IRC log of aria-apg on 2024-03-26
Timestamps are in UTC.
- 18:00:14 [RRSAgent]
- RRSAgent has joined #aria-apg
- 18:00:18 [RRSAgent]
- logging to https://www.w3.org/2024/03/26-aria-apg-irc
- 18:00:21 [Jem]
- Jem has joined #aria-apg
- 18:00:30 [jugglinmike]
- rrsagent, make log public
- 18:00:39 [howard-e]
- howard-e has joined #aria-apg
- 18:00:39 [jongund]
- jongund has joined #aria-apg
- 18:00:44 [jugglinmike]
- Zakim, start the meeting
- 18:00:44 [Zakim]
- RRSAgent, make logs Public
- 18:00:46 [Zakim]
- Meeting: ARIA Authoring Practices Task Force
- 18:00:59 [jugglinmike]
- present+ jugglinmike
- 18:01:05 [jugglinmike]
- present+ howard-e
- 18:01:17 [jugglinmike]
- present+ Jem
- 18:01:47 [jugglinmike]
- scribe+ jugglinmike
- 18:01:59 [jongund]
- jongund has joined #aria-apg
- 18:02:34 [Jem]
- present+
- 18:04:29 [Jem]
- https://github.com/w3c/aria-practices/wiki/March-26%2C-2024-Agenda
- 18:05:02 [Matt_King]
- Matt_King has joined #aria-apg
- 18:05:07 [Matt_King]
- present+
- 18:06:07 [jugglinmike]
- jugglinmike has joined #aria-apg
- 18:06:17 [jugglinmike]
- scribe+ jugglinmike
- 18:06:32 [jugglinmike]
- Topic: Publication status
- 18:06:52 [jugglinmike]
- present+ Matt_King
- 18:06:59 [jugglinmike]
- Matt_King: Current target is April 1 for next publication.
- 18:08:41 [jugglinmike]
- Matt_King: We want to move it to April 8
- 18:08:46 [jugglinmike]
- howard-e: That works for me
- 18:08:54 [jugglinmike]
- Jem: I'm comfortable with that, too
- 18:09:27 [jugglinmike]
- Matt_King: Alright, then. Hearing no objections, we'll move the publication target to April 8
- 18:09:54 [jugglinmike]
- Topic: Combobox select vs focus issues
- 18:10:03 [jugglinmike]
- github: https://github.com/w3c/aria-practices/issues/2962
- 18:10:20 [jugglinmike]
- Matt_King: We discussed this at the end of last week's meeting.
- 18:10:34 [jugglinmike]
- Matt_King: It didn't feel like we got to a place where we made decisions
- 18:12:49 [jugglinmike]
- Matt_King: It still seems like there's room for asynchronous discussion. I'm just not sure that we have a clear definition of the problem.
- 18:13:02 [Jem]
- https://github.com/w3c/aria-practices/issues/2962#issuecomment-2007922864
- 18:13:04 [jugglinmike]
- present+ jongund
- 18:13:23 [jugglinmike]
- jongund: We were concerned about being consistent with standard select boxes
- 18:13:52 [jugglinmike]
- Matt_King: Basically, it does seem like a key question is whether or not selection should follow focus in the list box
- 18:14:13 [jugglinmike]
- Matt_King: In the select-only combobox, the selection does not follow the focus. In the list with auto-complete, the selection DOES follow the focus
- 18:14:51 [jugglinmike]
- Matt_King: That's the primary difference between the two. Is there a strong, good rationale for that?
- 18:15:13 [jugglinmike]
- Matt_King: I can explain the rationale for selection not following focus in the select-only combobox, and I think it is explained in the pattern
- 18:15:52 [jugglinmike]
- jongund: If you are going through using the up and down arrows to use the list, but if you hit the "tab" key, that's where the difference shows up
- 18:17:11 [jugglinmike]
- Matt_King: Reviewing the select-only combobox again, under "keyboard support", we have "listbox popop", and it says "tab sets the value to the focused option (it doesn't say 'selected' option) and then closes the combobox and performs the default action"
- 18:17:29 [jugglinmike]
- Matt_King: We have that behavior even though selection doesn't follow focus
- 18:17:41 [jugglinmike]
- Matt_King: For "down arrow", it just says it moves visual focus
- 18:19:49 [jugglinmike]
- Matt_King: This probably should say that "When it's visually indicated as selected (i.e. with a checkmark) [...]"
- 18:20:24 [jugglinmike]
- Matt_King: Oh, there is a copy-paste error here. It says, "only when an option is referenced by aria-activedescendent [...]" but that is not true. That part of this is truly a bug
- 18:20:52 [jugglinmike]
- Matt_King: That's a very fast-fix bug, it's editorial, and it's important.
- 18:21:21 [jugglinmike]
- Matt_King: I think that the past discussions, we had a lot of people (including Sarah and James) agreeing that tabbing out should set a value
- 18:21:34 [jugglinmike]
- Matt_King: Just like what it says here in the documentation of "tab" in the example.
