18:45:20 <RRSAgent> RRSAgent has joined #aria-at
18:45:20 <RRSAgent> logging to https://www.w3.org/2021/06/03-aria-at-irc
18:45:36 <Zakim> Zakim has joined #aria-at
18:45:45 <Matt_King> zakim, clear agenda
18:45:45 <Zakim> agenda cleared
18:46:12 <Matt_King> rrsagent, make log public
18:46:33 <Matt_King> MEETING: ARIA and Assistive Technologies Community Group
18:46:47 <Matt_King> CHAIR: James Schoels
18:50:11 <Matt_King> present+
18:50:17 <Matt_King> rrsagent, make minutes
18:50:17 <RRSAgent> I have made the request to generate https://www.w3.org/2021/06/03-aria-at-minutes.html Matt_King
18:54:47 <Matt_King> rrsagent, make log public
18:54:56 <Matt_King> rrsagent, make minutes
18:54:56 <RRSAgent> I have made the request to generate https://www.w3.org/2021/06/03-aria-at-minutes.html Matt_King
18:58:54 <michael_fairchild> michael_fairchild has joined #aria-at
19:01:15 <Matt_King> present+ Alysa
19:01:24 <Matt_King> rrsagent, make minutes
19:01:24 <RRSAgent> I have made the request to generate https://www.w3.org/2021/06/03-aria-at-minutes.html Matt_King
19:01:49 <Alyssa_G> Alyssa_G has joined #aria-at
19:02:31 <Matt_King> rrsagent, make log public
19:04:53 <Matt_King> TOPIC: 1. Support command sequences in at-commands.mjs (
19:05:03 <Matt_King> rrsagent, make minutes
19:05:03 <RRSAgent> I have made the request to generate https://www.w3.org/2021/06/03-aria-at-minutes.html Matt_King
19:05:19 <michael_fairchild> scribe: michael_fairchild
19:06:02 <Matt_King> rrsagent, make minutes
19:06:02 <RRSAgent> I have made the request to generate https://www.w3.org/2021/06/03-aria-at-minutes.html Matt_King
19:07:16 <michael_fairchild> JS: This PR is about command sequences. A tester might need to press several commands in sequence to complete a test.
19:07:48 <michael_fairchild> Matt_King: we want to test the output of all of the commands in this sequence. gather the output from all 3 commands.
19:09:15 <michael_fairchild> JS: format in the CSV: multiple commands in command sequence will be delimited by commas
19:10:14 <michael_fairchild> JS: this is presented to testers in the app by separating the commands with the string "followed by". So a user might see/hear the phrase "followed by" many times.
19:11:26 <michael_fairchild> MK: What is the label for the output field? If this is the label for the output field, this should be pretty clear. I like the clarity of "followed by" before every command in a command sequence.
19:11:55 <michael_fairchild> MK: you could shorten it by saying something like "followed by down arrow 3 times", but I'm not sure how easy that might be
19:12:36 <michael_fairchild> JS: so we want the output field to reflect the new wording
19:13:23 <michael_fairchild> JS: is shortening the command sequence so it says something like "followed by down arrow 3 times" a high priority?
19:13:40 <michael_fairchild> MK: No, I don't think so. That case will be fairly rare.
19:13:56 <michael_fairchild> JS: Agree that it will be rare, especially with the setup scripts.
19:15:42 <michael_fairchild> ST: Are there already tests that have optional commands?
19:16:04 <michael_fairchild> JS: Yes, and we want to get to a place where we don't have optional commands.
19:16:55 <michael_fairchild> JS: one concern about readability is that voiceover commands are very long.
19:17:12 <michael_fairchild> MK: saying "then" might be better than "followed by"
19:17:32 <michael_fairchild> JS: we could use a comma to help separate the commands
19:18:43 <michael_fairchild> MK: concern about punctation settings and the tester pressing the comma key
19:22:06 <michael_fairchild> (further discussion about formatting a command sequence)
19:26:47 <michael_fairchild> landed on adding a comma prior to each "then"
19:27:09 <michael_fairchild> JS updated the issue with the feedback from this conversation
19:28:12 <michael_fairchild> topic: Impact of hint text/help messages on test results
19:28:24 <michael_fairchild> https://github.com/w3c/aria-at/issues/365
19:28:34 <michael_fairchild> JS: this was raised again by Rob
19:29:09 <michael_fairchild> JS: The original concern was the VO didn't announce comboboxes as editable, but the hint from VO said to type text
19:29:31 <michael_fairchild> JS: I think where we left this is that we needed more examples and Rob provided them
19:30:14 <michael_fairchild> JS: Radio buttons. When you move to a radio button that isn't checked, it didn't say that it was not checked. But the hint text suggest that you could check the radio button.
19:31:58 <michael_fairchild> MK: Jaws might hint that you can press left or right arrow to hint at horizontal menu bar (vs vertical)
19:32:13 <michael_fairchild> MK: in this case aria-orientation is only affecting the hint text
19:34:35 <michael_fairchild> MK: the question is should an assertion pass if the hint text satisfies it
19:34:47 <michael_fairchild> JS: no
19:35:22 <michael_fairchild> MK: But then you have an issue with automation because you can't distinguish between what is and what isn't a hint.
19:37:05 <michael_fairchild> JS: no because it gives a misleading impression of how well it communicates a role, state, property, etc.
19:38:52 <michael_fairchild> JS: and I think a lot of people turn off hints
19:39:10 <michael_fairchild> MK: I think the most common verbosity level is medium, not beginner
19:42:20 <michael_fairchild> MF: concern about making an exception to testing with default settings
19:43:41 <michael_fairchild> MK: Users shouldn't be able to wish away some aspects of what is announced, such as the role or state. Those should always be conveyed in some way. By turning off hints, you wish away some very important information.
19:45:52 <michael_fairchild> MF: can we ask screen reader devs why some information is only conveyed by hint text? That might reveal helpful information for this conversation.
19:47:09 <michael_fairchild> JS: I agree it would be good to ask vendors about this
19:47:29 <michael_fairchild> MK: another option is that we could mark these as partial support.
19:50:00 <michael_fairchild> JS: it sounds like with JAWS and VO, we will have to capture two pieces of output, because they send the hint text as a second output
19:50:33 <michael_fairchild> MK: you could also run it with the hint text on and then again with it off
19:52:09 <michael_fairchild> Hadi: how is hint text configured in JAWS
19:52:46 <michael_fairchild> MK: it would be complicated for human testers too, and configuring JAWS to turn of hints might be difficult
19:53:11 <michael_fairchild> JS: it sounds like it would be far easier to test without hints, because the difference between them goes away
19:54:01 <michael_fairchild> JS: Unless we want to record if something was supported by hint text
19:55:33 <michael_fairchild> MK: Human judgement combined with automation would be great.
19:56:07 <michael_fairchild> MK: Idea: we could have a button on the page to change the verbosity of the screen reader
21:36:00 <Zakim> Zakim has left #aria-at