Web Events WG Voice Conference -- 15 Oct 2013

<smaug> uh, I need a minute or two

<scribe> Scribe: Art

<scribe> ScribeNick: ArtB

<cjolif> +cjolif

Tweak agenda

AB: I submitted a draft agenda on October 11 http://lists.w3.org/Archives/Public/public-webevents/2013OctDec/0012.html. Since then, IBM notified us of some work they have done re PointerEvent and TouchEvent mapping so I propose we include that when we discuss mapping.
... any objections to that?

RB: no, this is great

AB: any other change requests?
... Christophe, please give a short intro

CJ: Sebastien is more knowledgable here but I can fill in for him since he can't make it today

… our context is Dojo

… need to map TE to PE

… to help standardize the way we deal with mouse, touch, pointerevnts

… PE isn't there today

… so need to make a mapping to complete our impl

… We noticed this group intends to work on this mapping

… so we hope our work can help this group

… The mapping is in a spreadsheet

… We think our work will be helpful

Publishing Working Group Note of TEv2 spec

AB: We previously agreed to publish Touch Events v2 spec as a "WG Note" which signifies work on that spec has stopped. Unless I hear otherwise, I'll assume that is still our `Plan of Record`.
... is that still the case?

RB: fine with me

… think we should publish something

DS: I think a Note is reasonable

AB: personally, I don't recall the compelling reasons to publish a v2 Note, that is, I'm not convinced there is a `problem` that is solved by publishing this Note. On the other hand, I don't see anything about it that is `harmful`.

RB: so, what if we don't publish anything?

… is what we have adequate?

… we don't want it to get implemented

AB: we never even published a FPWD of v2

… Note is clear signal the work has stopped

DS: if we think v2 reflects some impl, then I think a Note is probably better

<sangwhan> Do we want to have certain implementations that portions of v2 as is? Shouldn't the v2 specifics be reverted if the work is officially stopped and the spec is considered scrapped?

… it gives us a chance to add some context and rationale

… publishing a Note is light weight

… so I feel more comfortable with publishing a Note

RB: Olli, does v2 match FF impl?

OP: not sure; we need to talk to Matt

<sangwhan> For the sake of interop I'm not sure having certain implementations provide subsets of features of v2 is a good idea

… I don't care if Note or ED

… Both say "don't implement it"

RB: the only part of v2 we implement is the radius stuff

… having those mentioned somewhere is important

… but if FF enters touch{enter,leave} that would be important

OP: but we don't

RB: think it is in the MDN docs

AB: without Matt, not comfortable reversing our PoR

… so that is what we will do

AB: so what changes are High Prio?

RB: remove section 5.7 and 5.8

… assuming no one has implemented those

AB: Sangwhan?

SM: no, we don't implement those

RB: some additional props on the Touch object

… want to leave the radius props

AB: we also need to update some of the Web IDL that was buggy before Boris noted those errors
... any other changes?

RB: instead of publishing everything as is, we could just publish those additional pieces

… i.e. 3.1

… would be a lot less work

DS: seems reasonable to me

OP: no one will look at the Note so I don't think it matters much

… I recommend minimum amount of work

SM: do we have some boilerplate for the top?

DS: there is some precedence e.g. Web Storage

AB: good Q; we need a propoasl for the obsolecence text

RB: think we should listen to Matt

AB: I think that's a good idea

<smaug> http://www.w3.org/TR/webdatabase/ is pretty clear about its status

<scribe> ACTION: barstow followup with Matt re if TEv2 Note should be `full spec` or just the `extra pieces` that have been implemented [recorded in http://www.w3.org/2013/10/15-webevents-minutes.html#action01]

<trackbot> Created ACTION-105 - Followup with matt re if tev2 note should be `full spec` or just the `extra pieces` that have been implemented [on Arthur Barstow - due 2013-10-22].

Touch Events / Pointer Events mapping

AB: we agreed a while ago to create a TE and PE mapping document. The following e-mail I sent last week summarizes this topic http://lists.w3.org/Archives/Public/public-webevents/2013OctDec/0011.html.
... we also now have a related input from IBM https://docs.google.com/spreadsheet/ccc?key=0AvdBn9Kvx22qdGRnRXNPb0ZBTUl3SEkwdUdtaW9pWWc&usp=sharing#gid=2
... before digging into any details, perhaps it would be helpful if we stepped back a bit to talk about the `Problem Statement` here, f.ex. what is the problem we want to address, who is the audience, what is the scope, etc.
... Rick, your thoughts?

RB: we know there is some tricky interaction here

… want to reduce pain for the Web developers

… want browser authors to have some info about how to handle these interaction

… how does css touch action property work

… how is default handled

… there are some other diffs between PE and TE that need to be documented

… f.ex. with mouse interaction

… it's unclear how browsers should handle the various cases

… when the browser supports both TE and PE

… The various polyfills need answers to these Qs

… (Dojo, Polymer, etc.)

… When sites bits of a page that support one or the other, things are complex

SM: also want to know about PE and caret browsing

… there are some places where caret browsing doesn't work

RB: it's going to be even worse for browser that support both of these events

AB: Sebastien, please give us a summary of your work

SP: we started with some common user tasks

… first was click

… 2nd was swipe

… from each action, we observed the behaviour from various browsers

… some browsers support only TE, some only PE

… We noted the sequence of Touch events and Pointer events

… need to map the sequence of TEs to PEs

… The second step was to identify patterns of TEs to PEs

… Every browser has a different sequence of events

… but we found a way to generate consistent Pointer events

… Not sure if this is what you were looking for

AB: what do people think?