- 18:21:58 [jugglinmike]
- Matt_King: ...in the table "listbox popup keyboard support"
- 18:22:08 [jugglinmike]
- Matt_King: I know we had strong alignment on that behavior
- 18:22:39 [jugglinmike]
- Matt_King: I think the person who wrote this issue is actually asking for that to be the behavior for the editable combobo with autocomplete
- 18:23:13 [jugglinmike]
- Matt_King: If we were to change the list-autocomplete combobox to use "Selected" the way that select-only combobox does, I wonder what the ramifications would be
- 18:24:44 [jugglinmike]
- Matt_King: In the editable combobox with list autocomplete. The only way that you can get to every state with the arrow keys is if the edit box is empty
- 18:24:54 [jugglinmike]
- Matt_King: If the edit box has a value, then you can't
- 18:25:22 [jugglinmike]
- Matt_King: If I type "a" and choose the first one ("Alabama"), now, there's no way to open the list box and see all 50 states as long as "Alabama" is in the list box
- 18:25:31 [jugglinmike]
- Matt_King: That's a fundamental difference
- 18:26:34 [jugglinmike]
- Matt_King: jongund do you think there should be a command (say, "Alt + Down arrow") that allows you to open the listbox with everything present?
- 18:27:09 [jugglinmike]
- Matt_King: If it didn't filter, though, then that would break people's expectations
- 18:27:55 [jugglinmike]
- Matt_King: aria-selected doesn't seem to play a role in this. If you don't have a selected value, then selection just follows focus
- 18:30:15 [jugglinmike]
- jugglinmike: Maybe we should make the filtering dependent on whether a selection event occurred.
- 18:32:17 [jugglinmike]
- jongund: If aria-selected is following focus, then it doesn't really provide you with any additional information in this example
- 18:33:54 [jugglinmike]
- Matt_King: Yes. I think what jugglinmike and Jem are saying is that you would not filter the list if a selection event occurred and a value had been selected
- 18:34:10 [jugglinmike]
- Jem: Yes
- 18:34:14 [jugglinmike]
- jugglinmike: Yes
- 18:34:44 [jugglinmike]
- Matt_King: if you were to tab into this and "Alabama" was already there, and you pressed "down arrow", you would get all the options and see "checked" next to Alabama
- 18:35:16 [jugglinmike]
- Matt_King: And I suppose that if you do type a character, then the whole value should be selected automatically so that your key presses are replacing the selected value
- 18:38:44 [jugglinmike]
- Matt_King: Is this realistically the behavior that people want? If there is a selected value in the box, then should the value still serve as a filter?
- 18:42:07 [jugglinmike]
- jongund: It seems to me that aria-selected should correspond to some visual indication of a selection. Since in this case, there is no such visual indication, there doesn't seem to be an appropriate use for aria-selected
- 18:42:07 [jugglinmike]
- Matt_King: That's a good point
- 18:44:26 [jugglinmike]
- Matt_King: In the "combobox" section of the ARIA spec, there's nothing in the prose about aria-selected
- 18:44:47 [jugglinmike]
- Matt_King: So what jongund is suggesting would be in-line with the spec
- 18:46:49 [jugglinmike]
- Matt_King: Under "listbox", there's nothing in the prose about it
- 18:47:14 [jugglinmike]
- Matt_King: but it's under "inherited states and properties"
- 18:48:08 [jugglinmike]
- jongund: It looks like the "option" role REQUIRES aria-selected
- 18:50:13 [jugglinmike]
- Matt_King: In the latest version, aria-selected isn't required, it says it's "supported"
- 18:50:41 [jugglinmike]
- Matt_King: We changed this. In the current editor's draft, it's "supported" and not "required". It does not have an implicit value
- 18:51:12 [jugglinmike]
- Matt_King: This is a little tricky, but it's not meant to be!
- 18:52:19 [jugglinmike]
- Matt_King: What jongund suggested (removing aria-selected completely), I think it would be in-line with the other example, as well.
- 18:52:33 [jugglinmike]
- Matt_King: This meeting has turned into a deep-dive into the combobox, hasn't it?
- 18:52:58 [jugglinmike]
- Matt_King: It feels like a little bit of work is needed to prepare a proposal for all the changes that need to happen across the combobox
- 18:53:51 [jugglinmike]
- Matt_King: I think my proposal for at least two of the comboboxes is that we remove any use of aria-selected because (as jongund pointed out) it's not doing anything
- 18:54:26 [jugglinmike]
- Matt_King: I think that would be true in this combobox and the list+autocomplete combobox. I'm not so sure about the other comboboxes, though
- 18:54:57 [jugglinmike]
- Matt_King: I'd like to take the minutes from this meeting and audit the other combobox examples and look at this behavior in all of them
- 18:55:42 [jugglinmike]
- Matt_King: I think we need a complete proposal for the changes this would entail to them all. Though I don't think we necessarily should do that all in a single pull requests (especially considering the corresponding test changes it will likely entail)
- 18:56:18 [Jem]
- https://www.w3.org/TR/wai-aria-1.2/#option
- 18:56:29 [jugglinmike]
- Matt_King: I want to re-read everything that I and others have written about "option"
- 19:00:19 [jugglinmike]
- Zakim, end the meeting
- 19:00:19 [Zakim]
- As of this point the attendees have been jugglinmike, howard-e, Jem, Matt_King, jongund
- 19:00:21 [Zakim]
- RRSAgent, please draft minutes v2
- 19:00:23 [RRSAgent]
- I have made the request to generate https://www.w3.org/2024/03/26-aria-apg-minutes.html Zakim
- 19:00:30 [Zakim]
- I am happy to have been of service, jugglinmike; please remember to excuse RRSAgent. Goodbye
- 19:00:30 [Zakim]
- Zakim has left #aria-apg
- 19:00:30 [jugglinmike]
- RRSAgent, leave
- 19:00:30 [RRSAgent]
- I see no action items