[aapi] UAI TF Meeting Tue 25 Feb 2014 from Joanmarie Diggs on 2014-02-25 (wai-xtech@w3.org from February 2014)

From: Joanmarie Diggs <jdiggs@igalia.com>
Date: Tue, 25 Feb 2014 16:08:28 -0500
To: "wai-xtech@w3.org" <wai-xtech@w3.org>
Message-ID: <530D064C.5050500@igalia.com>

Link: http://www.w3.org/2014/02/25-aapi-minutes.html

Plain text follows:

[1]W3C

[1] http://www.w3.org/

- DRAFT -

Protocols and Formats Working Group Teleconference
25 Feb 2014

See also: [2]IRC log

[2] http://www.w3.org/2014/02/25-aapi-irc

Attendees

Present
Joseph_Scheuhammer, Bryan_Garaventa, Joanmarie_Diggs,
Rich_Schwerdtfeger, [Microsoft]

Regrets
David_Bolter

Chair
Joseph_Scheuhammer

Scribe
joanie

Contents

* [3]Topics
1. [4]Having a core UAIG, one specific to HTML, and one
to SVG. Also, when to start edits to the UAIG 1.1 core
document.
* [5]Summary of Action Items
__________________________________________________________

<trackbot> Date: 25 February 2014

<clown> agenda: this

<bgaraventa1979> Bryan Garaventa

<richardschwerdtfeger> what is the call in passcode?

<clown> richardschwerdtfeger: 2274#

<clown> [6]http://www.w3.org/WAI/PF/aria-1.1/

[6] http://www.w3.org/WAI/PF/aria-1.1/

<scribe> scribenick: joanie

Having a core UAIG, one specific to HTML, and one to SVG. Also, when
to start edits to the UAIG 1.1 core document.

RS: We have an outline we were going to do....

<richardschwerdtfeger>
[7]http://www.w3.org/WAI/PF/wiki/Outline_Core_User_Agent_Implem
entation_Guide#Core_User_Agent_Implementation_Guide

[7]
http://www.w3.org/WAI/PF/wiki/Outline_Core_User_Agent_Implementation_Guide#Core_User_Agent_Implementation_Guide

RS: The bulk of the 1.1 Implementation Guide would go into the
core
... What we can do is create implementation guides for HTML 5.1
and SVG 2 based on the core
... Instead of having to go back and duplicate role mappings,
we can refer back
... There are features in the host language we'll have to
define there
... Examples are name and description computations

<richardschwerdtfeger>
[8]https://svgwg.org/svg2-draft/struct.html#implicit-aria-seman
tics

[8] https://svgwg.org/svg2-draft/struct.html#implicit-aria-semantics

RS: In the document structure section of the specification, we
define how ARIA integrates into the specification
... We also have a new host language semantics section
... In each of these we have a defined role and how it gets
mapped into the accessibility API
... The other thing we are doing, for those elements which can
be mapped, we try to say they have a presentation role unless
something else gets added to them to cause them to be mapped
... Because everything in SVG is persistent, we don't want to
map everything to the tree unless we have to
... When do we create a role in the accessibility tree is going
to be very important

<richardschwerdtfeger> presentation role provided no associated
‘title’ element, ‘desc’ element, ‘aria-label’ attribute,
‘aria-labelledby’ attribute, or ‘aria-describedby’ attribute;
otherwise, group role

RS: Beyond that, there are no restrictions
... If you put some sort of text inside a circle (e.g. title,
label) it would get mapped to group.
... We are going to be defining host language semantics
... We can then point to the core specification
... The HTML and SVG DOM stuff are all getting merged
... EVen the event handlers
... Other interesting things about SVG are
... You can embed an iframe in SVG, also a canvas element
... I have to think about what that means
... Is the fallback HTML or SVG?

JS: So the canvas element is the same?

<clown>
[9]https://svgwg.org/svg2-draft/embedded.html#CanvasElement

[9] https://svgwg.org/svg2-draft/embedded.html#CanvasElement

RS: That's correct.
... The good thing is we have one DOM and a common set of
attributes.
... The question I have is, does that make sense?
... Are we going to take ARIA 1.1 and add it to the core?

JS: Yes
... We already have a promise to the GNOME Accessibility Team
... I want to get that going
... That draft (ARIA 1.1) could start out as the core
... Then you could start taking things out or whatever you need
to do

RS: I like the way the HTML and SVG specs expand and collapse
elements
... Are you going to do that?

