Pointer Events WG Voice Conference -- 09 Apr 2013

<scribe> ScribeNick: ArtB

<scribe> Scribe: Art

Getting started: tweak agenda; determine scribe

AB: I posted the draft agenda yesterday http://lists.w3.org/Archives/Public/public-pointer-events/2013AprJun/0021.html.
... One addition I propose is the "Constructor question and mouseEvents compat" thread started by Mozilla's Wesley Johnston http://lists.w3.org/Archives/Public/public-pointer-events/2013AprJun/0009.html since Matt indicated a spec change might be needed to address this comment.
... any objections to including that during the LC comment processing discussion?

[None]

AB: any other change requests?

<smaug> ArtB: sorry, I can't as I said here few minutes ago

<smaug> I'm in a train

<jrossi21> http://lists.w3.org/Archives/Public/public-pointer-events/2013AprJun/0012.html

DS: I think we should add Benoit's comment <http://lists.w3.org/Archives/Public/public-pointer-events/2013AprJun/0012.html>

AB: any objections to that?

[ None ]

Last Call comment processing

AB: the comment tracking document is https://dvcs.w3.org/hg/pointerevents/raw-file/tip/dc.html.
... Jacob, I think it would be useful, especially if/when there is a CR transition call, if each issue that resulted in a change includes a link to the related changeset(s). Can you please add that?

<scribe> ACTION: Jacob add links to changesets for LC comments that resulted in changes [recorded in http://www.w3.org/2013/04/09-pointerevents-minutes.html#action01]

<trackbot> Created ACTION-33 - Add links to changesets for LC comments that resulted in changes [on Jacob Rossi - due 2013-04-16].

AB: Jacob, you will also need to include changes since LC publication in the Revision History Appendix.

JR: ok, will do

Sergey's Comments

AB: Sergey's response to the group's resolution to Yanex's 5 comments is in http://lists.w3.org/Archives/Public/public-pointer-events/2013AprJun/0019.html. He replies to #1, #2, #3 and #5, presumably he is OK with our response to #4 (tiltX/tiltY).
... during the March 26 call we agreed to not make any changes although we did add #1 (#17 in LC tracking doc) and #2 (#16) to the v.next list http://www.w3.org/wiki/PointerEvents/UseCasesAndRequirements.
... in http://lists.w3.org/Archives/Public/public-pointer-events/2013AprJun/0022.html indicates he doesn't object to Sangwhan's reasoning re #2 so I think we are OK there.
... regarding the other comments, I'm not sure that much new information is being added such that people are convinced to change or move their position. This isn't an especially good situation but not uncommon either.
... I do note that Sergey has not formally Objected to the group's decision
... my gut feel here is that best way to proceed is to move to the implementation phase where we can then get feedback from not only implementers of the spec but developers too.
... let's break this up into two parts: 1) detailed comments regarding Sergey's comments; 2) more general comments on how we proceed.
... any detailed comments or thoughts on Sergey's comments?

RB: setPointerCapture gives us the ability to emulate touch event better

… so it can actually be essential with processing iframes

… which cannot be done with touch events

AB: does anyone think we need to change to our previously agreed resolutions on Yandex's comments?

JR: no, I don't think there is any new evidence to reopen our previous decisions

RB: I agree but want to add that pointerCapture is something that is worth debating

… I hope we can continue to encourage discussion and get feedback from devs

… especially wrt composition

JR: yes, I agree we need more feedback, especially from framework people

… and that's true for lots of other APIs in the platform e.g. Web Components

… It could be that Web Components helps with the capture issues

<smaug> hmm, does the spec say how pointer events work with (i)frames

RB: this is a bit diff for propagation
... stopImmediatePropagation

… we need to watch for composition scenarios that may be problematic

… e.g. embedded maps

… This is a subtle and important class of problems that we need to get right

JR: for single canvas element should be able to set capture to itself

… and then if lost can be determined

… think we have some tools

SG: nested widgets means inner widget needs to be able to say I'll handle this

… and now there is no way to do that

… i.e. say "I handled this event"

… with jQuery have to handle bubbling differences with special property

… but there are problems when adding new libs

… The innermost thing that wants to handle should handle it

DS: concerned our rationale is being discussed among ourselves

… and not directly to Sergey

… I think we should invite him to a telco

… It could help smooth out some differences

MB: we should also consider that most devs won't make the effort that Sergey did

… i.e. they won't read the spec and send comments

… they will get something that doesn't work and then just back to iOS

<smaug> (most devs use script libraries which probably get things right)

… so agree we need to do more outreach to devs

DS: yes, could put some doc in webplatform.org, MDN, etc.

… need some tutorial info

… think talking directly could help with the dialogue

DS: do people agree with a call

<mbrubeck> +1 on inviting Sergey to participate in the telcon

JR: he certainly represents a class of scenarios

… and we need to reach out to them

… not sure if a telco will help

… but think documenting/marketing these scenarios is justified

AB: I don't object to it, but I don't think we should necessarily block on such a call

DS: sure

<scribe> ACTION: shepazu invite Sergey to a call [recorded in http://www.w3.org/2013/04/09-pointerevents-minutes.html#action02]

<trackbot> Created ACTION-34 - Invite Sergey to a call [on Doug Schepers - due 2013-04-16].

