Meeting minutes
Status of current draft review testing
James_Scholes: Let's start with JAWS and Firefox, Color viewer slider
James_Scholes: Hadi and Isa are assigned. Isa's testing is complete. Hadi has not started, but we had a testing session with him on Monday
James_Scholes: NVDA and Firefox, Color viewer slider
James_Scholes: Dylan and Isa are testing. There are three conflicts
James_Scholes: VoiceOver with Chrome, Color viewer slider
James_Scholes: Dylan is assigned, but hasn't started yet. Isa's testing is complete
Hadi_Rangin: I will complete testing for Color Viewer Slider with JAWS and Firefox by the end of the week
Matt_King: It seems like we might be able to wrap up color viewer slider within the next week
James_Scholes: Moving on to Navigation menu button
James_Scholes: Isa is the only person assigned for JAWS and Firefox. Those tests are complete
howard-e: Louis at Bocoup will be off next week, but will be available the following week
James_Scholes: The same situation for other AT/Browser combos
James_Scholes: Modal dialog
James_Scholes: JAWS and Firefox has one conflict
James_Scholes: NVDA and Firefox have no conflicts
James_Scholes: VoiceOver has only one Tester--Isa. She has completed testing. We'll need another tester
Murray_Moss: I'd like to volunteer for VoiceOver testing
Matt_King: Thank you! James_Scholes will get you set up as a Tester in ARIA-AT App
Progressing candidate test plans to recommended phase
Matt_King: In the Working Mode (that's the business process of the Community Group), test plans move from "draft review" to "candidate review", and once we feel we have sufficient consensus from the active AT implementers, we move them into the "recommended" phase
Matt_King: Last week, we were talking about the process of moving from "Candidate" to "Recommended" in the "sandbox" server
James_Scholes: Isa is putting together a report about what high-level issues we would need to raise with the vendors. We hope to have the complete by either tomorrow or Friday
Triage automation issues
Should one tester be able to assign verdicts to AT responses collected by another? · Issue #936 · w3c/aria-at
jugglinmike: These three issues are related to latent capabilities that automation is surfacing. Up until now, testing has included two processes, collecting test data, and assigning verdicts.
jugglinmike: We are still defining what assigning a verdict means
These questions also help us define if we can think of automation as another user in the system.
MK: So then there are two roles, one to collect responses, and one to be a response analyzer
We need to also define the skill set and experience required for each role
Link to GitHub Issue for this question: w3c/
jugglinmike: In an earlier meeting this week, James pointed out that currently we have some implicit requirements for testers (having an AT installed and be a daily user of it) that prove they have the experience required to run tests
MK: What are the dependencies of making these decisions? Will the answers effect the data architecture required?
jugglinmike: If the answer to any of these questions is yes, then yes they are critical. This is about being proactive and not painting ourselves into a corner
These are the additional questions Mike Raised: w3c/
jugglinmike: Answering the questions would help define the working mode and help us decide the kinds of tasks we want to work on developing.
jugglinmike: The working mode would then be more substantiated by answering these questions
MK: Can we prioritize one of these questions and get everyones feedback?
jugglinmike: The third question depends on the first. The second is dependent
jugglinmike: The idea would be the more we decouple these two tasks of collecting responses and assigning verdicts, the more potential there is to flush out conflicts.
jugglinmike: We need to make sure we are getting the same result with the AT, before we decide how to split up assigning verdicts
MK: I'm wondering about a way forward. Should we discuss 938 first?
jugglinmike: Maybe, its hard to take these apart. Its really about how much we want to commit to treating these as separate functions
MK: I dont see a world where we don't have a human collect a result for each test.
<garcialo> I've got to head out. Go ahead and let me know next steps and whatnot with the test stuff.
MK: I will work to add answers to these questions, James and Isa should work on responses to these as well
jugglinmike: We also need to figure out if we need to address automation in the working mode
MK: Yes lets discuss on Monday