Web Events WG Voice Conference -- 11 Jan 2011

<scribe> ScribeNick: ArtB

<scribe> Scribe: Art

Date: 11 January 2011

Tweak the agenda

AH: I work for Opera in Sweden

<mbrubeck> sorry I'm late, dialing in now

AB: agenda posted yesterday ( http://lists.w3.org/Archives/Public/public-webevents/2011JanMar/0001.html ). Any change requests?

DS: I have some additions

<mbrubeck> That's me

DS: there is a proposal from Nokia to talk about
... there is another issue we didn't cover in D3E that we may want to capture here
... and that is mouse capture

AB: we can include them during the specs topic

Use Cases and requirements

AB: other than the UCs and requirements Cathy sent to the list last month ( http://lists.w3.org/Archives/Public/public-webevents/2010OctDec/0005.html ), there have been no additional UCs and requirements inputs. I encourage everyone to submit them.
... Cathy, are the UCs and requirements in your input addressed by the spec you submitted to the WG yesterday?

CC: yes, I believe they are

AB: ok; just wanted to clarify
... the requirements are required in later stages of the W3C recommendation process

Landscape wiki

AB: Doug and I added a couple of additional resources to the landscape wiki ( http://www.w3.org/2010/webevents/wiki/Landscape ). Everyone should consider that wiki as community property and edit/update accordingly.
... any other comments on this wiki?

CC: in the landscape is a link to Ilkka's proposal
... and it is same doc that I submitted to the list yesterday
... as such, the wiki should be updated to just point to the doc I submitted

<scribe> ACTION: barstow update the landscape doc so that Ilkka's spec is replaced with Cathy's spec of Jan 10 [recorded in http://www.w3.org/2011/01/11-webevents-minutes.html#action01]

<trackbot> Created ACTION-2 - Update the landscape doc so that Ilkka's spec is replaced with Cathy's spec of Jan 10 [on Arthur Barstow - due 2011-01-18].

Specs

AB: Doug, what's the status of your initial drafts?

DS: I created some initial drafts
... not quite ready to be checked in
... hope to get something to review by the end of this week
... I may use CVS rather than Mercurial

AB: that is fine with me

<shepazu> mouse capture: http://www.w3.org/Bugs/Public/show_bug.cgi?id=9557

AB: Cathy submitted an input to the list yesterday ( http://lists.w3.org/Archives/Public/public-webevents/2011JanMar/att-0002/TransformActionProposal.html ).

CC: the spec we are proposing addresses the representation level of user interactions
... physical touch events from UA to web apps is very low level
... our spec introduces higher level action such as zoom, pan
... have a transaction start event
... and then update events that include data like the action e.g. pan/zoom/etc.
... Allows authors to use high level events

AB: any intial questions or comments for Cathy?

DS: I think this is a good match for the existing transforms for SVG and CSS

CC: ok

MB: do you have a specific/concrete interaction that would trigger these events?
... e.g. 3 fingers on a touch screen

CC: that's up to the UA
... e.g. multiple touches, shaking device
... meant to be open to the impl

MB: have you doen any impl work with this?
... want to know how it works with the DOM

CC: we did some impl in Starlight project
... and it uses multi-touch
... my email to the list included a link

SM: scale and rotate are the same events

<mbrubeck> The link http://opensource.nokia.com/Starlight currently redirects to http://wiki.forum.nokia.com/index.php/Nokia_Open_Source

SM: wondering why they weren't separated

DS: I believe the intent is to allow any given gesture to comprise more than one compent
... could do more than one thing
... if scroll up and to the right, don't know if user is trying two things or one

<mbrubeck> If you think of using these events to implement a UI like http://hacks.mozilla.org/2010/08/firefox4-beta3/

<mbrubeck> then it makes sense that it would rotate and scale simultaneously.

<cathy> The link to starlight should be: http://starlight-webkit.org/Starlight/

MB: the example above does two things at once

DS: yes; this would scale and rotate
... so Matt, do you agree with that aspect of the TransAction spec?

MB: yes, I think this is a reasonable constraint

DS: someone asked if we are going to have literal mapping of 2 fingers on screen and one somewhere else and tie them to specific gestures
... we need to map between a defined gesture and an action such as zoom/pan/rotate
... it would be appropriate for us in our non-normative docs to outline specific gesture such as 3 finger down means something like zoom/pan/rotate

AB: thanks for clarifying that

Mouse Capture

<Sangwhan_Moon> http://www.w3.org/Bugs/Public/show_bug.cgi?id=9557

<shepazu> http://www.w3.org/Bugs/Public/show_bug.cgi?id=9557

DS: if you grab the scrollbar and move it up/down and then the mouse strays off the scrollbar
... or if you scroll down on a dropdown list
... We talked about defining this in D3Events
... and then thought Mouse Capture should be defined in this WG
... have been implemented in IE and perhaps Gecko

<mbrubeck> https://developer.mozilla.org/en/DOM/element.setCapture

<mbrubeck> "Introduced in Gecko 2.0"

<mbrubeck> (i.e. Firefox 4)

DS: Authors will want to do similar things for their own custom UI elements (widgets)
... the bug report is about setCapture and releaseCapture
... My intent, unless there is an objection, is to define setCapture and releaseCapture

AB: you would define them in one of the specs you already commited or a new spec?

