<Matt_King> https://github.com/w3c/aria-practices/wiki/January-28%2C-2020-Meeting
<sarah_higley> Matt: our next meeting is Feb 4th
<sarah_higley> Matt: RIght now there are 60 issues and PRs in the milestone, closed 15
<sarah_higley> Matt: I'm doing some work to trim things down, but there are some large blocks of work that need tackling
<sarah_higley> Matt : I'm estimating we'll end up closing 40 things
<sarah_higley> Matt: there's currently 24 bugs, and I want to see if we can eliminate all of them
<sarah_higley> Matt: Right now we're prioritizing combobox, and there are some of these that need prioritizing
<Jemma> https://github.com/w3c/aria-practices/milestone/2
<sarah_higley> Matt: Is there anyone who would be interested in taking on bug management, I'd like to have an offline discussion
<sarah_higley> Matt: I'd like to find a way to smash bugs faster
<sarah_higley> Jemma: Pesticide?
<sarah_higley> Matt: any questions about what we're doing between now and March, or how we're doing it?
<sarah_higley> Jemma: I can help, but can't lead the project
<sarah_higley> Matt: let's talk about that in the meeting after this meeting
<Jemma> https://github.com/w3c/aria-practices/pull/1120
<sarah_higley> Carolyn: I'm still doing the accessibility review, I have a little ways to go
<sarah_higley> Carolyn: how much are we caring about Edge with Narrator? It doesn't read the live regions, and we just sort of say that?
<sarah_higley> Matt: we don't need to test all the screen reader functionality when we do the accessibility review, that can go to ARIA-AT
<sarah_higley> Matt: but we need to make sure that all the things screen readers need are there
<sarah_higley> Matt: the aria-roledescription stuff really screws stuff up
<sarah_higley> Carolyn: actually it worked for me
<sarah_higley> Matt: It says it, but is screwing up other announcements. NVDA was ignoring it on group
<sarah_higley> James: when I've tested this, I've only done it on other roles, not groups
<Jemma> https://github.com/wnolfm/a11y-slider/issues/12#issuecomment-539091266
<sarah_higley> Carolyn: it works fine as far as I could tell, it says "slide"
<Jemma> https://github.com/wnolfm/a11y-slider/issues/12
<sarah_higley> Matt: oh, I didn't tab, only used the reading cursor
<sarah_higley> Matt: so for our accessibility reviews, we don't need to go deep on screen reader testing
<sarah_higley> Carolyn: with Narrator the live region doesn't work, and arrow keys don't work
<sarah_higley> Matt: it would be a problem if arrow keys don't navigate the tabs and scan mode is turned off
<sarah_higley> Matt: the goal is to make sure there aren't any bugs in our code
<sarah_higley> Sarah: I can test Narrator again, since I know some more of its quirks
<sarah_higley> Sarah: I have questions about having live region there at all, since we don't have it with regular tabs
<sarah_higley> Matt: the idea is a screen reader user can simulate an auto-rotating experience, but with manually moving through the tabs
<sarah_higley> Matt: the idea is that it allows users to skim content. We do it with the next and previous buttons in the other carousel because mostly people just click next, next, next
<sarah_higley> Matt: and a screen reader user would have to move back and forth, and it doesn't support skimming
<sarah_higley> Sarah: it doesn't communicate any information about the structure of the content in there
<sarah_higley> Matt: it worked for me ok
<sarah_higley> Sarah: I wonder if that's because this pattern is very simple, and doesn't have any of the buttons, headings, etc. that are common in carousels
<sarah_higley> Matt: we're planning on doing a usability test at Facebook on this pattern
<carmacleod> There's a paragraph about the use of the live region in the "Screen Reader Announcement of Slide Changes"
<carmacleod> On https://raw.githack.com/w3c/aria-practices/issue955-add-tabbed-carousel-example/examples/carousel/carousel-2-tablist.html
<sarah_higley> Matt: and that example has a lot of other functionality
<sarah_higley> Sarah: yeah, I think I can see that. So that usability test will have a good amount of structured information in it?
<sarah_higley> Matt: yes, that's the goal and should be done in February?