RB: I just saw this yesterday

… this is great data of existing behavior

… I've been working on a doc re behavior and noting differences

… IBM's doc is more thorough

SP: want to get consistent behavior across browsers

… wondering if this could be standardized

… to get consistent behavior

RB: re Col F, do you think that is consistent with the PE spec?

SP: I think what we have is consistent with the PE spec

… we are also implementing some stuff that is not specified

… f.ex. how click will behave

… preventdefault is implemented differently depending on the gesture

… still is ongoing work

… but I think it is close to the PE spec

RB: when you find things that disagree with PE spec, please let us know

… agree no info re single and double click for PE is not documented

… there are some IP reasons why behavior like that is not specified in W3C

OP: need to define order of touch and pointer events

RB: yes, and how one cancels the other

OP: what about FF on mobile?

SP: we didn't include it yet but it could be added

RB: one conclusion, there is a lot of detail that needs to be documented

… for browsers and polyfills

… must get consistency

… I think this doc is helpful

… but this is just one piece

… Must also define the interleaving of PEs and TEs

SP: yes, the goal is for the dev to only have to care about Pointer Events

RB: so, interleaving isn't an issue in that case

SP: correct

… although there will of course be cases like that

… but that isn't a primary task for us

RB: would like you to help with our mapping doc

SP: yes, we can do that

RB: please feel free to reach out to me re Chrome's behavior

<rbyers_> Here's my doc on touch event details across browsers:

<rbyers_> https://docs.google.com/a/google.com/document/d/12k_LL_Ot9GjF8zGWP9eI_3IMbSizD72susba0frg44Y/edit

https://docs.google.com/document/d/12-HPlSIF7-ISY8TQHtuQ3IqDi-isZVI0Yzv5zwl90VU/edit

AB: ooops; sorry

<rbyers_> Sorry, this is it: https://docs.google.com/a/chromium.org/document/d/12k_LL_Ot9GjF8zGWP9eI_3IMbSizD72susba0frg44Y/edit

SP: we will check this

RB: so, how do we move forward

AB: I'd like to get this work done by mid November

RB: I think a lot of the details that are important can't be documented until we get more experience

… we could always shorten our scope

… don't think we'll understand interleave and other subtel issues for a while

… need to get PE out and deployed

AB: if we used something like a wiki for the mapping, we could close the group and still let the mapping doc evolve

DS: I see some conflicting requirements here

… I think we need to keep moving toward closing this group

… I don't see IBM's doc something that maps directly to a W3C spec

… If the doc is more prescriptive i.e. "we expect browsers to do X in scenarios Y/Z"

… We can also revisit defining the mapping in the PE Working Group but we must be very careful with that approach

DS: if it's an informative doc, we can use a wiki or WebPlatform.org

<sangwhan> What if it is both?

… but if it is more prescriptive for browsers, it should be Note

<sangwhan> (as in, reference for implementors but also helpful for web developers)

… I suspect we can get an extension

RB: I think we want a doc that is advice for browser implementers

DS: but that is not what IBM has

RB: yes, agree

SP: agree too; we can also do some more work

DS: do we think we can define the "desired behavior"?

… if so, we should be able to define that fairly quickly

RB: Column F is mostly prescriptive

… the larger problem is interleaving and touch-actions

… think we are going to need a new CSS property

… need to allow opt in

… not sure how to get that specified (e.g. which group)

DS: yes, CSS WG is one option

… although, other groups can do so if the work is done jointly

DS: perhaps we can revisit the mapping document decision within the PEWG

… if we need a specification, one possibility is DOM4

… now part of the HTMLWG

RB: the PEWG's charter doesn't explicitly state that interaction with Touch events is out of scope

<scribe> ACTION: doug ask PEWG to revisit the decision to document the PE and TE mapping [recorded in http://www.w3.org/2013/10/15-webevents-minutes.html#action02]

<trackbot> Created ACTION-106 - Ask pewg to revisit the decision to document the pe and te mapping [on Doug Schepers - due 2013-10-22].

<smaug> sangwhan: htmlwg perhaps?

<sangwhan> smaug: :D

DS: think it would be helpful if we had a very specific proposal for the PEWG

… f.ex. be able to say "this is the sequence of events …."

RB: still need more impl and deployment experience

OP: we need at least 2 impls

… that will take some time

AB: what about moving the mapping to a CG

RB: that's OK with me

DS: yeah, that's worth pursuing

… understand we can't do detailed work on the mapping without more impl experience

… perhaps the CG is the right approach

AB: let's work on a proposal for moving mapping to a CG or PEWG

DS: ok; not clear which one is better

<scribe> ACTION: barstow work with Doug on where to define the PE/TE mapping (CG, PEWG, etc.) [recorded in http://www.w3.org/2013/10/15-webevents-minutes.html#action03]

<trackbot> Created ACTION-107 - Work with doug on where to define the pe/te mapping (cg, pewg, etc.) [on Arthur Barstow - due 2013-10-22].

SM: webkit bug for PE hasn't been progressing

… stuck since the Blink fork

AB: thanks Sebastien and Christophe for joining our call

DS: yes, thanks very much

AoB

AB: anything else for today?
... meeting adjourned

- DRAFT -

Web Events WG Voice Conference

15 Oct 2013

Attendees

Contents

Tweak agenda

Publishing Working Group Note of TEv2 spec

Touch Events / Pointer Events mapping

AoB

Summary of Action Items

Scribe.perl diagnostic output