JS: I wanted to get that done for 1.0
... But the script doesn't work with the table in the UAIG
... The table markup is slightly different

<richardschwerdtfeger>
[10]http://www.w3.org/TR/2012/WD-html-aapi-20120329/

[10] http://www.w3.org/TR/2012/WD-html-aapi-20120329/

CS: (Confirms what Jospeh stated)

JS: The document you put in Rich looks like what I was refering
to

RS: I can expand them but not collapse them

JS: I should ask Jason about this

RS: I don't think you can expand or collapse them with the
keyboard

(Discussion on how there is an option at the top to toggle all)

JS: I think they are using re-spec(?sp)
... You write the spec with staright HTML in one file
... And you include this re-spec javascript and it does all the
formatting, etc.
... What we do for the ARIA documents...
... We have a core document and an XSLT processor
... And it puts everythign together, adds table of contents,
creates a full HTML document

RS: Does anyone have the source for it?

CS: You'll have to ask Jason

JS: I have the javascript. It's in the document itself.

RS: We need to get the original

JS: It's probably in W3C's mercurial repo

CS: Jason will know and Steve Faulkner will know

(Discussion about this might not being the right/same document)

<richardschwerdtfeger> [11]http://www.w3.org/TR/html-aapi/

[11] http://www.w3.org/TR/html-aapi/

<clown>
[12]http://rawgithub.com/w3c/html-api-map/master/index.html

[12] http://rawgithub.com/w3c/html-api-map/master/index.html

RS: I didn't know you could put W3C docs on github

CS: Anyway, email Jason and Steve

JS: This one works with the keyboard

RS: We definitely want this one
... So who is going to be the editor of the HTML 5.1 one?

CS: I just asked for that to be put on the Task Force agenda
... Steve doesn't have time, Jason doesn't have time
... We need people

RS: I'll talk to Steve about this
... We need to figure out who is going to edit this
... I don't have bandwidth for this. I might to SVG

CS: I have bandwidth to provide details about how IE does
things

JS: I have bandwidth to work on the core right now

RS: Will we have the UIA for 1.1?

CS: I cannot answer that yet

RS: They have a single table too

JS: When I looked at it you gave it the ID of the table element
and it would do that table
... So you could do it for multiple tables

RS: Given that most of this will be in the core spec....
... This should go in the core, right?

JS: The HTML one?

RS: The table mapping for the APIs
... What they have here is like anchor
... It goes to a link role

JS: To you want this anchor list to be in the core?

RS: No
... You are going to have an equivalent in the core spec
... All the API mappings will be

JS: What they are now

RS: But it wouldn't be in the HTML spec

CS: It would be in the HTML mapping guide because HTML and ARIA
aren't the same thing

JS: I think Rich is saying you'd put a link to the core

(Discussion on specific examples to clarify the point)

CS: I disagree that mapping from HTML to accessibility APIs
should have to go through ARIA

RS: But that is what we are doing
... This is needed for SVG
... I don't know if we'd call it "ARIA"
... But it doesn't make sense to duplicate it
... It makes sense to have it in one place

CS: But if I'm implementing HTML, why do I have to look in two
specs?

RS: Trying to keep them in sync is a nightmare
... That is why we want to have one core spec that define
things in terms of ARIA symantics
... What we want to do is have a core implementation guide that
defines the mapping for those symantics in it
... We'd refer back to the core spec where we can
... But not with things like name computation

CS: What about places where HTML and something else has a
mapping, but ARIA does not?

RS: Then we'd have to take the HTML version

JS: You could populate the HTML version by pulling the mapping
out of the core, e.g. via AJAX
... Use the UAIG core as the database

RS: Yeah, you could do that

JS: As you load the mapping table, it fetches content (done in
the browser)

CS: One of the things I'd like to do is see what features are
missing from my API
... Having a place where they are all together would be helpful

RS: Did you file the bugs after we went through CR?

CS: Yes
... I want to be able to explain both to IE and authors is the
relation between HTML and UIA
... The levels of indirection makes that harder

RS: I don't think there are that many levels of indirection

CS: Having a view that puts this together would be very helpful
... I think the script Jason has would be able to do that

(more discussion about the above)

CS: Are you ok with us owning this html piece as well

JS: Who is "we"?
... I'm not able to be the editor

CS: I'm trying to get the same level of productivity?

JS: Would it be part of this call?

