IRC log of aria on 2021-11-04
Timestamps are in UTC.
- 16:59:39 [RRSAgent]
- RRSAgent has joined #aria
- 16:59:39 [RRSAgent]
- logging to https://www.w3.org/2021/11/04-aria-irc
- 16:59:41 [Zakim]
- RRSAgent, make logs Public
- 16:59:43 [Zakim]
- please title this meeting ("meeting: ..."), jamesn
- 16:59:46 [MarkMcCarthy]
- present+
- 16:59:48 [jamesn]
- present+
- 16:59:50 [StefanS]
- present+
- 16:59:50 [pkra]
- pkra has joined #aria
- 16:59:51 [joanie]
- present+ Joanmarie_Diggs
- 16:59:53 [jamesn]
- meeting: ARIA WG
- 16:59:54 [pkra]
- present+
- 16:59:58 [jamesn]
- chair: JamesNurthen
- 17:01:10 [jamesn]
- agenda+ [New Issue Triage](https://bit.ly/3CH8cfG)
- 17:01:10 [jamesn]
- agenda+ [New PR Triage](https://bit.ly/3wcnVRy)
- 17:01:10 [jamesn]
- agenda+ [Deep Dive planning](https://bit.ly/aria-meaty-topic-candidates)
- 17:01:10 [jamesn]
- agenda+ [Updating ARIA 1.2 due to IDL implementations (exit and re-enter CR?)](https://github.com/w3c/aria/issues/1598)
- 17:01:11 [jamesn]
- agenda+ [A role for indicating whether a given ruby represents phonetics ](https://github.com/w3c/aria/issues/1620)
- 17:01:11 [jamesn]
- agenda+ [Updated spec text to reflect the processing of hidden elements when c…](https://github.com/w3c/accname/pull/137)
- 17:01:11 [jamesn]
- agenda+ [Clarify how "otherwise interactive" relates to overriding the none/presentation role](https://github.com/w3c/aria/issues/1628)
- 17:01:12 [jamesn]
- agenda+ [clarify img naming steps](https://github.com/w3c/html-aam/pull/322)
- 17:02:01 [spectranaut]
- present+
- 17:02:10 [spectranaut]
- scribe: spectranaut
- 17:02:16 [spectranaut]
- zakim, first item
- 17:02:16 [Zakim]
- I don't understand 'first item', spectranaut
- 17:02:22 [spectranaut]
- zakim, next item
- 17:02:22 [Zakim]
- agendum 1 -- [New Issue Triage](https://bit.ly/3CH8cfG) -- taken up [from jamesn]
- 17:02:33 [sarah_higley]
- sarah_higley has joined #aria
- 17:02:37 [sarah_higley]
- present+
- 17:03:18 [spectranaut]
- https://github.com/w3c/aria/issues/1636
- 17:04:04 [spectranaut]
- jamesn: let's leave this to simmer for a bit and discuss next week
- 17:04:39 [spectranaut]
- jamesn: https://github.com/w3c/aria/issues/1634 move to practices?
- 17:06:13 [spectranaut]
- jamesn: https://github.com/w3c/accname/issues/146 agenda+ for next week.
- 17:06:37 [spectranaut]
- https://github.com/w3c/aria/issues/1632
- 17:06:44 [spectranaut]
- jamesn: why not support on role="group"?
- 17:07:03 [spectranaut]
- aaron: seems like it get revoked, maybe not intentionally?
- 17:07:15 [spectranaut]
- aaron: fieldset is a group, and fieldset you can still disable
- 17:08:07 [spectranaut]
- stefan: original idea declare an entire group, all children get disabled. old discussion
- 17:08:26 [spectranaut]
- jamesn: milestone 1.3? anyone want to be assigned to determine a proposal?
- 17:08:35 [spectranaut]
- sarah_higley: scott already has a PR for it
- 17:08:47 [spectranaut]
- jamesn: is this a deuplicate?
- 17:09:09 [spectranaut]
- sarah_higley: related to aria-disable 1130 issue...
- 17:09:32 [spectranaut]
- jamesn: milestone 1.3, lets revist after the other issue is resolved.
- 17:09:41 [sohara]
- sohara has joined #aria
- 17:09:43 [spectranaut]
- https://github.com/w3c/aria/issues/1630
- 17:10:11 [spectranaut]
- jamesn: spec or practices?
- 17:10:44 [spectranaut]
- sarah_higley: we just moved an issue over to APG related to overflow tabs, seems related to this and secondary tabs.
- 17:11:00 [spectranaut]
- sarah_higley: you can assign to me
- 17:11:18 [spectranaut]
- jamesn: milestone 1.3 not to lose
- 17:11:29 [spectranaut]
- https://github.com/w3c/aria/issues/1629
- 17:11:45 [spectranaut]
- scott: you can assign this issue to me, but not a 1.3 issue....
