<jamesn> 5:08 PM agenda+ [New PR Triage](https://github.com/search?l=&q=is%3Aopen+is%3Apr+repo%3Aw3c%2Faria+created%3A%3E%3D2019-07-18+repo%3Aw3c%2Faria+repo%3Aw3c%2Faccname+repo%3Aw3c%2Fcore-aam&type=Issues)

<jamesn> 5:08 PM agenda+ [Agenda for TPAC 2019](https://github.com/w3c/aria/issues/1001)

<jamesn> 5:08 PM agenda+ [Revising what ARIA attributes should be considered global](https://github.com/w3c/aria/issues/999)

<jamesn> 5:08 PM agenda+ Naming Techniques Discussion

<scribe> scribe: melanierichards

https://github.com/w3c/aria/issues/1022

jamesn: 1.2?

joanie: Scott pointed out biggest problem is handled in HTML-AAM, but there's still some in ARIA on me

https://github.com/w3c/aria/issues/1021

jamesn: way too big to tackle in 1.3
... do we even want to milestone, or punt further?

https://github.com/w3c/aria/issues/1019

jamesn: later in agenda, 1.2?

https://github.com/w3c/aria/issues/1017

jamesn: putting a question label instead of a milestone

Scott_O: I can take a stab at that, it's related to toast component

https://github.com/w3c/aria/issues/1015

carmacleod: talked about it last time, it's editorial
... probably won't be done til September

jamesn: all are in the agenda

[Agenda for TPAC 2019](https://github.com/w3c/aria/issues/1001)

<Jemma> I think Matt King wanted to talk about list and group so I will add that to the agenda

jamesn: TPAC becoming urgent, we need agenda items. And if we need to coordinate with other groups, we need to know today or tomorrow
... please take a look at this today if possible

carmacleod: can I put F2F label on DOM order issue?

jamesn: you can put the candidate label on it, anyone can put a candidate on an issue

melanierichards: annotations on the agenda?

jamesn: planning on having something to share with the group soon, will be discussing annotations with the annotations WG at TPAC. Unfortunately Aaron won't be there

jongund: aria-details?

jamesn: probably related to the discussion but feel free to add if you have a specific question

<Jemma> APG meeting minutes did not record that ARIA list structure will be one of the TPAC agenda. https://github.com/w3c/aria/issues/1011

jamesn: will try to do annotations for best possible time for call-ins

[Revising what ARIA attributes should be considered global](https://github.com/w3c/aria/issues/999)

<carmacleod> https://github.com/w3c/aria/pull/1023

jamesn: I took a first pass at a PR w/r/t global ARIA attributes based on what Scott wrote
... doesn't change anything on aria-describedby and aria-details. Didn't do anything with aria-controls.

mck: the comments you're talking about, are they all in #999?

jamesn: yes

mck: well we did come to some decisions w/r/t aria-controls in APG...maybe I didn't make a PR for this. One thing we talked about was clearly defining the controlling thing. We were going to bring to larger group
... if the group agrees with that, it would be pretty easy to restrict aria-controls to things that can control

jamesn: [reads out aria-controls definition]
... I think this is a tangent, we have a discussion about how aria-controls isn't well supported by ATs and may want to make other changes restricting use
... would like to go through what we have and put aria-controls as a later topic
... want to do the well understood ones first

mck: ok, if we leave it in here and deal it with it later that would be good

jamesn: my PR is there for discussion, not going to be merged at this point in time

mck: the ones that I'm iffy on are (just an inkling)...aria-haspopup, is that something that you might ever want to make something focusable only because it has a popup, but it has a static structural role. Is that valid markup we want to support?

carmacleod: is a contextmenu a popup?

mck: that's kind of what a popup is, can change semantics on certain placements. I don't know if you're supposed to put aria-haspopup on a tree item if it has a context menu, we don't do that in APG...menu buttons and comboboxes are really clear about needing aria-haspopup, otherwise I'm not clear on when you need it if you have a context menu
... do we have a clear understanding of the use of aria-haspopup?

carmacleod: still not sure about context menu, which could go on anything

Scott_O: that's not communicated right now

mck: nothing in aria-haspopup itself that indicates that it's only for certain kinds of things. James, what change did you make?

jamesn: what Scott put in his comment

Scott_O: I looked at this as interactive or form controls are where people are usually aria-haspopup to load menus, listboxes, etc

mck: notable that you skipped tab, I've noticed it pretty extensively on tab
... anything in the browser has a context menu...

