<jamesn> agendabot, this is public-aria
<agendabot> jamesn, OK, using https://lists.w3.org/Archives/Public/public-aria/
<jamesn> agendabot, find agenda
<agendabot> jamesn, OK. This may take a minute...
<agendabot> clear agenda
<jamesn> pesent+
<carmacleod> scribe: carmacleod
4 new issues
https://github.com/w3c/aria/issues/1328
<jcraig> other treeitem issue search results: https://github.com/w3c/aria/issues?q=is%3Aissue+is%3Aopen+treeitem
jamesn: I think we have discussed
this several times before - would be good to look up those
conversations and include them here
... for triage - if there are API extensions, we can't really
do this in 1.3
StefanS: even without API extensions, if we could have a way to wrap the content of a cell to indicate that there are multiple children, that would be useful
jcraig: is this about keyboard interaction or communicating semantics?
StefanS: we need a common understanding about what could be done - agree that keyboard interaction and communicating about content both apply
jcraig: agree this is not 1.3 - at least 1.4
https://github.com/w3c/aria/issues/1325
jamesn: this doesn't seem very different from a grid, so something we could conceivably be done in 1.3
sarah_higley: if you want it to be exactly like Windows Explorer, then there would be more work, because ListView is both list and columns
Scott_O: trige 1.3 and have discussion?
<sarah_higley> ... I have an existing grid implementation that acts mostly like Windows Explorer, if this ends up being an APG thing
<sarah_higley> already explored it because Microsoft web UIs try to emulate this all the time :)
aaronlev: Chrome has concept of listgrid - primary keyboard navigation is use arrows to navigate rows, can use tab key within row
<Zakim> jcraig, you wanted to mention 1) the Mac mapping for single column tables (aka "grids") and 2) the author confusion re: naming
jcraig: the Mac mapping for this
is just single column tables (aka "grids")
... a concern is author confusion re: naming... if any new
roles are needed, the name should clearly communicate a
difference between the new role and the existing list or grid
roles
sarah_higley: agree with jcraig - I don't think we need a new role - I've implemented something like this before - can implement something for APG to try out
jamesn: thank-you so much, Sarah!
https://github.com/w3c/aria/issues/1324
jcraig: agree that we should get rid of those
StefanS: it's just confusing
jamesn: we do say that it can be used for layout
<Jemma> can either Stefan or James C point out the layout example at APG?
jamesn: 1.3 milestone
<jcraig> layout examples are linked from the issue
https://w3c.github.io/aria-practices/examples/grid/LayoutGrids.html
https://github.com/w3c/aria/issues/1323
jamesn: editor's draft of spec already says should use -1
jamesn: proposals are due Oct 19,
call for breakout sessions is open, if anyone wants to talk
about anything, add to wiki
... anyone planning to have a session?
Sina: can anyone attend any of the sessions
jamesn: yes
Deep Dive #3 on Trees
<jcraig> scribe:jcraig
carmacleod: quoting spec: "selection normally follows focus and is managed by UA"
if selected needs to be on option, and could be subbed for aria checked, we need to know if the UA is managing focus, to avoid dual management
aaronlev: history is that we
tried to match Windows widgets
... two cases are 1) tab into single select list,
focus/selection would both move... and 2) focus would indicate,
but selection didn't occur until you used an arrow, then focus
and selection would match (like default radio button
groups)
... core aam has some guidance on single selection
containers... scattered in several places
<carmacleod> https://w3c.github.io/core-aam/#mapping_events_selection
<jamesn> https://www.w3.org/TR/core-aam-1.2/#mapping_events_selection
aaronlev: prevents weird situations and saves author some trouble
jamesn: author not doing [sic] aria selected
carmacleod: proposal was that: author didn't have to use aria-selected
aaronlev: if saying in one way, why the need to say it in two ways
sarah_higley: better to be explicit and keep it in
don't give authors the opportunity to remove aria for the sake of their dev practice, could lead to confusion about when it was okay to remove
sina: consequence if we force selected being spoken by some SRs
sarah_higley: it doesn't speak it
aaronlev: perhaps because the SR is not reading it on single-selection widgets
sarah_higley: multi sleect does not come into play
carmacleod: all UAs doing this?
aaronlev: I think so
may depend on the platform, there is one case (#2 above I mentioned above) where the initial selection state is false
jamesn: seems reasonable
sarah_higley: as long as it does not say authors MUST NOT or SHOULD NOT. I will be a reviewer.
carmacleod: mentioned related RFC statement in Core AAM
jamesn: sarah_higley, myself, and aaronlev will review
aaronlev: we don't have to disallow this. seems unusual but I can see someone may have a reason to do it.
sarah_higley: lets you escape without selecting the previously selection (scribe help please sarah_higley?)
sina: I thought we assumed selection would follow focus
carmacleod: yes, if you don't set
aria selected
... can joanie review as well? if we should change it to a UA
req, we'd like your feedback too.
<Zakim> aaronlev, you wanted to explain the history on this one
<scribe> scribe: carmacleod
StefanS: I don't think it makes
sense to have checked and selected on the same control
... problem is that people see a checkbox and they think they
need aria-checked, but this is not so
jamesn: why do we allow aria-checked on a listbox
<Jemma> https://github.com/w3c/aria/issues/1052#issuecomment-530635835
Scott_O: last time we talked about it, it seemed like a bi-product of inheritance
jamesn: why do people use
checked?
... on an option?
sarah_higley: I think people see a checkmark and assume they need to add aria-checked
joanie: I looked at the spec, and I see that aria-checked is a supported attribute, not just inherited
<joanie> aria-checked is an explicitly supported state on option in ARIA 1.0
sina: how do i select contiguous with shift+down and then type space to check all?
sarah_higley: just get shift+down to check and select. don't need dual state
sina: I haven't encountered an example before where selecting changes state
sarah_higley: the 2nd example does that
sina: ok, in that example, I press shift and then down, down, down. I don't expect it to do anything that is stateful. I didn't tell it to change the state with space key or anything
sarah_higley: all different mediums should behave the same, i.e. touch
jcraig: touch is tricky
isabel: I have used selected and checked separately
sina: I just want to be able to select a bunch of things and then check them
<pkra> gotta go. bye everyone.
<sarah_higley> sorry I dropped so suddenly, had a bit of an internet snafu 😅
This is scribe.perl Revision of Date 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/jamesn/jcraig/ Succeeded: s/a concern is author confusion re: naming/a concern is author confusion re: naming... if any new roles are needed, the name should clearly communicate a difference between the new role and the existing list or grid roles/ Succeeded: s/linked from the issue/layout examples are linked from the issue/ Succeeded: s/quoting spec:/carmacleod: quoting spec:/ Succeeded: s/now/not/ Present: jamesn Joanmarie_Diggs StefanS pkra carmacleod Scott_O Irfan Jemma Sina jcraig aaronlev mark_mccarthy sarah_higley juliette isabel Regrets: CurtBellew BryanGaraventa MelanieSumner MattKing JonGunderson Found Scribe: carmacleod Inferring ScribeNick: carmacleod Found Scribe: jcraig Inferring ScribeNick: jcraig Found Scribe: carmacleod Inferring ScribeNick: carmacleod Scribes: carmacleod, jcraig ScribeNicks: carmacleod, jcraig Agenda: https://lists.w3.org/Archives/Public/public-aria/2020Sep/0028.html 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]