<sarah_higley> Matt: I put that note in there to discourage people from doing that in normal tabs, because usually the tab label is descriptive enough
<sarah_higley> Matt: but the carousel labels are slide 1/2/3/4/5/6
<sarah_higley> Sarah: maybe the better change would be to make the tab buttons themselves have descriptive names
<sarah_higley> Sarah: especially since the live region here is set up to be buggy -- both because it's using CSS display changes instead of DOM insertion, and the content changes happen at the same time as a tab button focus change
<ZoeBijl> hello
test?
<ZoeBijl> something something internet
Sarah: scribe note: an IRC malfunction lost some transcription: Summary: Zoe suggested adding a transparent outline style so the captions in high contrast mode had visible focus styles
Matt: let's capture that in the PR. Code review is not checked off
Carolyn: I think Sarah just needs to re-review
Matt: I guess not today, but this is super closed. Let's try to get it done this week
<Jemma> https://github.com/w3c/aria-practices/issues/970
Matt: this is about the pattern that goes along with our carousel examples
https://github.com/w3c/aria-practices/issues/970
Matt: I was on the verge of
creating issues for things that were in here, but I didn't do
it. I thought it would be better to discuss the comments here
and decide which were issues that we really wanted to
change
... let's go through comment by comment
Jemma: I made a summary of the
feedback
... we were asked if we could have a carousel example without
auto-rotation
... #4: aria-description support issues
... they're suggesting using aria-label instead of
description
... two more feedback: one is display: none issues, I think
this is the one Sarah addressed already
Matt: let's just work through
each of these
... the composite widget thing I think we should probably
respond to. I don't see us making a composite widget for
carousel
... in our carousels, aren't we using display: none? Oh, are
they saying the pattern should specify that people should do
that?
Jemma: the issue I think is that we should not use that. That's what they're suggesting
<Matt_King> link to pattern: https://w3c.github.io/aria-practices/#carousel
Matt: the APG pattern for
carousel right now I don't know if we have a warning that
people have to make sure that the content of slides is
completely hidden, not just visually hidden
... I think that's what they're saying in the display: none
comment
<Jemma> "This leads to inconsistent behaviour between differing implementations of the sliding effect.
<Jemma> In example:
<Jemma> • The provided example hides non visible slides by using display: none, which excludes them (or there focusable children) from the page tab sequence
<Jemma> • In a more common approach, using overflow: scroll and manipulating the scroll position to move slides in and out of the visible area, the slides (and there focusable children) are not removed from the page tab sequence (unless hurting performance with scroll event listeners)"
Matt: under roles, states, and
properties we don't have anything that talks about how to hide
slides accessibly
... in general, we don't talk about those types of engineering
issues in the patterns. Is this one worth making an exception?
Should there be a note?
... if we do it here, we should probably do it in the tabs
pattern
Jemma: why should we make an exception for this one?
Matt: I'm asking if we should. I
do think it's a very common implementation that you can read
all slides that are offscreen, and it's very annoying
... I think I've also seen it with tabs
Sarah: for what it's worth, I've had someone try to tell me that all slides should be in the DOM all the time, and that was more accessible
Matt: Zoe, what was the reason we changed to display: none?
Zoe: because I interpreted the spec in a different way than the authors did, and tried to be the bigger person and made the change in our code
Matt: we had a discussion about
hidden and visibility:hidden and display: none
... maybe we should have a note about those differences in the
APG
... then we could link to it when we say something should be
hidden from all users
... then we could have a very short note in the tabs and
carousel pattern that links to that
... so it sounds like this one is worth turning into an issue.
Composite pattern is not
<Jemma> https://html.spec.whatwg.org/multipage/interaction.html#the-hidden-attribute
Matt: #2 grouped carousel interaction
<Jemma> "I'm trying to understand the delta between the two implementations of the slide picker controls. The primary difference to the control seems to be the addition of rotation and next/previous controls though I'm not sure that requires a completely divergent keyboard interaction model. That seems problematic and confusing for both implementors and users. Are there concerns regarding navigation redundancy by exposing next/previous and individual
<Jemma> buttons to fire content?"