AB: so it seems like we could record a resolution like: RESOLUTION: the group's previous decisions on Yandex's comments stands (see https://dvcs.w3.org/hg/pointerevents/raw-file/default/dc.html for details)
... any objections to such a resolution?

[ None ]

RESOLUTION: the group's previous decisions on Yandex's comments stands (see https://dvcs.w3.org/hg/pointerevents/raw-file/default/dc.html for details)

Mose Bottacini's comments

AB: Asir asked Mose to reply to his response by April 8. It appears we still have no response from him http://lists.w3.org/Archives/Public/public-pointer-events/2013JanMar/0171.html
... I believe re already recorded a Resolution that the group agrees with Asir's reply so I think the comment tracking should reflect that we contacted him, we got no response and that we won't block on this.
... does anyone object to not blocking on this comment?

[ None ]

RESOLUTION: the group's previous decision on Mose's comment stands (no change to the spec)

Constructor question and mouseEvents compat

AB: Wesley's comment is http://lists.w3.org/Archives/Public/public-pointer-events/2013AprJun/0009.html. There were replies by Jacob http://lists.w3.org/Archives/Public/public-pointer-events/2013AprJun/0017.html and Matt http://lists.w3.org/Archives/Public/public-pointer-events/2013AprJun/0034.html.
... Jacob already checked in a non-substantive fix for comment #1. What, if anything, do we want to do here re comment #2?

MB: issue #2 had 2 parts

… some confusing wording re primary pointers

… second issues is spec does not make it clear whether more than one simul pointer creates compat events

… I haven't tested this yet on IE

JR: the behavior here is that multiple pointers are fighting for the mouse cursor

… we think our behavior is compatible

… we looked at more arbitration scenarios

… they get pretty complex

… Some form factors may benefit from diff arbitration rules

MB: from the user perspective, then just use one mouse

… (if don't like jumping around)

… Perhaps the sec 8 algorithms can be then be left as is

… and arbitration is done at a separate level

… and we can just change the text to make it more clear

… re primary pointer

JR: yes, adding some clarity would be fine

MB: if you search for primary pointer there are a couple of other places that need clarification

JR: can you do that Matt?

MB: yes

RB: thanks Matt

<scribe> ACTION: Matt update the spec to reflect primary pointer clarification [recorded in http://www.w3.org/2013/04/09-pointerevents-minutes.html#action03]

<trackbot> Created ACTION-35 - Update the spec to reflect primary pointer clarification [on Matt Brubeck - due 2013-04-16].

Boris Zbarsky comments on April 5 .

AB: Boris submitted some comments on April 5 http://lists.w3.org/Archives/Public/public-pointer-events/2013AprJun/0016.html. For those that don't know Boris, he works for Mozilla.
... the good news is that he didn't identify any WebIDL issues. However, he made two other comments and no one has replied to them yet. I think we should treat his comments like LC comment.

JR: I didn't reply yet

<asir> do we need to record a resolution for the previous issue?

… but I did make a slight change re his first comment

<asir> okay - thank you

RESOLUTION: to address Wesley's comment, Matt will update the spec with some non-substantive changes

JR: the events are already marked as async in the table

… I need to update some text too

… I'll reply to that 1st comment

… Re the 2nd comment

… there is some performance implications with blocks

… want to be able to do the processing off thread

… inline elements are more challenging

… Based on the scenarios we looked at, most are block level elements

… for canvas and svg where you may have an interactive game, those are also block level

… In practice, we have not seen scenarios for inline elements

… I think if it was supported, the UX would be poor

… So mostly this way for performance reasons

DS: is the rationale documented?

JR: not in the spec

DS: think it would be useful

JR: I can propose something in my reply to Boris

DS: to overcome this perceived limitation, an author can make it block level with CSS, just in time

RB: there are some UCs where you wouldn't want to do that

… [ missed the scenario ]

… But I agree with Jacob - let's keep it simple for now

… and if that changes, we revisit

JR: at this point, I think it would go counter to our performance goals

RB: I don't think this is important enough to delay implementations

<scribe> ACTION: jacob reply to Boris [recorded in http://www.w3.org/2013/04/09-pointerevents-minutes.html#action04]

<trackbot> Created ACTION-36 - Reply to Boris [on Jacob Rossi - due 2013-04-16].

AB: is there consensus to address this by non-normative text?

DS: we need to include rationale

JR: yes, and that would be non-normative

DS: we need to add it

<rbyers> The scenario I described was an inline image inside of text (maybe even a button) which wants to have special touch handling. I said an inline image carousel, but more realistic is probably an inline button (eg. footnote) which wants to have some special touch behavior (eg. swipe to open) without changing scrolling behavior on the text...