- 17:11:56 [spectranaut]
- jamesn: does this need normative changes?
- 17:11:59 [spectranaut]
- scott: no
- 17:12:34 [spectranaut]
- jamesn: maybe brennan can do it?
- 17:12:43 [spectranaut]
- jamesn: let's ask him
- 17:13:00 [spectranaut]
- jamesn: I won't assign it to you yet, but we can come back and assign you later on if need be
- 17:13:10 [spectranaut]
- zakim, next item
- 17:13:10 [Zakim]
- agendum 2 -- [New PR Triage](https://bit.ly/3wcnVRy) -- taken up [from jamesn]
- 17:13:54 [spectranaut]
- jamesn: https://github.com/w3c/aria/pull/1635 ready to merge
- 17:14:03 [spectranaut]
- https://github.com/w3c/html-aam/pull/354
- 17:15:29 [spectranaut]
- https://github.com/w3c/html-aam/pull/350 needs review
- 17:15:47 [bkardell_]
- bkardell_ has joined #aria
- 17:15:57 [spectranaut]
- scott: will need implementation when lands. no aria-spec impact, just html
- 17:16:06 [spectranaut]
- jamesn: we should have implementation review
- 17:16:16 [spectranaut]
- jcraig: I'll review
- 17:16:29 [spectranaut]
- aaron: I'll review
- 17:16:46 [spectranaut]
- joanie: I'll review
- 17:16:57 [spectranaut]
- zakim, next item
- 17:16:57 [Zakim]
- agendum 3 -- [Deep Dive planning](https://bit.ly/aria-meaty-topic-candidates) -- taken up [from jamesn]
- 17:17:13 [aaronlev]
- aaronlev has joined #aria
- 17:17:16 [spectranaut]
- jamesn: next week I can't do the deep dive
- 17:17:35 [spectranaut]
- jamesn: anyone else on holiday?
- 17:18:00 [spectranaut]
- jamesn: I guess we should meet next week!
- 17:18:26 [spectranaut]
- jamesn: because we have so much to do, but no deep dive. if you have a topic for the following week, please send
- 17:18:30 [spectranaut]
- zakim, next item
- 17:18:30 [Zakim]
- agendum 4 -- [Updating ARIA 1.2 due to IDL implementations (exit and re-enter CR?)](https://github.com/w3c/aria/issues/1598) -- taken up [from jamesn]
- 17:18:48 [jamesn]
- https://github.com/w3c/aria/pull/1633
- 17:18:49 [spectranaut]
- jamesn: I created a pr based on our discussions during the meeting last week
- 17:19:23 [spectranaut]
- jamesn: I would like more reviews
- 17:19:43 [spectranaut]
- jamesn: we have no implementation in a shipping spec of IDL
- 17:19:51 [spectranaut]
- jcraig: only in the browsers
- 17:21:15 [spectranaut]
- jamesn: can aaron ping dominic?
- 17:21:38 [spectranaut]
- aaronlev: yes
- 17:22:15 [spectranaut]
- jamesn: I'll try pining dominic and anna in the thread?
- 17:22:45 [spectranaut]
- jamesn: this is the most urgent thing we are doing, it is the blocker for 1.2, I would appreciate as much review as possible on this!!!!
- 17:24:36 [spectranaut]
- zakim, next item
- 17:24:36 [Zakim]
- agendum 5 -- [A role for indicating whether a given ruby represents phonetics ](https://github.com/w3c/aria/issues/1620) -- taken up [from jamesn]
- 17:25:25 [spectranaut]
- jamesn: is there still work to do on this?
- 17:25:37 [siri]
- siri has joined #aria
- 17:26:06 [spectranaut]
- jcraig: seems like he is open to us closing it
- 17:26:33 [spectranaut]
- jamesn: can I leave it to you and aaronlev to finalize these issues and close it out?
- 17:26:43 [spectranaut]
- jcraig: I'm going to close the issue, and if we need to reopen we can
- 17:26:55 [spectranaut]
- zakim, next item
- 17:26:55 [Zakim]
- agendum 6 -- [Updated spec text to reflect the processing of hidden elements when c…](https://github.com/w3c/accname/pull/137) -- taken up [from jamesn]
- 17:27:22 [jcraig]
- s/the issue/issues 1619 and 1620/
- 17:28:01 [spectranaut]
- aaronlev: I feel like the innertext was a nice simplification until we realized there were too many details
- 17:29:01 [spectranaut]
- aaronlev: firefox doesn't implement it quite correctly
- 17:29:14 [spectranaut]
- jamesn: having inconsistent implementations.... is a problem
- 17:29:24 [spectranaut]
- jamesn: is your solution for chrome a solution other people could use?
