Meeting minutes
<Jem> https://
Setup and Review Agenda
APG Issue Triage
<Jem> https://
Jem: how best to triage this issue?
jamesn: i don't think this is an APG issue, probably a VO issue. only wanted it here for tracking. filed with apple via feedback assistant
Matt_King: did you get any link from apple?
jamesn: not with feedback assistant
Matt_King: we have labels for AT. if we say this isn't for practices... it'd need to be added to the slider project so people know
jamesn: most people seem to think it's a VO issue
Matt_King: i've experienced this, I assumed it wasn't our code, but I was trying with safari so thought maybe it was something unique to webkit. is thre anything about webkit that's different?
jamesn: reproduces in chrome too
jamesn: when VO isn't running, the page doesn't scroll. so to me it seems like VO is doing something
Matt_King: should we close this then, since it's not for APG?
jamesn: fair enough. if I hear back from apple i'll let everyone know
Matt_King: this will certainly affect aria-at
<Jem> https://
jamesn: I don't want to escalate as it doesn't reproduce on input type=range
Matt_King: 2040 --
Matt_King: sounds like something to assign to infrastructure?
Jem: yeah
jongunderson: this can be closed
<Jem> next issue is https://
Matt_King: added infrastructure label, should be part of the infrastructure project
Matt_King: maybe 2040 didn't get done because it wasn't in the project? it's assigned to Jes
Jem: so this one's about how are we going to best indicate when an exmaple was last updated?
jongunderson: some of our examples are 5-6 years old, and it's hard to tell what's using our most up to date code or not
Matt_King: this should be transferred to apg-redesign
Matt_King: i think i got it transferred
Jem: looks good, next item is..
<Jem> https://
<Jem> https://
MarkMcCarthy: If I'm understanding right, I can't reproduce the issue
jamesn: me either
MarkMcCarthy: I'm using Firefox on Windows
jamesn: i'm using a Mac
jamesn: oh, using Edge on Mac, if I Esc out of the context menu, my selection becomes the thing I right clicked
MarkMcCarthy: oh I see it now too, doing it that way
jamesn: i'm not sure it's a WCAG violation though
Matt_King: but the bug report says context menu doesn't appear, AND selection changes, but you two see a context menu
jamesn: is expected behavior that right click doesn't select? Selecting text, right click, text remains selected (on github for instance)
Matt_King: tempted to say we can't reproduce the problem as described, might need more details
Matt_King: seems it's working as intended
jamesn: it's not a change of context, really, right? maybe kinda... hmm
jamesn: maybe it'd just -work better- if it didn't do that
MarkMcCarthy: i'm inclined to agree
jamesn: i don't think it's enough to disorient a user
MarkMcCarthy: right
jamesn: pointer cancellation though... <skimming the SC> I don't think it's that either
<CurtBellew> Something has come up and I need to drop off.
jamesn: maybe a bug but not a WCAG violation
jamesn: maybe we should look at native date pickers
jongunderson: maybe we ignore the right click then?
jamesn: native picker in edge ignores the right click
MarkMcCarthy: native Windows date picker (e.g. Word) seems to ignore right click also
Jem: so maybe assign to jongunderson and he can explore from there?
Jem: cool. next issue is..
<Jem> https://
Matt_King: we do have a generic show/hide, it's the figure caption one
<Jem> https://
Matt_King: let's ask if this meets the needs described
Jem: got this mentioned and linked on github
Tab PR
<jongunderson> https://
Matt_King: what's the linked issue?
jongunderson: 2044
<Jem> https://
[general group grumbling about github's recent changes]
MarkMcCarthy: so what do we need for this PR?
Matt_King: is this only editorial?
jamesn: in my opinion yes
Matt_King: i'll merge this after the meeting, all set
jongunderson: been doing some testing with the right click thing while chatting, it's not detectible with a regular click event
jongunderson: it'd complicate the code to add more to listen for that right click
Matt_King: can you add a comment explaining that particular bit? we'll use that as our decision.
MarkMcCarthy: i agree, not in favor of adding complexity for something so small
jamesn: +1
jongunderson: not to mention it'd affect almost every example
Accordion Example
https://
<Jem> https://
<Jem> "Should tabbing/using arrow keys skip over disabled headings? The example uses aria-disabled instead of the disabled attribute for the button, though it appears to me the two accomplish the same goal. I imagine for illustrative purposes, the former was used. Nevertheless, using the disabled attribute disallows focusing, so I'm wondering if the example should be updated for consistency?"
Matt_King: why would we NOT want focus to go there? updated for consistency with what?
Matt_King: did we actually take away arrow key behavior?
<Jem> https://
siri: i'm not seeing a disabled attribute here
Matt_King: OHHH we updated the example. THIS one doesn't have what the issue is describing
Matt_King: do we have another? besides Sarah's PR?
<Jem> https://
Matt_King: is it okay with the group to say we did do as it's asking, I think Sarah's example includes this? any reason we would NOT want the heading in the tab order?
<Jem> we have two accordion examples in progress.
siri: well if it's disabled it should not have focus
Matt_King: in accordions, there's an element that visually appears like an icon/button/etc. in addition to the heading text
MarkMcCarthy: there's a little arrow (and +/- on Sarah's example) showing if something is open or closed, but there's no button or anything
Matt_King: maybe lets look at Sarah's PRs
Jem: this is the more complex one, has a button wrapped with an h3
<siri> https://
jongunderson: aria-disabled is on the button, it's true when expanded
Matt_King: does it look any differnt visually?
jongunderson: no
<siri> yes
Matt_King: so siri's point that it's a little confusing that you can tab to it and not doing anything is kinda confusing
bryan: i've never liked disabling an accordion button like this, vs. one that you're actually not supposed to interact with
bryan: yeah, there are some accordions where you can expand and collapse everything, and others where there's always one open. but that doesn't mean the open one has its button disabled
Matt_King: but the button that collapses it doesn't do anything, so it's better to be disabled - right?
Matt_King: rather than removing the button?
bryan: I see accordions all the time that DON'T use aria-disabled, rather than this kind
Matt_King: because they always allow you to collapse it...right?
bryan: no - you can collapse it by expanding another section
bryan: -expanded is meant to convey the current state, not whether or not you can do something with it. imo
<jamesn> https://
bryan: in other words - I think that, for an active accordion, it's hard to tell if there's a REAL disabled button, rather than this kind of "fake" disabled button
Matt_King: but IF an accordion section is disabled, what does it mean? you cannot expand it?
bryan: yes, it's inactive.
Matt_King: okay, it's collapsed and you can't expand it - so now what about the expanded one? can you NOT change the expanded state unless you go to ANOTHER section header and expand THAT ONE?
Matt_King: ergo, it's disabled.
bryan: okay, but I guess what I'm saying is this is more confusing than what's out in the wild
jongunderson: i think part of this is a UI design preference. APG can't solve those. it CAN say "doing it THIS way, then program it THIS way."
<siri> + Jon
jongunderson: we aren't going to (and can't make) a pattern for everybody or dictate what they use
jamesn: I did include Microsoft's component if anyone needs an example.
MarkMcCarthy: this will be a good discussion for next week