AB: proposed RESOLUTION: re Boris' comment, we will add non-normative rationale to the spec

RESOLUTION: re Boris' comment, we will add non-normative rationale to the spec

Benoit's comment

<shepazu> http://lists.w3.org/Archives/Public/public-pointer-events/2013AprJun/0012.html

DS: Jacob replied

… I would characterize them as v2

… is that fair?

JR: yes, we have previously discussed these and agreed v2

DS: have he received a reply yet?

JR: no

DS: he is writing a framework

… this is good input

… we should have a roadmap to deal with things like pointer lists

JR: pointer lists is on the v2 list
... once we start to spec this, there could be broader device query API needed

DS: yeah, we had related comments with D3E

… OK, I'm satisfied

AB: is there consensus to not change the spec and these features are for v2?

JR: yes

<rbyers> sounds right to me

… sustaining previous resolutions

RESOUTION: re Benoit's comments, we consider those feature requests as part of v2

<scribe> ACTION: jacob reply to Benoit re the group's 9-Apr-2013 decision [recorded in http://www.w3.org/2013/04/09-pointerevents-minutes.html#action05]

<trackbot> Created ACTION-37 - Reply to Benoit re the group's 9-Apr-2013 decision [on Jacob Rossi - due 2013-04-16].

RB: I think there is also a bit of confusion re "same time" for events

… it could be worthwhile to say we think there is no real problem for this but we can look into it for v2

… but it would be good to make sure his concerns are clear

<scribe> ACTION: rick reply to Benoit re clarity for his concerns [recorded in http://www.w3.org/2013/04/09-pointerevents-minutes.html#action06]

<trackbot> Created ACTION-38 - Reply to Benoit re clarity for his concerns [on Rick Byers - due 2013-04-16].

[ I missed your comment Scott - sorry about that! Please add them here ]

JR: there are devices that give effectively parallel input

… moves can happen within 1/100 of a second

… and they can be "simultaneous" moves

RB: and pinch may not necessarily be two touches at the same time

JR: if Rick is going to reply to Benoit, then I don't have to

RB: ok, that's fiine

Moving Pointer Events spec to Candidate Recommendation

AB: we still have a few things to do before we can determine if we have consensus to publish a CR.
... the main open actions followups
... if we can get closure on comments real soon, we should be a position to have a CfC during our next call on April 16.

<scott_gonzalez> For tracking "framed" touches, you can use requestAnimationFrame() or similar and manually track the events that occur between frames.

AB: does that sound about right?

JR: yes

AB: I think it is time to agree to something like "We're done with LC comments. All new comments will be considered during CR."

JR: yes, I agree

RB: yes

AB: proposed RESOLUTION: the group agreed that any new comments for Pointer Events v1 will be consider CR comments
... any objections?

<asir> Closing Last Call issues list makes sense

<jrossi21> ACTION: jrossi to update comment doc to reflect these last comments [recorded in http://www.w3.org/2013/04/09-pointerevents-minutes.html#action07]

<trackbot> Created ACTION-39 - Update comment doc to reflect these last comments [on Jacob Rossi - due 2013-04-16].

[ None ]

RESOLUTION: the group agreed that any new comments for Pointer Events v1 will be consider CR comments

Testing: status, need volunteers for test submissions;

AB: Cathy did a lot of work on the Test Assertions since our last meeting, so thanks Cathy for that! http://www.w3.org/wiki/PointerEvents/TestAssertions.
... Scott checked in some tests yesterday, so thanks Scott! https://dvcs.w3.org/hg/pointerevents/file/default/tests/pointerevents-v1/submissions/jQuery.
... there are still lots of holes so please review the TA tables and pick some.

JR: there is the TTF event by Microsoft and Adobe on Friday and Saturday

… please come if you can

… Matt and I will be there

<mbrubeck> Not sure if I'm coming yet, but will try

SG: someone from jQ will be there

AB: that's excellent!

JR: some groups have a separate list for tests?

AB: I don't think that's necessary

<mbrubeck> I agree it's not needed for this group.

… and if we have probs, we can revisit

Any other Business: implementation status

AB: any new implementation status to share?

RB: re the WebKit/Blink fork

<mbrubeck> Wes Johnston at Mozilla is in early stages of implementation work.

… If anyone has Qs or concerns, I'm happy to talk about that

MB: on Gecko side, Wes Johnston has started some implementation work

… it's a bit of a side project but hopefully more people will get involved

DS: are you aware of any PE issues wrt Blink?

RB: at Google I/O I am giving a talk and hope to report some progress

<asir> I think you cannot hear me for some reason

AB: next call is April 16.
... meeting adjourned

- DRAFT -

Pointer Events WG Voice Conference

09 Apr 2013

Attendees

Contents

Getting started: tweak agenda; determine scribe

Last Call comment processing

Sergey's Comments

Mose Bottacini's comments

Constructor question and mouseEvents compat

Boris Zbarsky comments on April 5 .

Benoit's comment

Moving Pointer Events spec to Candidate Recommendation

Testing: status, need volunteers for test submissions;

Any other Business: implementation status

Summary of Action Items

Scribe.perl diagnostic output