about this more.
Automation workstream
Seth: There are a number of shared tasks between automation and AT-app, including the hybrid test format.
Seth: that's the main thing happening right now
Seth: we will start the engineering in late April
Seth: week of the April 26th
Matt: Seth, can you coordinate some issue discussions (test format improvement and supporting automated tests from the same format)?
Seth: yes
Test development
James: We have been working through the 8 patterns.
James: potential blockers: we are waiting on some APG example updates for sliders, and file tree-view. We are also waiting on the toolbar example.
Matt: as far as the APG, the updates for the sliders will be landing soon, including a multi thumb slider.
James: should we add the multi thumb slider to our list of test plans?
Matt: yes, especially because the tree view example might take some time.
James: for the sliders that are using aria-orientation, where are we at on that?
Matt: we had a discussion about labels, and decided to add role=none on the SVG element
Jon: and role=slider is on an g element inside the SVG with role=none
Matt: for toolbar, I have a question for Seth. What month of the year do you think we will have the ability to include tests that are not in the APG, such as atomic ARIA or atomic HTML tests?
Seth: It's mainly just a matter of making sure that the scripts to pull in example are generic enough. Should we prioritize this?
Matt: toolbar might require this.
Matt: at some point, we might decide that the APG tests don't actually belong in the APG repo
Matt: so I think the base capability here is making where we pull tests from configurable
Seth: there is additional engineering work, so we would want to have some conversations about this but it should be driven by a test writing need
Matt: there is a need
James: from a test writing perspective, this is currently possible, but perhaps not ideal.
Matt: we could come up with a directory naming structure for our tests
James: we could make this as simple as possible. In references.csv, we have two rows for design pattern and example. We could update that format to define a source of something like 'apg' or 'other'. Or if there is no design pattern link, then it is not an APG example.
Matt: when we get to toolbar, maybe that's when we implement that iterative approach
Matt: but I do see this rippling through everything, even to the reporting
Seth: I think there is a way to sneak them in right now, then there next step is to update our metadata fields are more generic, then we can update the UI
Matt: I have merged the stuff that I can merge, but some PRs have conflicts in the keys file
Matt: those are real conflicts that actually need to be resolved
(further discussion on resolving conflicts)
Running and reviewing current tests
Matt: now that we have merged these tests, we want the CG to run these tests and provide feedback.
Matt: so this is an opportunity for people to do testing. Should we formalize this process, or should we just ask people to pick their own tests and run them?
Matt: if it's helpful, we could formalize the process
Michael: I think formalizing the process would be good
Jon: I do encourage more people to get involved. If there are assignments, it might help people get involved.
Matt: Should we send an email to the list and recruit people for specific tests?
Rob: Should I test with all of the screen readers, or is testing with only one okay?
Matt: we are happy with whatever you can do
Matt: it would be good if we could have someone help coordinate this
Michael: should we discuss this on our next call?
Matt: yes, James is that okay?
James: yes