SV_MEETING_TITLE -- 02 Oct 2008

<jallan> title: UAWG telecon

<KFord> tractbot, start telecon

<KFord> Chair: Jim_Allan

<KFord> Scribe: KFord

<jeanne> http://lists.w3.org/Archives/Public/w3c-wai-ua/2008JulSep/0148.html

F2F Agenda

<jallan> http://www.w3.org/WAI/UA/2008/10f2f#Agenda

Jeanne: Should we include other time zones.

<scribe> ACTION: Jeanne to update F2F with GMT times plus local times. [recorded in http://www.w3.org/2008/10/02-ua-minutes.html#action01]

<trackbot> Created ACTION-25 - Update F2F with GMT times plus local times. [on Jeanne Spellman - due 2008-10-09].

<sharper> http://www.timeanddate.com/worldclock/meeting.html

KFord: Any thought of splitting document review into two days?

JBrewer: Jeanne, can you update document with review priorities we wanted to use?

<scribe> ACTION: Jeanne, update document review sections with review priorities. [recorded in http://www.w3.org/2008/10/02-ua-minutes.html#action02]

<trackbot> Sorry, couldn't find user - Jeanne,

Jeanne: Suggest to use afternoon for issues time.

JAllan: I suggest we put document review to mornings and technical to afternoon.

<scribe> ACTION: Jeanne update agenda to reflect change in plans. [recorded in http://www.w3.org/2008/10/02-ua-minutes.html#action03]

<trackbot> Created ACTION-26 - Update agenda to reflect change in plans. [on Jeanne Spellman - due 2008-10-09].

Jallan: we don't have keyboard on F2F. Can we power through remianing items.

Keyboard access

SHarper had proposal of six rewrites to keyboard override.,

<jallan> 4.1.xx Simon rewrite http://lists.w3.org/Archives/Public/w3c-wai-ua/2008JulSep/0149.html

Jallan: Jan indicated on mail list he liked option 1.
... I agreed this was the best.

Sharper: Can someone explain the recognized part?

JAN: It means that the user agent has to know this is a key binding. For accesskeys it knows what to do. For AJAX controls/javascript, the browser will likely not know what this key is for e.g. data entry, command.

Jallan: Further explanation, there are black holes where the user agent doesn't know what the key is for, again AJAX and such.

SHarper: This makes sense, my thinkin g was to start with accesskey. But do we need to know more than this. But do we need to do more.

For example, if ctrl+a is pressed, and the user has mapped to ctrl+k, does it really matter.

SHarper: the key was still pressed so the user wants this to be ctrtl+k.

JAN: Restates SHarper explanation, talking about at a higher level, process the keyboard action before anything else happens. Doesn't matter why.

Sharper: My last suggestion was to reflect this.

JAllan: I'm not sure if this is correct but my recollection from last F2F was that scripts, then accesskey then chrome gets keys. Not sure we can do anything about this.

JAN: I see what you are saying but UA could still do this. This would be just a top level mapping.

<jallan> KFord: browser makes choices of where keys go.

<Jan> KF: I need to check but in IE, we always take alt-D...before AJAX

<Jan> KF: I'm not 100% but I know we spent lots of time making this work

JBrewer: IF WaI were to issue guidelines that siad change what you've learned, probably wouldn't fly.

<Jan> yes

<Jan> Scribe: Jan

SH: JR is right...I'm remapping at high level...alt D can still work
... Not even saying what needs to pick things up first
... Just a matter of what keys I want to press to activate what's there

JA: While you were talking I wish Al was here...
... Lots of keyboard mapping ATs out there.
... Maybe we are asking too much of base browsers

JB: Triple A?

JA: THink it's double A

JB: I'm concerned about the increasing complexity here
... Weren't we also going to check with PF on this
... To make sure we weren't going to collide with them

JA: I remember proposal...but no action item was created

JR: why a collision?

JA: PF working on access key stuff

<KFord> JAN: We were talking about accesskey mapping and we were talking at a higher level than content. We are talking before keys are used for anything.

<scribe> ACTION: JA to Talk to PF about the issue of key remapping [recorded in http://www.w3.org/2008/10/02-ua-minutes.html#action04]

<trackbot> Created ACTION-27 - Talk to PF about the issue of key remapping [on Jim Allan - due 2008-10-09].

<KFord> SHarper: Yes, this is what I'm talking about.

SH: Not specific to access keys

JA: But it is specific

<KFord> JAN: I think in UAAG 1 there was a separate checkpoint addressing key mappings for the browser frame. This is a way to combine these.

JR: In old UAAG 1.0...there were weparte checkpoints for remapping keys to chrome and keys to content

<KFord> KFord: I will also talk to one of our developers here.

<KFord> JAllan: I will send to the group and then send to PF. Let's table this for now.

<jeanne> http://www.w3.org/WAI/UA/2008/WD-UAAG20-20080925/WD-UAAG20-20080925.html

<jeanne> http://www.w3.org/WAI/UA/2008/WD-UAAG20-20080925/WD-UAAG20-20080925.html#principle-operable

<jeanne> http://www.w3.org/WAI/UA/tracker/actions/open

<KFord> KFord, interesting phone echos, cats and other such items.

http://lists.w3.org/Archives/Public/w3c-wai-ua/2008JulSep/0127.html

<judy> http://www.w3.org/WAI/UA/tracker/actions/open

<KFord> Meeting pauses to review open action items at above URL and close as appropriate.

<KFord> ACTION: Jeanne update agenda to include breaks/lunch. [recorded in http://www.w3.org/2008/10/02-ua-minutes.html#action05]

<trackbot> Created ACTION-28 - Update agenda to include breaks/lunch. [on Jeanne Spellman - due 2008-10-09].