- 17:29:45 [spectranaut]
- aaronlev: I think so, you can make an accessible object for a dom node even it is not in the final platform tree
- 17:29:53 [spectranaut]
- aaronlev: it might be harder in other browsers
- 17:30:19 [spectranaut]
- aaronlev: because of the number of accesisbility APIs we support, with only one tree in blink, and we have to send from render to browser process... we have ways of skipping nodes in output
- 17:30:30 [spectranaut]
- aaronlev: when we are computing name we can use fuller tree
- 17:30:44 [spectranaut]
- aaronlev: I think firefox will have to move to that model as well
- 17:31:12 [spectranaut]
- jamesn: does the fact that we don't have consistent implementations mean we can get to CR, or it's ok because it has always been broken?
- 17:31:27 [spectranaut]
- aaronlev: chrome, we can do it based on the spec now or based on some different innertext thing
- 17:31:41 [joanie]
- q+ To ask what exactly what the spec says now means
- 17:31:46 [spectranaut]
- aaronlev: probably firefox and webkit can support current spec, but its probably harder
- 17:31:58 [spectranaut]
- aaronlev: we are flexible if you want to do something else
- 17:34:05 [jamesn]
- ack joanie
- 17:34:05 [Zakim]
- joanie, you wanted to ask what exactly what the spec says now means
- 17:34:10 [spectranaut]
- jamesn: currently everything is broken in a different way
- 17:34:47 [spectranaut]
- joanie: what do you mean when you say what you implemented what the spec says... I think the spec says don't go down the tree!
- 17:35:06 [spectranaut]
- joanie: apple does what the spec currently confusingly says
- 17:35:36 [spectranaut]
- joanie: the spec is in accname, if the div is hidden, it's innertext, but any children that are not directly references by the idea of aria-labelled by, they are not included
- 17:36:38 [spectranaut]
- jcraig: brian wrote the first version of the spec change, I clarified, pull 137 -- if chrome is doing that specifically, but this PR mentions innerText, chrome is not using innerText?
- 17:36:52 [spectranaut]
- jcraig: aaronlev can you reply with PR changes to match your implementation...?
- 17:37:36 [pkra]
- sorry, have to drop off.
- 17:38:19 [spectranaut]
- matt_king: joanie, your summary... I thought the whole purpose was for this to be understandable by users, web authors who are trying to make accessible labels. my interpretation of what aaronlev was saying is that chrome is doing what people thought the spec is meant to say. What the spec actually says is bizarre and unexpected.
- 17:38:39 [spectranaut]
- matt_king: I thought the point was to clear up and include the tree
- 17:38:54 [spectranaut]
- aaronlev: you might want to have a hidden label or description..
- 17:39:46 [jcraig]
- q+
- 17:40:26 [spectranaut]
- bryan: .... a real world example I couldn't catch!
- 17:40:45 [spectranaut]
- matt_kind: good example
- 17:40:59 [spectranaut]
- matt_king: is there a concern with the new proposal?
- 17:40:59 [jamesn]
- ack jcraig
- 17:41:16 [MarkMcCarthy]
- s/matt_kind: good/matt_king: good
- 17:42:26 [spectranaut]
- jcraig: if part of the label is displayed, and part not displayed, how do we computer? Once the ancestor is not displayed, we don't care why the sub level elements are not displayed.
- 17:42:43 [spectranaut]
- jcraig: I don't think we should exposed those substrings willy nilly
- 17:43:05 [spectranaut]
- jcraig: what is the PR change we can review, the devil is in the details, I can't tell you what I think until we see the PR
- 17:44:04 [jamesn]
- q+
- 17:44:07 [StefanS]
- q+
- 17:44:09 [spectranaut]
- aaronlev: I don't think the double hidden is a real case. if people used aria-label... probably because it's in the dom already. if you have double hidden, you can expect predictable behavoir
- 17:44:12 [jamesn]
- q+ bryan
- 17:44:51 [jamesn]
- q-
- 17:45:15 [jamesn]
- q++ bryan
- 17:45:21 [jamesn]
- q-+
- 17:45:42 [aaronlev]
- q+
- 17:45:48 [spectranaut]
- sarah_higley: we reference tool tips via labelled-by or described by, it's hidden until you have focus, but we want the labelledby or describedby to be consistent. tooltips contain whatever text you want, some of it is display none, the displaynone in the displaynone tool tip I would expect as an author for it to NOT be used
- 17:46:01 [spectranaut]
- matt_king: I can't imagine a display-none in a display-none
- 17:46:11 [spectranaut]
- sarah_higley: it might be aria-hidden....
- 17:46:22 [spectranaut]
- matt_king: I can understand that
- 17:46:48 [jamesn]
- q+
- 17:46:59 [spectranaut]
- jcraig: sometimes password fields have a bunch of rules, but only the unfullfilled criteria will be visible
- 17:47:40 [spectranaut]
- stefan: is there a general idea of how to handle these nested situations generally???
