Minutes from Pointer Events WF F2F at TPAC 25 Oct 2018

Dear all,

minutes from today's meeting are here 
https://www.w3.org/2018/10/25-pointerevents-minutes.html and pasted below.

Please send any amendments/etc by next week Friday.

W3C
- DRAFT -
Pointer Events WG F2F TPAC 2018
25 Oct 2018
Agenda

Attendees
Present
patrickhlauke, smaug, NavidZ_, jdai, TravisLeithead
Regrets
Chair
patrickhlauke
Scribe
patrickhlauke
Contents
Topics
stale branches
implementation status
PEP
Summary of Action Items
Summary of Resolutions
<scribe> Scribe: patrickhlauke

initial hellos, apologies for being a bit slack over the last few 
months, but we've pulled together to push towards the final stage

PE2 is currently in PR. feedback is already coming in from ACs. 2 
positives, no objections/substantive requests for change so far. all 
things going well, we'll be able to publish by end of calendar year. our 
group in its current form is chartered until end of December 2018

when PLH will join us, we'll discuss further steps for the working 
group. rechartering to WG, with possibility of putting the WG dormant 
until we're confident that we can move forward with substantive work on PE3

stale branches
NavidZ_ looked at those branches, and they're indeed stale. we can remove

implementation status
Patrick: wanted to get a feel at high level what the implementation 
status is, as for the last tests there were still a few failing 
outstanding items, but they are being worked on

smaug: we still have some issues with pen/touch

other issue is click event handling, which followed the (old) blink 
behavior which is against the spec as well

NavidZ_ gives example of a page which wasn't using pointer capture, but 
this was not yet implemented fully so the page relied on the other 
(non-standard) behavior. we're hoping to implement the behavior from the 
spec

<NavidZ_> Here is the bug for Chromium regarding targeting of click when 
pointer capture is active

<NavidZ_> https://bugs.chromium.org/p/chromium/issues/detail?id=689158

Patrick: mentioning recent work on Bootstrap for carousel swipes, and 
noticing that Firefox was very fickle when setting touch-ACTION: pan-y, 
but user moves by a few pixels up/down and FX immediately firing 
pointercancel. do we want/need to say anything about thresholds in the 
spec? fundamentally it's a user agent issue, which we've shied away 
from. maybe a non-normative note?

NavidZ_ gives similar example of threshold for browsers starting to 
handle scroll

Smaug: the situation is also similar to drag'n'drop

<smaug> https://bugzilla.mozilla.org/show_bug.cgi?id=1502010

Patrick: we could avoid having problems with our spec mandating UA 
behavior (which we've tried to avoid, see also staying away from 
defining certain gestures etc) by having a "negative" note along lines 
of "this spec does NOT define how UAs should do this...but here's the 
issue". possibly an informative note in 
https://w3c.github.io/pointerevents/#details-of-touch-action-values 
which already has similar notes. I will file a PR for the group to

discuss. need to also find out from PLH if we're still allowed to make 
non-normative tweaks of this nature? would assume yes, as following the 
AC review there may still be some things that need tweaking?

Patrick: also, there's some ReSpec errors...

NavidZ_ looked at the errors, and it brings up question why the Pointer 
Capture section is marked as non-normative. We'd need to explore this 
(though may be too late to make normative, in light of publication/above)

NavidZ: status in Chrome implementation

problem with making sense of touch-action on rotated elements, due to 
compositor

working on Firefox and Edge, but not Chrome. but also, no bugs have been 
filed against Chrome for real-world sites failing

soon we're going to automate manual tests

Patrick: we don't have anybody here from Microsoft, but from my 
understanding/conversations, Edge status is essentially stuck at PE 
Level 1. hopefully when/if we recharter, and get an MS rep, it will be 
interesting to see if there's desire in Edge to move to parity with 
Firefox/Chrome

NavidZ_ one of the things that would encourage adoption/work in MS etc 
would be showing developer desire to use PE

there are some cases like painting apps that can be done (better) with 
PE, but maybe for next Google.io we could have some more demos of cool 
things devs can do

Patrick: i think part of the problem is the PE is a cleaner/more 
sensible way of handling multiple pointer types (not registering mouse, 
AND touch, AND any proprietary extra stuff - but just a single set of 
PE, which can then still inside the function differentiate based on 
pointerType). but beyond that, it doesn't have a major WOW feature

NavidZ_ pointer capture - touch events had implicit capture that 
couldn't change, but PE lets you explicitly set and unset capture

Smaug: Firefox had a mouse pointer capture API (?) in the past already, 
which is why we managed to implement it for PE easily

NavidZ_ some nicer/more explicit way of handling RAF, canvas painting, 
workers, and raw coordinates from pointers etc

wonder how gaming might be affected (positively) by having more low 
latency / controlled

faster raw events

Patrick: difficult as the spec isn't doing anything "sexy". it's 
useful/makes sense. e.g. less code forking for different inputs.

big stumbling block has historically been getting developers to use 
while also telling them to polyfill for iOS, since Apple's silence on 
this spec meant that potentially devs would have to polyfill 
indefinitely. made it less of a sell compared to old tried and tested 
"mouse events plus touch events" (particularly after touch events were 
extended to allow for Pencil support/distinction).

PEP
Patrick: broad overview of what's happening with PEP appearing 
abandoned. asking what we should do...should/can Google or Mozilla take 
it over? Should I reach out to jQuery Foundation to see what status is? 
it's a good reference polyfill, would be shame to see it fall by 
wayside. some more complex PE Level 2 stuff may not necessarily be 
polyfillable (e.g. some of the more complex touch-action stuff), but 
would be good to have at least as a

baseline

NavidZ_: going to check with the Google Maps team (as they used polymer 
PE which was the original codebase before being shifted out to PEP and 
taken over by jQuery Foundation)

Patrick: I will reach out jQuery Foundation (Dave Methin (?), Scott 
Gonzales) to see what their plans are (if any)

(quick break while we wait for PLH who wanted to join us)

PLH: so question is how do we make PE2 part of history? maintenance of 
PE2, PE3? not in a hurry to push for PE3

https://raw.githack.com/w3c/pointerevents/pe3-charter/charter/pointerevents-charter.html

(PLH gave us a good overview of our options at this point. extending 
current charter, rechartering, etc)

discussion about obsolete vs superceded

PLH suggests we should do a CFC to mailing list about asking if the 
group is ok for PE2 to supercede PE1

in future, specify in the spec intro that "this specification 
supercededs..."

(Travis chats with us about MS commitment/desire to move to adopting 
more PE2 stuff in Edge. NavidZ_ mentions the PE Extension work 
https://w3c.github.io/pointerevents/extension.html and Travis suggests 
this could be submitted to TAG for initial review 
https://github.com/w3ctag/design-reviews/issues/new)

Summary of Action Items
Summary of Resolutions
[End of minutes]
-- 
Patrick H. Lauke

www.splintered.co.uk | https://github.com/patrickhlauke
http://flickr.com/photos/redux/ | http://redux.deviantart.com
twitter: @patrick_h_lauke | skype: patrick_h_lauke

Received on Thursday, 25 October 2018 16:02:28 UTC