W3C

- DRAFT -

Pointer Events Working Group Voice Conference

07 Apr 2015

Agenda

See also: IRC log

Attendees

Present
Art_Barstow, Matt_Brubeck, Asir_Vedamuthu, Scott_Gonzalex, Olli_Pettay, Rick_Byers, Tim_Dresser, Mustaq_Ahmed, Patrick_Lauke, Jacob_Rossi, Doug_Schepers
Regrets
Sangwhan_Moon
Chair
ArtB
Scribe
ArtB

Contents


<scribe> ScribeNick: ArtB

<scribe> Scribe: ArtB

<jrossi2> I'll be a bin late... my MSFT credentials expired while on vacation #doh

<patrick_h_lauke> i'm calling in in a sec, but will mostly be on mute due to childminding

<smaug> Hello

<rbyers> Zakim rbyers is Chrome_Team

<patrick_h_lauke> i'm in

Agree on Agenda

AB: yesterday, Jacob proposed a draft agenda and it looks good to me <https://lists.w3.org/Archives/Public/public-pointer-events/2015AprJun/0002.html>. Any change requests?

[ None ]

Implementation Status

AB: besides Google's recent annoucement (f.ex. see <https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/ODWmcKNQl0I>), would like an update from Microsoft/Spartan and Mozilla.

RB: my blink-dev thread didn't include a lot of context

… so I can give some context now

… and take feedback too.

<rbyers> RB: 4 main reasons reasons

<rbyers> ... 1. interoperability ecosystem: interop benefit of focusing on touch events was no longer outweighing the costs (including partner/developer good will)

<smaug> (there is some old CSS proposal for tactile feedback)

<patrick_h_lauke> so, for first time, apple now have a multi-input type scenario and realising it's potentially giving problems...hmmm

<rbyers> … 2. scroll-start performance: blink's focus in 2014 was appy sites where a single UI thread was preferable (lots of touchmove driven effects interacting with animation etc.). In 2015 we've shifted focus back to sites that depend on threaded scrolling.

<rbyers> s/... 2/RB: 2/

<rbyers> RB: 3. richness concerns being addressed: scroll customization at houdini is the right path for p2r. Touch-action extensions as a tactical mitigation

… (p2r = PullToRefresh)

<rbyers> RB: 4. compat issues being addressed: mouse event model for touch

AB: thanks Rick. Does anyone have any Qs or comments about these 4 points?

AV: re point #4, is there something specific?

RB: IE has made some compat changes

… we need to tighten the spec

<patrick_h_lauke> new spartan behavior: see last row in http://patrickhlauke.github.io/touch/tests/results/#desktop-touchscreen-events

JR: in a future Spartan build there will be a flag to turn on/off the diff models (touch and pointer)

… we are still collecting data

<patrick_h_lauke> compared with IE11 which sends mouse compat events interleaved with their pointer equivalent

… and the group still needs to work out the details

RB: need to know where we have good interop and then update the spec to address the problem areas

DS: would you please explain Intent to Implement vs. Intent to Ship?

RB: in Blink process, don't do I2Ship until code is done or very close to done

<rbyers> Blink process: http://www.chromium.org/blink

… thus requires lots of impl work

<rbyers> http://www.chromium.org/blink#launch-process

DS: so does I2I kind of imply a future I2Ship?

RB: yes, normally that is the case

… the announcement serves the purposes to gather lots of broad feedback

… some people will evaluate the impl from a performance perspective

… and certainly interop

AV: is there any target for shipping?

RB: too soon to say

… we do a quarterly planning process

… and a lot of that is shifting to be public

… now doing Q2 planning; I can share some of the informaiton

… expecting this to be multiple quarters

AV: is there a link to the quarterly info?

RB: this is the first time we are doing this

… nothing available yet but I'll let you know when it is

DS: will there be something at upcoming Google I/O?

RB: no, unfortunately I don't think we will be far enough for this year

… clear benefit for performance

… as well as stylus support

… This work is important but we won't rush it.

… We need to balance a lot of priorities.

… Getting PE working for mouse events in WK is going to be a lot of work.

… Must be careful that every change doesn't break some web site expecting WK behavior

… Lots of iteration

<asir> Thank you Rick for sharing these info

JR: Spartan has PE support now

… including PE Constructor (which isn't in IE11)

… we have some flags included

… Want to align our impl with other browsers like Blink and Safari

… f.ex. pan and some other TE model changes

… [ Jacob provides some other implementation details ]

… Also enabled gesture based mouse model (up, down, etc.)

… Still collecting lots of data

… Going forward we will work on issues like those Rick identified

… think they are all solvable with PE

<patrick_h_lauke> jacob, would be happy/interested to help test/see some of the example sites that break

AB: are these details documented in Public?