- 17:47:53 [jamesn]
- ack ste
- 17:47:58 [jamesn]
- ack bry
- 17:48:07 [spectranaut]
- bryan: in the case where it is area hidden... it doesn't matter what is on the inside
- 17:48:14 [jamesn]
- ack aa
- 17:48:16 [StefanS]
- q-
- 17:48:35 [spectranaut]
- aaronlev: the problem with display none is that there is implementation problems. the browser will do many optimizations.
- 17:49:04 [spectranaut]
- aaronlev: a lot of styling is not applied, its hard to have a special rule about display-none subtrees
- 17:49:33 [jamesn]
- ack me
- 17:49:35 [spectranaut]
- aaronlev: if people want hidden sub trees in display-none... they would have to use aria-hidden, every other way of hiding something would not have an impact
- 17:49:44 [spectranaut]
- jamesn: sounds feasible
- 17:49:55 [jcraig]
- q+
- 17:50:21 [spectranaut]
- jamesn: while we may have one user agent that works, we don't have consistency, so authors who are using this stuff are already broken. I don't care about breaking authors who are already broken in a slightly different way.
- 17:51:01 [spectranaut]
- jcraig: aaronlev mentioned the only thing that can hid substrings in hidden elements is "aria-hidden=false", I don't think this is a fair expectation for web authors
- 17:52:01 [spectranaut]
- jcraig: if you landed on the tool tip, and it is displayed, then you follow the expectations of hidden vs not hidden. if you land on the tool tip and it is NOT displayed for some reason, the aria-describedby is going to follow different rules
- 17:52:36 [spectranaut]
- jamesn: displayed vs not displayed should not change the value
- 17:52:44 [jamesn]
- agenda?
- 17:52:58 [jcraig]
- q+ to defend innertext
- 17:52:58 [spectranaut]
- matt_king: I was going to say that might not be the best example.
- 17:53:37 [spectranaut]
- jamesn: there might not be much to gain with this discussion. someone needs to come up with another example/fix for us to change our mind on the current change proposed
- 17:54:51 [spectranaut]
- jcraig: if we make this change, we have three working implementations. aaronlev did point out some edge cases, I would propose that if that is the case, then we should point out that case in the implementation of innerText its self. it if it is the wrong value for accessibility, then maybe it's wrong for the web generally
- 17:55:04 [spectranaut]
- jamesn: I doubt innertext will change for this
- 17:56:19 [jamesn]
- agenda?
- 17:56:28 [aaronlev]
- maybe a deep dive topic?
- 17:56:49 [jcraig]
- +1
- 17:56:54 [spectranaut]
- bryan: I will put example in this PR
- 17:57:05 [spectranaut]
- zakim, next item
- 17:57:05 [Zakim]
- I see a speaker queue remaining and respectfully decline to close this agendum, spectranaut
- 17:57:06 [jcraig]
- s/its self/itself/
- 17:57:33 [spectranaut]
- queue?
- 17:57:41 [jamesn]
- https://github.com/w3c/aria/issues/1628
- 17:57:44 [spectranaut]
- ack jcraig
- 17:57:44 [Zakim]
- jcraig, you wanted to defend innertext
- 17:57:50 [spectranaut]
- zakim, next item
- 17:57:50 [Zakim]
- agendum 7 -- [Clarify how "otherwise interactive" relates to overriding the none/presentation role](https://github.com/w3c/aria/issues/1628) -- taken up [from jamesn]
- 17:58:12 [spectranaut]
- jcraig: one of the webkit implementers raised this
- 17:58:34 [spectranaut]
- jcraig: I think this means the spec is right but could use some clarification, and all the implementations are wrong in this case,
- 17:59:02 [jamesn]
- q+
- 17:59:07 [spectranaut]
- jcraig: if this element is interactive (including if its focusable), then the presentation role should be ignored.
- 17:59:56 [spectranaut]
- jcraig: onClick should make something interactive and none role should be ignored?
- 18:00:02 [spectranaut]
- jcraig: what about drag events?
- 18:00:38 [spectranaut]
- jcraig: should we take delegation into consideration? like click event on body? everything can't be excluded from presentation
- 18:01:01 [spectranaut]
- jamesn: lets discuss in the issue!
- 18:01:16 [spectranaut]
- zakim, make minutes
- 18:01:16 [Zakim]
- I don't understand 'make minutes', spectranaut
- 18:01:45 [spectranaut]
- RRSAgent: make minutes
- 18:01:45 [RRSAgent]
- I have made the request to generate https://www.w3.org/2021/11/04-aria-minutes.html spectranaut
- 19:14:56 [jongunderson]
- jongunderson has joined #aria