Meeting minutes
Setup and Review Agenda
Matt_King: Next meeting: March 18
https://
Matt_King: Any requests for change to agenda?
jongund: When will the color-usage work return to the agenda?
Matt_King: Soon--after we get the next publication out
Publication planning
Matt_King: Remi found a mistake in the AT support tables. I guess we missed one where it is not sharing the data--link example 3.
Matt_King: So he did not publish.
Matt_King: I'll fix that (it should be an easy fix), then it will be ready to go
Daniel: Can you mention me on the relevant "ARIA practices" pull request?
Matt_King: Sure
Matt_King: For support tables, the source URL is based on the name of the directory for the repository in ARIA-AT
Daniel: There's something on the pull request that we're talking about changing with the dialog example
Matt_King: I'll get that merged at the same time and include it
Matt_King: Once I've done that, should I add a comment to the publication pull request and then mention both you and Remi
Daniel: Yes, that's perfect
Matt_King: Great. It'll be good to have that done before CSUN
PR 3400: aria-actions example
github: w3c/
Matt_King: This is what CurtBellew is working on. We added it to the agenda at his request
CurtBellew: I implemented the suggestions that we had in 3400: the changes in the list itself and an alteration to make it look a little more like GMail
Matt_King: I wonder if we might want to consider putting the thing that says "content goes here" in a read-only edit (like we do in the color picker and file-tree views). Then it can be focusable. Then, for screen reader users, if you tab out, you could go there
Matt_King: I'm torn on that, though because in real life, it wouldn't be that way. But it could be!
Matt_King: We could leave the heading the way you have it, where it says "details" with a level-3...
CurtBellew: I could look into that. I have to do some other changes, anyway--there are some CSS issues here
CurtBellew: I did struggle with that issue: how do you make it clear that this is a master detail within the list box?
Matt_King: If we are making it look like GMail, then does GMail update its panel automatically on focus? I don't know because I always have to press "enter"
Matt_King: On the other hand, there are plenty of user interfaces like this where the right-hand panel updates automatically as you scroll through, so I could go either way
Matt_King: We have multi-select list boxes where selection doesn't follow focus. But I don't think we have a single-select where selection doesn't follow focus. And the fact that we don't have that might be enough justification to demonstrate it here
Matt_King: If we had selection not following focus, it would work a little more like example 2. But example 2 is a multi-select. So I think we probably would make it so both "enter" and "space" select.
CurtBellew: This sounds good to me. It was one of the things I thought was weird when I was testing it
Matt_King: Okay, then we could have "enter" and "space" do the selecting, and when you select one, it de-selects the other
Matt_King: So it's a little like tabs with manual activation in that way
jongund: Is there a good use case for selection not following focus?
Matt_King: Yeah, it's like the GMail example where you have a list with a separate panel where you open stuff
jongund: Okay. I guess there's cases where you don't want selection to follow focus
Siri: I was playing with GMail just now and seeing this
Siri: In Gmail where there is a checkbox with a star next to it. We have the option to select the star separately. Maybe removing the option to select the star would be good.
Matt_King: Star and checkbox mean two different things
CurtBellew: To set the star, you hit the action button, and that will change the state of the star on the row. You can use a mouse to change the state directly, but with keyboard, you need the action
CurtBellew: And that was kind of confusing. It looked like you can click it
Matt_King: Do we want to show a persistent empty star?
Matt_King: What if the persistent empty star was the same thing that got keyboard focus? That might be a little weird because then focus would jump, though. Was that the concern?
CurtBellew: No, it was more that we move from the row that has focus to the action buttons. I hadn't thought about that. You would focus on the row, then arrow key would take you to the star on the row, and then another arrow key would take you to the action buttons...