<KFord> kford: Review of action items concludes.

<KFord> Jallan: Now talking about 4.10 and 4.X.

<KFord> JAN: We talked about this before. The new issue here was to drop within and the new thing was to go within groups.

<KFord> ACTION: Jallan review 4.11. [recorded in http://www.w3.org/2008/10/02-ua-minutes.html#action06]

<trackbot> Created ACTION-29 - Review 4.11. [on Jim Allan - due 2008-10-09].

4.2 Provide access to event handleers.

<jallan> 4.2.1 All Available: The user can activate, through keyboard input alone, all input device event handlers (including those for pointing devices, voice, etc.) that are explicitly associated with the element designated by the content focus.

<KFord> me Jim are you pasting this?

<KFord> KFord: KFord today I don't know of a user agent that really does all of this. Further explanation.

<KFord> JAN: There are some practical problems with this. Gave example of complex mouse interactions.

KF: When I think about this...problem today...2 categories...supplemenetal functionality...
... User agent lets you get to link with TAB but not the mouse actions...
... ex. imagine I have website...I write some HTML...click here to go to my Free offer
... In script onclick go here...makes a link
... Keyboard user locked out
... Could we argue that UA should allow access to this
... This is already saying this to some extent....but I'm talking about onclick
... But supporting all of the other actions would be too much

JR: So you are saying onclick is best bang for the buck

KF: Yes
... Thinking about what's practical

JS: Think we could approach...user agent should "repair"...

<KFord> Jeanne: I think we should could approach this from the approach of repair.

JA: This one is really further expansion of 4.1.1...all functions should be avaialble from the keyboard

<KFord> \JAllan: I think this is just a narrowing of our top level requirement that all functionality be available from the keyboard.

KF: Not realistic always...because mouse has different interaction model
... But on flip side not saying I want to lock users out
... I'm not a mouse user....
... So what are things you do with mouse that are actually too tedious for keyboard users...maybe answer is don't limit people.

JB: One behaviour is significant level of muscle weakness/fatigue...becomes tiring to move hands alot or do a repeated motion...
... Condition can be extremely variable...sometimes even hard to move from mouse to keyboard
... Irony then is it is best to do all with keyboard or all with mouse
... THink this is sometimes misunderstood

KF: OK, but at what point would you think an AT should come into play vs. base funtionality
... e.g. all OS's have mousekeys
... We aren't asking for screen readers for eg.

JB: I think every command function should be interoperable with AT....no question

KF: Assumption that screen readers are mature enough to demand interoperability from UAs

JB: Prob with that this APIs aren't completely in place.

KF: Here's how you might write this (in extreme): you have to be able to do everything you can do with mouse with the keyboard or support the platform's AT interoperablity mechanism
... Doesn't have to build it in directly

JB: Isn't it already the expectation that applications should support existing on-board access features such as mousekeys?

KF: Not today...I don't think it's ok to say I don't need to do that because OS has mousekeys

missed JB's clarification

KF: I think if one of 2 ways works, you've satisfied basic level of access
... Works like "all" make it hard to see what UA should do

JA: I've been enjoying discussion...valuable
... Yeah we can fall back to mousekeys...but then we're missing out on effeciency
... Also think there is a host of things...e.g. tooltips or title tags ...which I think are events...can't come up with keyboard...
... I think that hampers keyboard users
... At the same time, it's a crossover issue...WAI ARIA...
... And also WCAG2

KF: OK...bit of tangent...I'm a fan of ARIA...but it opens lots of new problems...
... ARIA is dangerous...
... In 2 seconds I can call any peice of HTML a new role....
... So it's entirely implicit on author to make things behave like trees....
... And guess what...it won't act like a tree most of the time

JA: Right and outside TAB tree

KF: Don't want to be a fear monger
... I've made the same points in other places
... But concerns me greatly
... Same thing as custom controls in Win32

JR: Hope is a toolbox like DOJO will do all the work

KF: Not sure about that...I see examples of companies trying to be on leading edge...doing their own thing
... Part of me says hey user agent...if that thing was called a tree, make it work...but of course that's not practical\

JA: All these are relevant issues
... But back to 4.2.1...

<KFord> Jallan: Does our wording work?

<KFord> JAN: I don't think you can just say fire them all.

<KFord> JAN: Gave example of complex mouse events again.

JA: To me...maybe we needto change wording....but onmouseover same as onfocus is more like a technique

<KFord> Jallan: I think saying things like onmouseover should move focus and such needs to be in a technique.

<KFord> JAN: Need some technique to have some wiggle room.

JR: So some instead of all input device handlers

KF: What does the section 508 redraft say?
... Most access guidelines say it like this...but then we redefine somewhere else and say all doesn't mean all.

JB: I'd have to check 508.

JA: If I have some item...more than one action can fire

KF: I can build a whole flyout menu...
... But I would argue that content author has failed already

JA: I see it's endd of our time

JB: Interesting discussion
... Anyone want an action item on 4.1.2?
... Anyone want an action item on 4.2.1?

<KFord> ACTION: kford, JAllan to revisit access [recorded in http://www.w3.org/2008/10/02-ua-minutes.html#action07]

<trackbot> Sorry, couldn't find user - kford,

<KFord> ACTION: KF, Jallen to revisit keyboard access and definiton of all and such. [recorded in http://www.w3.org/2008/10/02-ua-minutes.html#action08]

<trackbot> Sorry, couldn't find user - KF,

<KFord> Title: W3C UAG 10/2/2008

- DRAFT -

SV_MEETING_TITLE

02 Oct 2008

Attendees

Contents

F2F Agenda

Keyboard access

4.2 Provide access to event handleers.

Summary of Action Items

Scribe.perl diagnostic output