<jrossi> http://blogs.msdn.com/b/ie/archive/2015/02/24/pointer-events-w3c-recommendation-interoperable-touch-and-removing-the-dreaded-300ms-tap-delay.aspx

JR: yes, http://blogs.msdn.com/b/ie/archive/2015/02/24/pointer-events-w3c-recommendation-interoperable-touch-and-removing-the-dreaded-300ms-tap-delay.aspx

AB: anything Matt or Olli re implementation?

MB: think the blink-dev annoucement will make this more urgent

Scrolling implementation consolidation will help to support pointer events on all the platforms

<mbrubeck> Implementation work on Firefox/Gecko is continuing apace

v.Next Features and v1 Errata

AB: we have the v.next feature list <https://www.w3.org/wiki/PointerEvents/UseCasesAndRequirements#Requirements:_Pointer_Events_v.Next_Specification>. There is also one open bug <http://tinyurl.com/Bugs-PointerEvents>.
... Jacob included 3 items in his draft agenda: 1) "up/down/left/right" pan values for touch-action; 2) implicit/explicit capture 3) compatibility mouse events model. Any comments on these items or proposals for additional items?
... It would probably be good for us to get consensus on the scope of v.next but it might be a bit premature right now.

JR: with respect to the charter, I think the scope is the same

… I don't think we or Rick mentioned anything beyond the current charter

… the items Rick identified should be on the table

RB: for hit testing, implicit vs explicit capture; also over/enter/leave during capture

… capture behavior must be well spec'ed

<rbyers> This thread: https://lists.w3.org/Archives/Public/public-pointer-events/2015JanMar/0041.html

<rbyers> ... we want the default behavior not to incur per-move hit-test costs

RB: I've talked to a lot of people including Android team and think we've identified the main issues

AB: anything else from Mozilla's perspective or have Rick and Jacob identified the main issues?

JR: given the items, we might need more than 6 months

… for a v2 spec, 6 months is probably not enough

RB: will definitely need some compat data for some issues

… it's likely impl work won't be complete within 6 months

<patrick_h_lauke> i've got a very happy screaming 2.5 year old running around, so probably missed if it's been talked about, but: what was the decision about mouse compat events and when they fire (in the PE spec vs in real life Spartan)? do we plan to already work on errata/update for PE spec? is this for v2/v.next?

DS: then need to think about a 6-month extension to a new charter that extends 12-24 months

JR: I am indifferent re extension vs. new charter

… I think the charter itself is the same, just need to add v.next/v2

DS: we can figure out the charter parts

… more important to do the right thing re the spec

… we will get a 6-month extension

AB: I agree with Doug re the charter -> let him and me figure out the best way and keep the group focused on addressing the technical issues
... any thoughts on how to move the work forward?

… meetings, PRs, …?

JR: think a next step is the up/down issue

… would like to move to github

… re TE compat, might be helpful to have a f2f meeting for discussions

RB: agree; let's get PE spec on Github and start with the easy issues

AoB

AB: let's take the technical discussion to the mail lsit

<scribe> ACTION: barstow move the PE spec to Github (with Jacob and Rick) [recorded in http://www.w3.org/2015/04/07-pointerevents-minutes.html#action01]

<trackbot> Created ACTION-149 - Move the pe spec to github (with jacob and rick) [on Arthur Barstow - due 2015-04-14].

AB: meeting adjourned

<patrick_h_lauke> thank you, will review transcript as jack gave me zero chance to actually listen ;)

Summary of Action Items

[NEW] ACTION: barstow move the PE spec to Github (with Jacob and Rick) [recorded in http://www.w3.org/2015/04/07-pointerevents-minutes.html#action01]
 
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.140 (CVS log)
$Date: 2015/04/07 16:46:31 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.140  of Date: 2014-11-06 18:16:30  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/Tim_Dinlocker/Tim_Dresser/
FAILED: s/... 2/RB: 2/
Succeeded: s/need balance/need to balance/
Succeeded: s/OP: think PE will help with some scrolling issues/Scrolling implementation consolidation will help to support pointer events on all the platforms/
Found ScribeNick: ArtB
Found Scribe: ArtB
Inferring ScribeNick: ArtB
Default Present: Matt_Brubeck, [Microsoft], Scott_Gonzalez, Olli_Pettay, Art_Barstow, Chrome_Team, +44.797.663.aaaa, patrick_h_lauke, Doug_Schepers, +1.770.402.aabb, jrossi2
Present: Art_Barstow Matt_Brubeck Asir_Vedamuthu Scott_Gonzalex Olli_Pettay Rick_Byers Tim_Dresser Mustaq_Ahmed Patrick_Lauke Jacob_Rossi Doug_Schepers
Regrets: Sangwhan_Moon
Agenda: https://lists.w3.org/Archives/Public/public-pointer-events/2015AprJun/0003.html
Got date from IRC log name: 07 Apr 2015
Guessing minutes URL: http://www.w3.org/2015/04/07-pointerevents-minutes.html
People with action items: barstow

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


[End of scribe.perl diagnostic output]