W3C

– DRAFT –
ARIA Authoring Practices Task Force

07 May 2024

Attendees

Present
arigilmore, howard-e, jugglinmike, Matt_King, siri
Regrets
-
Chair
-
Scribe
jugglinmike

Meeting minutes

ARIA Authoring Practices Task Force Weekly Teleconference

Setup and Review Agenda

https://github.com/w3c/aria-practices/wiki/May-7%2C-2024-Agenda

Matt_King: Any requests for change to agenda?

Matt_King: Hearing none, we'll move forward as planned

Matt_King: No meeting May 14 (it's Access U; I'll be there, and I'll be trying to recruit for this group)

Matt_King: Next meeting: May 21

Thank you Andrea

Matt_King: Andrea Cardona has accepted a new position outside IBM. She says good bye. We say Thank you ... and please come back if you can!

arigilmore: Her last day was about two weeks ago

Jem: There is an "invited expert" system to allow people to stay in the group if they change their employer

Publication status

Matt_King: due to Access U, I decided we'd move the publication from today to the 21st of May

Matt_King: I thought keeping it today would just be too much stress

Matt_King: If you look at this milestone, none of the pull requests in it are ready, but several close

https://github.com/w3c/aria-practices/milestone/31

daniel-montalvo: I'm not aware of anything that will keep us from publishing on the 21st, but I'll let you know if anything comes up

Feed example update

howard-e: pull request #2775 wasn't working because the pull request was out of sync with the repository which generates the preview

howard-e: To resolve it, I went to the generated pull request and just updated it with the main branch of the "build" repository and re-ran the relevant GitHub Action

howard-e: I think this will continue to happen with some of the older pull requests. We should probably document the process of manual updating so that it's not only me who can correct this problem

Matt_King: That's some great news that it's in sync. I tested this, and the "CTRL+End" thing in the preview is working great

Matt_King: I think arigilmore's work on this is done and that this is practically "ship ready"

arigilmore: Thanks to the team for all the help. I'm really glad that this will be in the next release

Matt_King: I can't believe this one took so long!

Update to AT Support tables

preview link: https://deploy-preview-317--aria-practices.netlify.app/aria/apg/patterns/radio/examples/radio-activedescendant/

Matt_King: five changes have been made based on feedback from this Task Force and the ARIA-AT community group

Matt_King: that includes

Matt_King: a column header in the first column

Matt_King: a new title for the second and third columns

Matt_King: removed the word "supported" from every cell for a cleaner presentation

Matt_King: We changed the order of rows to alphabetical (thanks to howard-e and team--it looks like that got fixed this morning)

Matt_King: The only thing left to do is to finish the content which explains the meaning of the rows and columns

Matt_King: We'll probably have to work asynchronously on that to finish in time for the 21st of this month

Matt_King: Just like we discussed last week, we'll add the link to just below the heading that's above the table

Jem: Looks good to me, Matt_King. Great job!

daniel-montalvo: This "must have" versus "should have" may be problematic. This assumes the spec language. We should try to explain where these are coming from.

daniel-montalvo: I've been having a look at the AT documentation. I think if you go down that path, then you can understand.

daniel-montalvo: I still kind of prefer using other language

Matt_King: We're intentionally using RFC 2119

Matt_King: the idea is that, in the long term, by aligning the terminology, we might be able to one day reverse-engineer a spec from the tests

daniel-montalvo: I'm glad to learn that there has been agreement between the main stakeholders

daniel-montalvo: I would appreciate this to be in the "about" page

Matt_King: Agreed. It will definitely be there

Jem: "A" sounds awesome

<Jem> s/Test-driver spec/AT test driven/

Support for experimental content

Matt_King: We have this "preview" pull request for the experimental content

w3c/aria-practices#2977

Matt_King: That's based on a fake experimental copy of the "tree view" example

Matt_King: I'm wondering whether or not we would be able to merge all of these changes except for the fake preview..