Scott_O: and that's not communicated now. We're not saying everything needs "aria-haspopup"

mck: put it there only if you customized the context menu? but that could get equally annoying

Scott_O: it's originally for menu role and then got expanded
... not all SRs have updated to respect all the different roles where supported

stefan: I want aria-haspopup as general as possible

mck: then we have a problem of telling SR devs what the expectations are for where to support aria-haspopup
... intent when we added in ARIA 1.1 was ONLY to support combobox

jamesn: there are some cases where something looks like a button that launches a popup (menu button) but it's really a dialog

mck: well but in that case SR mode gets confused if you say it's a menubutton but it opens a dialog

stefan: adding a role for dialog button?

mck: that would add more work for API developers

jamesn: so I'm hearing that some things need to happen on aria-haspopup before we can decide on global or not
... we need to clarify its use

[general agreement]

mck: if we want SRs to do more, we have to be really clear about that, otherwise it adds noise like aria-controls

[Scott_O volunteers to file this issue]

jamesn: put on 1.2
... what is our expectation of AT when aria-haspopup is on different roles. Which roles support should fall out of that
... let's park the rest of the discussion,let people read it, and leave comments in the discussion. Keep the discussion in the issue if you can, please only comment on PR if very specific to how PR was done

Naming Techniques Discussion

jamesn: we now have a couple additional naming strategies e.g. encapsulation and legend
... jongund proposing a couple others

jongund: the heading one we discussed at SF F2F

<jongund> https://raw.githack.com/w3c/aria/issue899-accname-from-heading/index.html#namecalculation

jongund: [reads off list of roles that would get name from heading child]
... the other concern was nesting, can heading only be used for closest ancestor vs being the label for multiple elements

jamesn: I don't see main working
... mains generally don't need a name

mck: there's certain situations where you want to name it in single page apps

Scott_O: I don't think you always want articles inheriting from headings

jamesn: some ATs exposing all articles

Scott_O: I thought it was interesting that James C brought this up because VO doesn't announce unnamed articles

<jamesn> https://raw.githack.com/w3c/aria/issue899-accname-from-heading/index.html#namefromheading

https://www.irccloud.com/pastebin/5qsAvYN9/name%20from%20heading

mck: I don't think we should have implicit namings of regions

jongund: you could explicitly name but you wouldn't have to manage it

<jongund> https://raw.githack.com/w3c/aria/issue899-accname-from-heading/index.html#region

mck: you could have a region with several paragraphs, and then a heading, and according to this, that heading could become the name for that region. That could be very misleading.I'd rather have no name.

jamesn: I think having region here is risky
... if we're looking into doing things like this, I'd rather us hammer down the ones that have pretty good consensus first

mck: I think this problem with region could be a problem with others as well (article)

carmacleod: looking at the table element in HTML-AAM, it says very clearly first child, not first descendent
... maybe we should make that clearer

<carmacleod> https://html.spec.whatwg.org/multipage/tables.html#the-caption-element

jongund: so change it to be first child element?

jamesn: do we have broad consensus to not have it on some things?

mck: well if we do first child element than I'm fine with it on region

jamesn: we're at the hour, we didn't get to caption. Can people take a look and leave comments for Jon?
... this could all massively complicate Acc Name, are we ok with that direction?

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)

Present: Joanmarie_Diggs Scott_OHara MarkMccarthy melanierichards carmacleod jamesn JemmaKu Stefan janina Matt_King Neil_Soiffer
Found Scribe: melanierichards
Inferring ScribeNick: melanierichards
Found Date: 01 Aug 2019
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.)