<scribe> scribe: MarkMccarthy

<Sarah_Higley> Are there any docs about scribing, or the commands for Zakim and RRSAgent?

Future meetings

mck: I'm in a conference next week, so can't run meeting. Jemma can't either, but we're in a good place to take a breather
... if we need input from the group, I'll find an asynchronous way. No meeting on 16, meet again 23rd

jongund: Conflict on 23rd and 30th

mck: so we'll talk about priorities for the rest of the year, 1.2 near the end of the year, so won't continue with 1.1 stuff coming soon. if there are agenda topics or priorities, send them in!

Publication schedule

mck: we're going to cut the 1.1 branch tomorrow night and send out a CfC
... we're going to go through the top priorities in the following topics

aria-details be part of descriptions section?

mck: jongund is making a case it shouldn't

jongund: right now I don't think it's very well supported. if we look at the AAM, it doesn't seem to be...
... the biggest problem I see is in the spec, there's conflicting information. one of the primary features is interoperability, 1.1 spec says "some api mappings will override -describedby".
... macOS doesn't map it at all. so I don't know that we could reliably recommend it
... it's not part of the accname calculation, maybe BGaraventa has some more info
... i think there's some things that need to be worked out in spec first, premature to include it as part of accname description and calculation

mck: okay, Sarah, last week you thought you could do some testing?

Sarah_Higley: I was actually on vacation last weekend, but I can have results by next week if that's okay

mck: well jongund, are you suggesting that it shouldn't be in 1.1 spec at all?
... because we have concerns about how it works in practice

jongund: i think it's in there because digital publishing wanted it

mck: we tried to reach out to George and some other folks to dig up the history; the reason it's in there now is because it does have an effect on description
... it takes preference over -describedby, but we don't know what a screen reader would do with -details as we don't know about support

<Sarah_Higley> https://www.scottohara.me/blog/2018/11/08/aria-details.html

Sarah_Higley: Scott O'Hara has some details (linked above)

BGaraventa: as far as accname computtaion goes, I haven't included that, primarily becuase of confusion on what should be included
... I heard the content from -details shouldn't be included in the same way that -describedby, right?

jongund: it's not supposed to generate a string like -describedby

mck: the spec says if an API can't support multiple references, -details takes precedence; content of -details should be used to help calc accname

BGaraventa: well, no. AT would say there's information, but it's unclear what would happen if there was a description that goes to a named region

mck: so, if there's related information ....

jongund: the only API that has a conflict is UIAutomate, the other mappings (IA2, MSAA, ATK) are othagonal, so they use lots of information to compute the descriptions/names

mck: what about AX?

jongund: it's not supported, so it doesn't work with MacOS. It can't interfere because the Mac doesn't know it's there
... I'm not exactly sure what UIAutomate does, but -details maps to the same IDREF as -describedby
... so basically it'll plug in the ref for -describedby

Sarah_Higley: -details doesn't appear to affect help text in UIA, unlike -describedby
... well, in practice

jongund: do we want something with this many problems in the AP right now?

mck: 1, we can remain silent, which might be a bad idea. 2, we can put it in some kind of warning or guidance (maybe in -describedby section), as maybe there's no danger of using it if it doesn't interfere

jongund: well they need to be warned some platforms don't get this info

mck: well, maybe the only options are to include a warning if it interferes, or remove it entirely
... if it appears it never interferes, we don't need a warning

Sarah_Higley: I don't know for sure that it doesn't interfere, but that UIA puts this information into help text.
... There might be different interpretations, not sure if there's enough screen reader support

mck: do any web browsers support UIA?

Sarah_Higley: chromium edge will probably add UIA support, alongside IA2
... probably better to ask someone on the chrome edge team

mck: i don't really know what to do, not sure we have enough information
... would it be better to remove it completely, or leave as is?

jongund: remove completely

Sarah_Higley: leave it out, for now at least

jongund: my testing shows that most screen readers ignore it. in 6 months, maybe things will change
... will there ever be a mapping for AX for example?

mck: okay, I think we should add aria-details to the TPAC agenda
... not hearing support for leaving it in, i'll get it out of the accname section and somewhere else for safe keeping
... we'll discuss this in the working group

naming and describing feedback

mck: I think I addressed all the feedback in issue 1050, other than offtopic comments,
... BGaraventa have you been able to read the entire draft yet?