CurtBellew: The empty star is to the left of the checkbox
CurtBellew: I'm assuming that the star button would still want to be an action
Matt_King: Yeah
Daniel: Actually, the favorite/star is the second thing when you're on the row
Daniel: If you add this in between the row and the actions, you'd have it twice
Matt_King: We're saying that we have a star in two different places when the row has focus
Matt_King: The second is click-able with the mouse but not accessible with a keyboard
Matt_King: If we removed the current "action" button and made the current icon an action button, it would not be part of the same group of buttons with "move up", "move down", and "delete"... But it wouldn't have to be
CurtBellew: Okay, I can explore that. It makes a lot more sense. I had some internal battles with having those two things clickable
Matt_King: I like this idea. Good call-out, Siri!
Matt_King: Mouse users would still have an additional capability, they could click the star without focusing on the row, but that doesn't matter
Matt_King: So we aligned on selection not following focus, "enter" and "space" following select (and that would change the display of the right panel), and then exploring how we can have only the "star" icon being present and that being an action item
CurtBellew: Then there's a third thing: read-only edit
Matt_King: Ah, right. That, too
Matt_King: Sounds great!
PR 3416: Add MCP server
github: w3c/
Matt_King: I didn't know that you could actually put something into the repo that would enable that repo to be an MCP source. This seems super-cool to me on the surface (if I understand the surface correctly)
Matt_King: We need some people who have actually done this sort of work to look at this and provide feedback. That wouldn't be me!
Matt_King: I would love to use this, and I would love it if a lot of people were using this. That would be great
Matt_King: Has anyone here done any work related to developing MCP sources?
Matt_King: Hearing none, should we take it to the ARIA working group?
Matt_King: Just to get a wider audience of people who could review and test this
Daniel: There's no official stance from the W3C on this sort of thing. I do know that some people in WAI have expressed opinions. I've used tools like this myself, but it's one thing to use them and another thing to promote their use.
Matt_King: I would assume that the goal is to assist people who are building components--to allow them to instruct their agents to use the guidance from the APG
Daniel: In that case, then my comments wouldn't apply as much
Matt_King: One of the problems with pointing an agent at the APG right now is that not all of the content in APG is necessarily relevant to building a specific pattern. My understanding is that this would help agents be smarter by focusing their attention on the relevant information. I could be wrong about that!
Daniel: I'll take a look at this, and I'll bring it to WAI (just to make sure we're in alignment for this type of change)
Matt_King: Sounds good
Matt_King: If we can support something that makes APG more useful to agents, I think that would be awesome. But it may be a bigger lift (e.g. maybe it has an impact on the W3 org), so I'm interested in hearing what you and Remi have to say, Daniel
PR 3387: Clarify guidance for Focusability of disabled controls
github: w3c/
Matt_King: Adam_Page couldn't make it today
Matt_King: I've added some feedback
Daniel: I've written up some suggestions, but I haven't been able to post them yet due to technical difficulties with the GitHub UI
Matt_King: I wonder if your comments are similar to my comments. I added a comment rather than trying to mess with the "diff" viewer itself. I quoted the text that I think should change and I added my suggestions there
Matt_King: (I'm finding the new "diff" viewer for GitHub pull requests almost impossible to use.)
Matt_King: It looks like I'm the only person who has been formally designated as a reviewer. If anyone else wants to add feedback, now is the time!
Matt_King: But if Daniel is also reviewing, I think that will be sufficient because this is strictly an editorial change
PR 3410: Use F2 to edit a grid cell
github: w3c/
Matt_King: There is a modification to the package.json file in this pull request
Matt_King: I'm not sure we want to accept any pull requests that fix examples and which also modify package.json
jongund: The "serve" module they added is related to serving static site content and listing directories
Matt_King: I'd like to see the change to "package.json" removed
jugglinmike: We might also encourage the author to remove the irrelevant whitespace changes since they add a lot of noise to the patch
jongund: The JavaScript is included inline in the HTML document itself. We'd want it to be in a separate JavaScript file
Matt_King: Should the JavaScript for a "popup" dialog be in its own file? Rather than mixed up with the existing JavaScript file?