Matt: I can answer that question,
the tabbed pattern makes the whole carousel one tab stop
... I think it was Bryan who suggested including the button
pattern because the tabs pattern would be too heavy for some
simple patterns
Jemma: do we really need a group
role for that?
... should that be a third example?
Matt: yes that should be a third
example, we don't have that yet. So maybe the best way to
address this would be to add a third example
... Bryan's suggestions all already got integrated into the
pattern itself
... yeah, we have basic carousel, tabbed carousel, and grouped
carousel
Jemma: could you re-explain the rationale for having a 3rd example?
Matt: because implementing the tabs pattern just adds complexity, and for carousels that don't have a lot of slides, it may be more complexity than is justified for some implementations
Jemma: so the third issue is a
carousel example without automatic image rotation
... the rationale for this one is based on WCAG 2.2 pause/play
stuff
Matt: so basically we have it,
both of our examples have pause
... so should we have an example where it cannot and does not
rotate at all
... are people going to do that in real life?
Jemma: yeah, Amazon and Microsoft
Matt: so in an example like that there would not be a pause button at all, because it would not rotate
Jemma: I think we can have a
rotating carousel that doesn't start automatically
... it's like a video that starts on mute, we can have it not
auto-rotate, so the default status is a paused status
... not having a pause/play button doesn't make sense to me
<Jemma> # 5 "Feedback: I would really like to see an alternative version of the Carousel example that does not auto rotate at all and the play/pause button is removed."
Matt: so there are three
different rotation ideas here: it can automatically rotate on
load and you can pause it, the second is it doesn't auto rotate
but you can start it, and the third is it doesn't rotate at
all
... we could make each of the examples different that way
... we could make it so that maye the tabbed one when it loads
is not playing but has a play button
... and when we have this third example with the button group
and have it not play at all
Jon: you could just make it a
command line variable and have more variations
... you could just have "carousel paused" in the command line
and javascript looks for that. You don't need a whole new
example for that
Matt: the URL thing is a little tricky. If you make it one of those views, one of those is going to be the default when you reload the page, and it'll always reset to the same one
Jon: when you reload the page,
it'll still be in the command line. It'll be whatever the last
one you selected was
... reload wouldn't affect it
... and we would still have unique links to all of the
variations of the examples
... it'll only change if you change the actual url you link
to
... I guess my concern is that the more variation we have, what
are we telling developers?
... are they looking for something that looks like what they
want, or are they looking for recommendations?
Jemma: I think our example doesn't violate WCAG 2.2
Sarah: WCAG lacks a lot with cognitive SC. I see pause/stop/hide as a bare minimum, and think auto-rotating is an antipattern
Jon: I think auto-rotating is bad, but marketing people win and we should have guidance to help developers do that
Carolyn: there's a CSS media
query called prefers-reduced-motion. I think we should honor
that media query as a bare minimum
... if that is set, we should not auto-rotate
Jemma: that is a WCAG failure too. You cannot violate OS settings
Matt: right, that's what Carolyn was saying
Zoe: you need to detect that media query with javascript
Matt: we can have some discussion
in the issues themselves, then get them back on the agenda
individually
... there's nothing that should hold up current work before we
merge it, but we can make changes later
Jemma: only thing we didn't make a decision on is the aria-roledescription issue
Jon: what if anything should change in the code?
Matt: nothing right now, let's wait until we have issues and decisions. The only current changes are what we discussed in the review section. Nothing else is changing before merging
This is scribe.perl Revision: 1.154 of Date: 2018/09/25 16:35:56 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00) Succeeded: s/gorups/groups/ Succeeded: s/aria-live/aria-label/ Present: Matt_King jongund jamesn sarah_higley carmacleod Jemma No ScribeNick specified. Guessing ScribeNick: sarah_higley_ Inferring Scribes: sarah_higley_ WARNING: No date found! Assuming today. (Hint: Specify the W3C IRC log URL, and the date will be determined from that.) Or specify the date like this: <dbooth> Date: 12 Sep 2002 People with action items: WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option. WARNING: IRC log location not specified! (You can ignore this warning if you do not want the generated minutes to contain a link to the original IRC log.)[End of scribe.perl diagnostic output]