Matt_King: ...so that we could proceed without changing anything visible in the APG, but that it would allow us the open a new pull request which could introduce experimental content

Matt_King: I'm wondering if we could get these changes in place without the experimental content just so that the infrastructure is available

Matt_King: I expect that the first experimental content for APG could take multiple weeks if not months

howard-e: I think that would be definitely possible

howard-e: It relates to a question I had in the pull request review

howard-e: If there is no experimental content, then that section at the bottom of the page is just empty

Matt_King: Could we change the script to only include that section if there is at least one piece of experimental content?

howard-e: Yes, that should be possible

Matt_King: If we move forward with this, that will allow us to begin moving forward with ARIA Actions

Matt_King: as far as I know, I think ARIA Actions is the most important experimental content coming up

Matt_King: Other things in ARIA 1.3 (e.g. annotations) come to mind, but we don't have enough people to work on that stuff right now, I think

New and unlabeled issues

https://github.com/w3c/aria-practices/issues?q=is%3Aissue+is%3Aopen+created%3A%3E2021-08-15+no%3Alabel++sort%3Aupdated-desc

How to properly implement typical e-commerce comboboxes

github: w3c/aria-practices#2999

Matt_King: I think this person is asking whether the elements in a dropdown list of a combobox can be links or work like links

Matt_King: In the example website, it looked to me like the combobox they have for their search type-ahead is working really nicely

Matt_King: If you follow that link, and you go to the search button, it opens a dialog which is just a combobox and a search button

Matt_King: I used the term which they suggested, "bags"

Matt_King: In the dropdown for the combobox, are those styled as links visually? Do they work like links for other people? They don't seem to behave like links for me; they seem to just execute a search

Matt_King: I don't know if maybe this is something that is non-obvious to a screen reader user

jugglinmike: The items function like links to me--they cause a navigation

Matt_King: Yeah, it works that way for me, too, but the screen reader doesn't tell me it's a link. I wonder if that's what people are concerned about

Matt_King: You can't open those links in a new tab as a keyboard user because the focus is on the search box

jugglinmike: I can open the links in a new tab as a mouse user by clicking with the middle button

Matt_King: I don't think there's a problem with this implementation as far as I can tell.

Matt_King: The only problem with the options not being links is that they don't have all the functionality of links. The dropdown certainly works as I would expect it to work

Matt_King: It's not clear to me why this is a special pattern

Matt_King: I don't know if using a grid in the drop-down (so that the link could be exposed as a link) is an improvement. I don't know if the dropdown items need to be exposed as links

Matt_King: Though I haven't tried this with multiple screen readers...

Matt_King: This looks just like APG's listboxes. It feels fine to me

Jem: I don't see any listbox role in this design

Matt_King: Somehow JAWS is going into forms mode, though

Jem: Ah, I see the "listbox" role, now

Jem: the role "option" is on the <li> element

Matt_King: the option is a child of group, which is allowed

Matt_King: There is a heading inside of a listbox, which is a validation issue

Matt_King: I don't know if this is a special pattern. It looks to me like this is the listbox with grouped options, and the options behave as links. There isn't a rule against options behaving as links as far as I know

Matt_King: Any executable elements is allowed to execute a command. One such command is "open this page" which is what a link does

siri: how does a user know that an item is a link if it isn't exposed as a link?

Matt_King: Well, what else would the user expect the items in a list of "search results" to do?

Minutes manually created (not a transcript), formatted by scribe.perl version 221 (Fri Jul 21 14:01:30 2023 UTC).

Diagnostics

Succeeded: s/comes uo/comes up/

Succeeded: s/Test-driver spec/A

Failed: s/Test-driver spec/AT test driven/

Maybe present: daniel-montalvo, Jem

All speakers: arigilmore, daniel-montalvo, howard-e, Jem, jugglinmike, Matt_King, siri

Active on IRC: arigilmore, howard-e, Jem, jugglinmike, Matt_King