Meeting minutes
TPAC report
<michael_fairchild> MK: wanted to see if we can use ARIA-AT as part of building the ARIA spec
<michael_fairchild> MK: We didn't get very far into that discussion, but there was interest. But there wasn't a lot of deep discussion.
<michael_fairchild> Seth: Also with the ARIA WG meeting, there was a lot of curiosity and questions about ARIA-AT, especially around scope.
ARIA WG Presentation: https://
<michael_fairchild> MK: Yes, lots of questions around iOS and Android
<michael_fairchild> MK: Seth went to the Browser Testing and Tools WG. The goal there was to find the right home for the automation API.
Browser, Testing, and Tools WG Notes: https://
<michael_fairchild> Seth: The question: is ARIA the right WG to host work on this API, or is a better place the Browser Testing and Tools WG? We gave an overview of the project including progress and goals, and floated the idea of them hosting the API.
<michael_fairchild> Seth: the response was very positive and the chair for the group was very open to continuing the discussion, and seemed optimistic.
<michael_fairchild> MK: There is also some questions about working group charters and where the automation fits best into which charter
App Development Update: Candidate Review queue
<michael_fairchild> MK: we are getting very close to another deployment
<michael_fairchild> Seth: yes, by next sprint we will have more to share about the deployment. We are working on candidate test plans, and allow vendors to log in and have the ability to review the candidate test plans. They can step through the tests and results, see related github issues. The test plan admin can see which reviews are complete.
<michael_fairchild> Seth: We think we can deploy close to the end of the next sprint.
<michael_fairchild> Seth: but no exact date right now
App design discussion: Test Plan Management
<michael_fairchild> MK: We will now have the ability to see when a plan is ready to go from candidate to recommended. But we won't yet be able to mark plans as recommended.
<michael_fairchild> Seth: We need to agree on what it means for something to be recommended.
<michael_fairchild> Seth: is the goal to recommend the test plan pattern itself rather than the plan applied to 3 different ATs?
<michael_fairchild> MK: the 3 AT scope is part of the recommendation. If we bring in a new AT, we would need a new version of the plan, which would need to start fresh for the new AT. But the tests for the existing AT should be fine as is.
<michael_fairchild> Seth: the top level change here, is that we need a way to set all of the material for a pattern to be recommended. We don't want to be in a place where combo box JAWS is recommended, but combo box Narrator is not.
<michael_fairchild> Seth: I want to discuss adding a future AT and I want to discuss where the button to change it to recommended exists in the UI.
s3ththompson: so we think we need a new page in the app called something like "Data Management" where you would be able to browse a table of all the data in the app. the test queue isn't sufficient because it's a dashboard of active items only, and it's actually showing Test Plan / AT / Browser (not just Test Plan). the candidate review queue isn't sufficient because it's also a dashbaord and it's showing Test Plan / AT...
s3ththompson: so we really need a new page that can slice data as we see fit, show a paginated archive, and expose raw data (e.g. all results ever recorded) for posterity
s3ththompson: there are some other conversations to be had down the road, e.g. What happens after we recommend Combobox but we add a new AT, like Narrator?
Matt_King: I could see that having to do with the discussion of versioning
s3ththompson: yes, potentially it's a question of versioning, potentially it's a question of having a better sense of expanding scope / coverage over time
Allow for more Test Commands in Test Format
s3ththompson: James is working on test plan for disclosure that has 7 keyboard commands. for arbitrary historical reasons, the CSV test format only supports 6 commands (listed by name `commandA` through `commandF`).
s3ththompson: I talked with howard, we should be able to expand in a backwards compatible way. it would be great to support up to 26, but not require test files to be padded with all those empty cells. I think we will be able to get to this next week
James Scholes: great, thanks. I will proceed with writing the test as if the parser supported 7 commands. we won't be able to preview the last command yet, but we can regenerate tests once that change lands
Matt_King: right, and this is all just our own parser, so if we needed to radically change something for a new AT down the line we always could
Testing Update
James: disclosure menu test plan nearly complete. Updates were a bit more complex than anticipated.
All current testing is on track. No new test assignments areneeded at this time.
Will be in contact with Lewis regarding the latest test plans.
No new feedback from Apple or Vispero.