WAI UA -- 24 Apr 2008

<scribe> Agenda: Chair: Jim Allan

Sctibe: Jan

<AllanJ> Scribe: Jan

<scribe> ACTION: JA to Look up accessible IRC [recorded in http://www.w3.org/2008/04/24-ua-minutes.html#action01]

1. Charter review with new updates

http://www.w3.org/WAI/UA/2008/draft_uawg_charter_20mar08.html

JA: JA, KF, and JR discussed some language

http://lists.w3.org/Archives/Public/w3c-wai-ua/2008AprJun/0063.html

<KFord> Proposed scope change to charter.

<KFord> 6 (NEW). Promote and support the development of user agent best practices and implementation guides for W3C technologies. Where appropriate, the resulting documents may be linked to by the UAWG or hosted by the Working Group as W3C Notes.

JA: Judy seemed to like it
... Any other issues?

KF: No prob with text but wanted to discuss ...
... Wonder if Aaron wanted us to do their doc

JR: I think he must know we couldn't...so he must mean we could host

GR: AL really pushing ahead on FF3
... So he's trying to get ARIA finished up

JA: I don't see that as part of the UAAG guidelines... they are more like techniques

<AllanJ> Jan: right, UAWG can provide space, or publish information as a NOTE

JR: So that's why I suggest it be a UAWG note

GR: Maybe we could follow a DHTML type model
... Taking place through xtech but doc is at dev.aol.com

dev.aol.com/dhtml_style_guide

GR: Becky Gibson and DOJO are also involved
... Talk about keybindings etc.

<AllanJ> dev.aol.com/dhtml_style_guide

Dean Hudson called in

DH: Work in voice over tech group in OS10, QA engineer, primarily responsible for vo quality on Mac platform...

JA: Glad to have you

KF: From Microsoft, work on IE team around accessibility

GR: Member of WAI-PF, HTML5, XHTML, Xfroms, vice chair for open accessibility WG

JR: Introduces self

JA: Web master at Texas school for blind and working with WAI since inception.
... I have an end user perspective
... Back to charter....

DH: I have looked a little bit

JA: We are jsut discussing new scope provision
... Any other comments?
... OK approved with no objections.

<scribe> ACTION: JR to Update charter and send to Judy [recorded in http://www.w3.org/2008/04/24-ua-minutes.html#action02]

2. Discuss/review ARIA implementer guide

JA: But all principal people not here

3. XHTML Access module

JA: In Agenda I included 4 messages...
... But GR is initiator of this

GR: XHTML2 working group is trying to push ahead a ACCESS module...defines standard way to access objects from keyboards
... I raised issue of ACTIVATE being boolean with no being default...
... But problem that more than one action can be associated with an object
... I proposed an "inspect" safety-type state that allows inspection of what focus will cause.
... Through discussion this week it's become clear that this involves UA interaction
... With ACTIVATE being boolean, with default no...ok for one action
... But what if some actions are activate=false and some true
... Some ATs can do inspect but what about people not using AT.

DH: What do you mean by activate

GR: Activate is like accesskey

JR: What about focus/activate split

<AllanJ> XHTML access module = http://www.w3.org/MarkUp/2008/WD-xhtml-access-20080418/

GR: Activate lets their be focus or not
... But mouse/pointer user is left out...

JA: THought activate was on a mouse click

<AllanJ> JR: example - OK button, can tab to get focus, with mouse can only activate

<AllanJ> GR: access module is specific to access key, and the author can choose whether the key press results in focus or action

GR: So as long as fixing it for keyboard....want to fix mouse as well
... Access can be defined for any target
... But we are dealing with potentially multiple events
... Since user agent has control over DOM, maybe user agent has to step in to show the handlers and offer.

DH: But user may not know which handler to choose

JA: We currently have checkpoint, "activate handlers"...
... And user has to be able to fire all actions with keyboard
... And so you're saying this should be different...with an inspect?

GR: Either have to do it in Access module or XML events2

JR: What's the user experience

<AllanJ> JR: practical user experience. with a pointer - mouse down, mouse over, the user doesn't know what will happen.

<AllanJ> JR: what's the purpose;

<AllanJ> GR: to defang the on-focus firing.

JR: I understand ensuring focus doesn't happen
... What do you mean by inspect?

GR: If multiple actions fire user may want to do some but not all

KF: So an example...on_mouse_down to activate link
... THen I have mouse_over

GR: Say you have extension where mouseover word gives translation and mousedown option to change language

JR: So you just know there is mousedown and mouseover but no meaning till code fires?

<AllanJ> the UA does not know what will happen.

GR: I won't see this using "read all"

DH: Maybve should be in ACCESS module
... Access module could send message to AT
... Another keypress could say press this to get menu etc

JA: What you say is true but if some extension to user agent then UA should know what is going on, but user's jscript will be blackhole
... UA doesn't jknow what will be happening
... Other issue is user should not have to self-discover a la Myst...UA should tell you when you get there

GR: This is why I always have verboisty on high - let's me discover more
... COuld be sound, alert, slashing screen etc.

DH: Right

GR: I think there should be way in Access module to get cascade
... Because even though accesskey is defined as one key from keyboard ...it might not be avaialble on a particular keyboard...going to have problems
... Access module says the keystrokes are only suggestions of UA...

<AllanJ> http://lists.w3.org/Archives/Public/w3c-wai-ua/2008AprJun/0061.html

JA: Just thinking about more studff in our guidelines....
... 9.5
... Then also 9.6

GR: So at very least they need to put pointers in to UAAG

JA: How should we work to liaison

GR: I really appreciared your message on the board
... Other thing we need to make sure is addressed...in absence of defined key, user agent SHOULD define a key
... User agent should display bindings....

<AllanJ> UAAG 1.0 11.3 allow the user to override any binding that is part of the user agent default input configuration

DH: And that would be in access module?

GR: So in there already is that the user agent should define a key

JA: UAAG is at the point where whatever author says fine, but allow reconfiguration
... We also have a P3, no automatic form submission

GR: THey want this to be abstract as possible but we do want to point out diff between mouse user and keyboard users experience

JA: When hit access key on top of page, something could happen on botom of page

DH: Interesting...mouse users experience this...go to web site, don't know loud music will start
... Lots of mouse users do sort of click around all over

GR: Mouse user is free agent
... Keyboard user is not, always restricted....

JA: Kelly?

KF: Absorbing.

JA: OK, we have 3 mins
... We will be picking up keyboard conversation again...
... So for next week...

<scribe> ACTION: JA to Summary of the ACCESS stuff - mapping relevant checkpoints in UAAG1 and UAAG2 [recorded in http://www.w3.org/2008/04/24-ua-minutes.html#action03]

JA: Because I think we cover a lot ofg this
... Then we'll try and get consensus and senfd over to XHTML

GR: Could be UAAG2 verbiage with UAAG1 mappings for normative power

JA: OK, well Dean great you could join us
... OK see you all next week

- DRAFT -

WAI UA

24 Apr 2008

Attendees

Contents

1. Charter review with new updates

2. Discuss/review ARIA implementer guide

3. XHTML Access module

Summary of Action Items

Scribe.perl diagnostic output