CS: (confirms)
... We'd have to find another editor

JS: We can ask

RS: I'm not sure if David will have time
... It sounds like we need a call with Jason

CS: I got some feedback from Jason about where things stand now
and his availability
... There was a timezone issue in the past (e.g. Toronto,
Siberia, New Zealand)

RS: So we just need editors for HTML 5 and the other one

JS: I'm willing to try to bring the HTML 5 issues to this call
here
... But I am afraid we will run out of time
... We can try

RS: We will have to coordinate things like name computation

JS: I think 90% of name computation is core

RS: It's going to be different for HTML 5 and SVG
... (gives example)

<clown>
[13]http://www.w3.org/WAI/PF/aria-implementation/#mapping_addit
ional_nd

[13]
http://www.w3.org/WAI/PF/aria-implementation/#mapping_additional_nd

RS: It's called placeholder

<richardschwerdtfeger>
[14]http://rawgithub.com/w3c/html-api-map/master/index.html#inp
ut-type-text-input-type-password-input-type-search-input-type-t
el-input-type-url-and-textarea-element

[14]
http://rawgithub.com/w3c/html-api-map/master/index.html#input-type-text-input-type-password-input-type-search-input-type-tel-input-type-url-and-textarea-element

CS: How does that figure into name calculation
... It's an aria-hint as far as I'm concerned

RS: (points out mobile use case)

CS: If there's not another label, the placeholder would go into
the name

CS: If an image has both an alt and a title, then the alt is
the name and the title is the description
... It would be similar with placeholder

RS: But this doesn't tell with the text before and after

JS: Is the above how placeholder works?

CS: Yes

RS: It's meant to be for the label
... (looks for form example)

CS: In the example from the HTML spec it's got "sample text"

RS: Where they really started using this first is Apple
... You'd put "city" and when you started typing that text
would go away

<cyns> The html spec example says

BG: Is the placeholder text supposed to disappear when it gets
focus?
... Or when you start typing?

RS: Basically what happens is that they put the text label in
the input field as grey text
... This is to save space
... Then you don't need to put an actual label
... When you start typing, the grey text goes away

CS: And sometimes that's a bad experience for sighted users too

RS: I'm not trying to defend it; I'm just describing what it's
been used for

CS: (Refers to text example from the spec she pasted in above)

<richardschwerdtfeger>
[15]http://www.w3schools.com/tags/att_input_placeholder.asp

[15] http://www.w3schools.com/tags/att_input_placeholder.asp

RS: They put first name and last name
... (said in reference to the example he pasted in)

JS: They're trying to save space

RS: That is what I was saying

JS: You cannot select the text

RS: It's considered a vanishing, static label

(More discussion about this user experience and use cases)

RS: The fact that it's in the name computation must give it
credence

CS: It's similar to the title attribute

JS: Firefox is setting the accessible name to the placeholder
text in that case

CS: That's what we'd do too, I think. But it's the name of last
resort
... Before there was placeholder, this was done via value and
CSS

RS: The question is, do you put this in a hint and also in a
name if nothing else is there?

JS: James Craig is writing the precedence as one of his 1.1
tasks

RS: I guess I can be the SVG editor, but I'd like to get some
eyes on this

CS: Are the HTML working group going to be ok pulling this into
core?

RS: I don't know if we've formalized it, but all the editors
working on that spec agreed this is the way to go
... We haven't agreed on who is going to work on what yet.

JS: To wrap up....
... In order to do anything we (or I) need help from Michael
... To find out how to fork the documents
... And how to do this work without interfering with what he is
doing
... Otherwise we are stalled

RS: Was he supposed to be on the call?

<cyns> CS: as long as there is a way to see all, I am ok with
breaking this into core/html/svg. We have had trouble with
staffing and progress on the HTML doc. pulling things into core
would make html - api mapping smaller, which might help with
staffing.

JS: I was hoping he would be. Before Christmas he was. Since
then he's been busy.
... I will ask him how we can go ahead.

RS: I am going to send him a note now.

JS: Let's call it the end of meeting.

Summary of Action Items

[End of minutes]
__________________________________________________________

Minutes formatted by David Booth's [16]scribe.perl version
1.138 ([17]CVS log)
$Date: 2014-02-25 21:03:43 $
__________________________________________________________

[16] http://dev.w3.org/cvsweb/~checkout~/2002/scribe/scribedoc.htm
[17] http://dev.w3.org/cvsweb/2002/scribe/

Received on Tuesday, 25 February 2014 21:09:09 UTC