W3C

- DRAFT -

SV_MEETING_TITLE

13 Jun 2016

See also: IRC log

Attendees

Present
AnnAbbott, jongund, JaEunJemmaKu, Bryan_Garaventa
Regrets
Matt, Leonie, Michiel
Chair
James Nurthen
Scribe
Birkir

Contents


<jamesn> https://rawgit.com/jongund/aria-practices/master/examples/treeview/treeview-1.html

<jamesn> https://rawgit.com/jongund/aria-practices/master/examples/treeview/treeview-2.html

<scribe> scribe: Birkir

Review updated tree pattern examples

<jamesn> https://rawgit.com/jongund/aria-practices/master/examples/treeview/treeview-1.html

<jamesn> https://rawgit.com/jongund/aria-practices/master/examples/treeview/treeview-2.html

Review updated widget from JGJemma and Nick

JG: the 3 of us are working on this, JG, AA and Jemma.

AA: I've been seeing a lot of presentational children discussion, should we put this on the backburner.
... IN this example a tree item owns other tree items.

Birkir: Probme may come up when calculating accessible name for a treeitem (when name comes from content).

JN: If we disallow tree items owning tree items it will cause major deofrestation (destroy tree widgets out there).

JG: Examples the items have no functionality.
... Eample 1 demonstrates there are actions associated with tree items.

Birkir: I can't tell if treeitem is actionable or just used to expand and collapse nodes.

JN: you can use aria-controls.

Birkir: only with Jaws/FF (no other combination announces aria-controls).

AA: There is nothing in the spec about aria-owns or controls.

JN: aria-controls is different, tree doesn't need it for its basic operation.

AA: I can see a developer going in and find the tree item does not apply (becuse the nodes are not actionable other than expanding and collapsing).

Birkir: CAn we expect all nodes in a tree to be actionable (other than the end nodes)?

Jemma: We can develop other examples where tree nodes are actionable (other than expand/collapse).

AA: Then I think the title is wrong (title="actions").

JG: There are actions, it updates the textbox.

Birkir: Usually the trees I ahve seen take you to a different page, or different part of same page.

JG: Could develop an example based on aria-selected (used a lot in filtering/bibliographic databases, where you can check and uncheck filtering items).
... If you check or uncheck a exampdable node, will it check/uncheck all its children, or do you ahve to check/uncheck them individually?

BG: I ahve a tree example where a checkbox affects all its children.

Birkir: How do you know if a treeitem is checkable?

BG: you can use aria-checked on tree items.

Jemma: Can you give us feedback on our examples please.

JN: I think we should wait until the discussion in the spec concludes.
... Safest way to do this is to put the treeitem role on the anchor tags.

Birkir: What about tree items that you can expand or activate?

<jemma> +q

BG: I have enver seen such a tree in practice, only the leafs ar actionable.

Jemma: Should we create widgets based on use cases, it is much easier if we work on a widget for a specific use case.

AA: It al comes back to the spec, the spec is not complete yet.

JN: But it is.
... Discussion is about presentational children, not about trees per se.
... What we got here is good, but it needs a minor little change to demonstrate the problems people are generally having.

BG: You can use the data-href (or data- anything) and a keypress.
... If you press enter on smething, you expect something to happen.

JN: What is the issue with it?

AA: I give up.
... You are not giving info to screen reader user.

BG: you can put a title attribute on treenode.

Birkir: I see 3 scenarios here. A tree item without aria-expanded I expdect activating it will cause an action. If I see an expandable tree item I expect that activating it will expand the node and put focus on the first node in the tree node. What I do not expect is when a tree item is expandable but it also has a different set of actions associated with it.

BG: We can use aria-haspopup on an expandable treenode and associate a menu of actions with it.
... This is why we proposed the aria-help property, which was shut down.

JG: Please provide feedback on the example documentation for the tree example.
... Let's look at "keyboard interaction and focus management".

JN: Seems appropriate.

<jamesn> https://www.w3.org/TR/html-aria/#sec-strong-native-semantics

JN: Why did you put all the aria-size and aria-posinset in?

JG: Because the browser does not calculate the set size for you, the spec says it "should", not that it "must>

JN: It is insane to have to do that, the browser knows how to do this.
... Would be great to add a button that removes the aria-setsize and aria-posinset attributes, so we can test it with and without.

JG: When you press arrow down and the node expands, if browser/a.t. does not notify user that the level changed, the user cannot figure out the structure.

Birkir: We need to be careful not to bloat our code with fixes for particular assistive technologies. If we code things according to spec, the browser or A.T. should do their job.

JN: How about we leave this in and conduct A.T. testing with and without the additional attributes.

JG: I will add buttons tha will toggle aria-level on and off, and aria-setsize/posinet.

<jemma> Synchronization of Visual and ARIA States

Jemma: REview the synchronization section (see URL).

<jemma> what do you think about "Synchronization of Visual and ARIA States"

JG: Should I put the treeitem role on the link?

JN: Yes, put role="presentation on the <li>s.

Birkir: If we use the <ul>/<li> structure, it is progressive enhancement, because if the scripting is not running, the hierarchy is still communicated.

JG: I got to run to another meeting.
... Please provide comments on the widget documentation.
... document on keyboard focus and synchronization. Avoid background images.

JN: I notieceddual keyboard focus rings.

Jemma: We can fix that.

JG: Should we remove all default action on expandable treenode elements?

JN/Birkir: no.

JN: The aria-level of the potted vegetables is wrong.

Review status of current example development work https://github.com/w3c/aria-practices/wiki/Design-Patterns-Status

Jemma: We are working on it, no updates.

JN: Can we take up the children presentational discussion.

BG: If you put e.g. role="group" on a bunch of cehckboxes A.T. announces it as a group. If children are presentational this role is not going to be announced.

JN: Treeitems have tree item children.

BG: If there is one tree item this works ok.

JN: Could change the spec.

rssagent make minutes

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.144 (CVS log)
$Date: 2016/06/13 18:22:13 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.144  of Date: 2015/11/17 08:39:34  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/ and co./Jemma and Nick/
Found Scribe: Birkir
Inferring ScribeNick: Birkir
Present: AnnAbbott jongund JaEunJemmaKu Bryan_Garaventa
Regrets: Matt Leonie Michiel

WARNING: No meeting title found!
You should specify the meeting title like this:
<dbooth> Meeting: Weekly Baking Club Meeting

Got date from IRC log name: 13 Jun 2016
Guessing minutes URL: http://www.w3.org/2016/06/13-aria-apg-minutes.html
People with action items: 

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.


[End of scribe.perl diagnostic output]