BGaraventa: I can do that this afternoon

mck: that'd be really helpful, especially in light of this discussion
... there's a link to the latest version in the issue
... BGaraventa, are you comfortable leaving the information about groups and author content the way it is? it seems like there's inconsistency in implementation
... i don't want to document that in AP, but I don't know what might be best

BGaraventa: I'm going to be responding to Aaron later today, and just that.

mck: if you see something that should change, based on that thread, let us know
... not sure if there are some other changes that should be made, I'm mostly comfortable with things as is

BGaraventa: i'll read through and comment where I can. I'll be out next week, back on the following, just so you know

mck: we're cutting the branch tomorrow, so that shouldn't be a factor

Date picker dialog

mck: jongund, i saw you removed -controls

jongund: yup and took care of -expanded

mck: okay cool. I did see we're referring the button in two different ways. when refer to it in the doc, we should use its accname?
... or how to refer?

jongund: maybe "choose date". I can make those changes. It's mostly in the keyboard support docs

mck: cool, that'd help
... one editorial thing, when I refer to a button in the middle of a sentence, do we normally put the name in quotes to make it recognizable? I want to make sure we're more consistent moving forward
... do we capitalize? put in quotes? i'm leaning towards always quoting them

jongund: quotes are good

MarkMccarthy: I like quotes

mck: using q element I suppose
... anything we do inside the code tags, we can use regular quote characters

Other outstanding PRs

mck: I think we're ready to go with disclosures, except one thing Sarah?
... we don't need to clean up history, because they get squashed

Sarah_Higley: I just noticed that
... I try to keep feature branch commits at the top

mck: I just use Github's auto merge feature, so it can run tests etc.
... if you prefer rebasing, I can do that, it's just not as convenient

Sarah_Higley: I can really work with whatever, but if i'm working on it myself I prefer to rebase so i can see history etc.

mck: there's a merge in this branch which resulted from a pull, so it didn't fast forward. it might've gotten squashed and had a few commit conflicts on my local
... not a big deal, just an extra merge
... i just pushed a rewrite of the accessibility features bullet before this meeting

Sarah_Higley: there was a brief dicussion about how to articulate that, so I was wondering what the best course of action might be
... i was initially trying to avoid being too specific about any certain behavior

mck: right, so I reworded it and I hope I captured it
... [discussing github preview linking issues]

<mck> https://raw.githack.com/w3c/aria-practices/issue1028-disclosure-menu-example/examples/disclosure/disclosure-navigation.html

mck: here's a preview link to the changes (above)
... under accessibility features, the first three bullets are unchanged, the fourth one about CSS is unchanged, number 5 is what was changed

<mck> 2. Optional navigation key support is primarily for the benefit of keyboard users who are not running a screen reader. Screen readers that have both reading

<mck> and interaction modes intercept these navigation keys and do not pass them to the page when the screen reader is in reading mode. When interacting with

<mck> this example, such screen readers are typically in reading mode because this example does not use a widget role that triggers screen reader interaction

<mck> mode.

mck: [changed text above]
... if people are okay with that wording, then i can use that to fix your comment on the following bullet (number 3)

Sarah_Higley: that looks good for 5.2

mck: one way to word this, that seems techy, is "because this example doesn't use a composite widget role, tab is expected to navigate..." so just changing the reason why
... but that might be a little too much unless people read and understand the spec

Sarah_Higley: I think everyone would benefit from having the most baseline keyboard interaction

mck: so I think I can take advantage of the changes to make what's in 5.3 clearer
... any other changes?

Sarah_Higley: only something optional, a couple mispellings (which I don't see now), so it should be fine

mck: disclosure example looks ready, spectranaut, were there changes for auto rotating carousel?

spectranaut: I think it's ready

jongund: yeah, it should be

mck: tooltip on toolbar should be ready, jongund did you look at it?
... I don't see the title attr anywhere, so that's good
... date picker spin button, how're we doing?

Sarah_Higley: I left some comments, which jongund responded to
... I noticed the year doesn't wrap, day and month do. argument was that it can be an example for how to do both.

jongund: i didn't add anything into documentation about that, we could add something though

mck: if we do add it, we need a test for it

Sarah_Higley: documentation says some dates wrap
... [reading docs about days and months for date picker]
... I think it's fine

jongund: the other comment was aria value text updating for year and it didn't need to, thanks for catching that Sarah_Higley

mck: do these spin buttons have visible touch targets?

Sarah_Higley: yeah

jongund: yep
... they're hidden with aria-hidden

mck: if using this on mobile, would you have to type?

jongund: there's no edit box, this is just three spin buttons

mck: screen reader says there's edit boxes

jongund: but there's no input element

mck: weird, JAWS says it's editable
... so everything's looking pretty good. thanks everyone, we'll have an announcement out soon.

jongund: if there are any changes needed until the 20th, I can do that.

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/oaky/okay/
Succeeded: s/remove as is/leave as is/
Succeeded: s/q tags/q element/
Succeeded: s/catching/catching that/
Present: jongund Sarah_Higley MarkMccarthy spectranaut ZoeBijl BryanGaraventa Bryan_Garaventa
Found Scribe: MarkMccarthy
Inferring ScribeNick: MarkMccarthy
Agenda: https://github.com/w3c/aria-practices/wiki/July-9%2C-2019-Meeting

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: 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.)