DS: not a great fit for either one of them
... could go in either spec

AB: perhaps a new/separate spec

DS: it's more low-level
... so that is probably more appropriate

AB: is there any spec in W3C that defines these?

DS: no, I don't think so

<timeless_mbp> ack me yielding to Sangwhan_Moon

SM: this may have some security implications

<timeless_mbp> Zakim: ack me yielding to Sangwhan_Moon

SM: and may not work with platform widgets
... could have some interop problems
... may need a visual indicator so user doesn't get confused

DS: I think we can deal with this in the security section of the spec
... agree there is at least one security issue here
... If a widget overlays everything on the screen and events are sent to a snooping server, there could be a prob
... Think a security note for implementations would address the problem

<timeless_mbp> I'm concerned about security. But i don't object to it being added to Low Level

AB: does anyone object to setCapture and releaseCapture being included in Dougs's low-leve spec?
... adding something does not mean discussion is ended

DS: yes, adding it the spec just starts the conversation

RESOLUTION: group agrees setCapture and releaseCapture events will be added to the low-level spec

DS: in the charter, says an author can map low level events to high level intenstional events
... e.g. a double-tap can mean something special in my app
... and override the UA's default behaviour which could be zoom/pan
... We won't define specific gestures
... but could define a way to define gestures

AH: this would lead to device specific pages

DS: that can be done today; can't stop it

AB: so you want to know if this would be useful?

DS: yes, and if so, want to know if there are solutions

<timeless_mbp> so,...

AB: you may want to pose this on the list

<timeless_mbp> i'd like to let an app expose a way to tell users...

DS: don't think we need a decision now but we do need to take a stand

<timeless_mbp> "I have these actions available" <graphical-list>

<timeless_mbp> "You can select an action"

DS: Don't want to over engineer the problem

<timeless_mbp> "Please gesture to define how to trigger this action"

DS: but I think people will ask for something like this

<timeless_mbp> that would enable the user to train a local gesture to map to a high level custom event

AB: agree, in the abstract, something like that would be useful

DS: customization would be especially useful for Accessibility use cases

<timeless_mbp> Accessibility for disadvantaged users and users of disadvantaged devices :)

ACTION schepers ask the mail list for specific proposals re custom low-level to high-level event mapping

<trackbot> Created ACTION-3 - Ask the mail list for specific proposals re custom low-level to high-level event mapping [on Doug Schepers - due 2011-01-18].

AB: +1 to Sangwhan re tricky to implement

DS: agree; perhaps a v2 feature

<timeless_mbp> fwiw, i worry that we might not encourage vendors to do this, even though i think it'll be rather important for certain devices (esp legacy)

AB: anything else before AoB

DS: AoB = any other business

AoB

AB: next call?

DS: don't want a call until after I have something published
... schedule a call for next week and we will cancel it if I can't a spec out to review

<mbrubeck> I added a link to a useful starlight doc/spec page on http://www.w3.org/2010/webevents/wiki/Landscape

DS: who volunteered to Edit?

SM: I did

MB: me too

SM: I prefer Mercurial but with a doc, CVS is fine

<timeless_mbp> (I'm still on record as offering support for Hg)

MB: prefer Mercurial but can live with CVS

<timeless_mbp> technically Hg can help if you have someone (like me) making random minor cleanup bits

MB: lots more flexible for Mercurial

DS: let me dig into this a bit

<timeless_mbp> shepazu: what OS are you on?

AB: there could be less work with Mercurial because CVS requires public keys

<timeless_mbp> seriously?

<timeless_mbp> shepazu: http://jasonfharris.com/machg/

AB: ok, tentatively we have a meeting on January 18; will be canceled if Doug does not get a spec out before then

<mbrubeck> (some of my first programming experience was on BeOS)

AB: meeting adjourned

- DRAFT -

Web Events WG Voice Conference

11 Jan 2011

Attendees

Contents

Tweak the agenda

Use Cases and requirements

Landscape wiki

Specs

Mouse Capture

AoB

Summary of Action Items

Scribe.perl diagnostic output