– DRAFT –
PEWG

Rob: is there any chance to move this meeting? Vlad would love to join, but this time doesn't work

[group discusses possible times/dates]

Patrick: decided to move to -1 hour current time slot]

ACTION: change the calendar invite etc to have PEWG starting one hour earlier in the day

Recharter w3c/strategy#515

PLH: two comments we had still: a) difficulty differentiating between what's in and out of scope for gesture

PLH: struggling to find differentiation

Patrick: roughly, out of scope is how the user agent determines if something IS a gesture (look at the position, and delta time, and if below a threshold treat it as a swipe). what is IN scope is: once the black box inside the UA has determined something IS a swipe, how is this swipe etc gesture exposed to the author

PLH: second question is do we actually have a spec / incubation that shows exactly what we want to specify

mustaq: doubletap (?)

PLH: also, are we thinking of having this as a separate spec for gestures?

Patrick: I was envisaging it being part of the PE spec itself

<mustaq> Extension touch-action spec: https://compat.spec.whatwg.org/#touch-action

Rob: I would also say it's easier to keep it all under the one spec, as we'll also likely need to define the interplay between gestures and regular singular pointer events

mustaq: sorry, meant pinch-zoom, not double-tap

PLH: that's touch-action, not gestures

Rob: touch-action lets you PREVENT gestures...

<mustaq> https://w3c.github.io/pointerevents/#the-touch-action-css-property

Patrick: from my point of view, I was thinking something along the lines of https://hammerjs.github.io/

<smaug> https://searchfox.org/firefox-main/source/dom/webidl/SimpleGestureEvent.webidl#8

Patrick: gesture events from webkit implementation https://developer.mozilla.org/en-US/docs/Web/API/GestureEvent

<mustaq> https://caniuse.com/?search=pinch-zoom

Patrick: are we perhaps jumping ahead too far? should we incubate something small first, with an aim to implement that as part of Level 4

PLH: tentative deliverable - still in scope, can work on it, but we need to do incubation first

Olli: do we need to already have something written down, even just as a markdown file, to kick off the incubation?

PLH: at best can put as tentative deliverable, at worst we need to write a rough proposal to point to

PLH: tentative MAY fly , but would stand better chance if we had rough proposal

Patrick: if it's easier/better, we can probably work on a rough proposal

Patrick: something that includes rotation and scale as per what WebKit has implemented. then additionally swipe

PLH: what about pan and zoom

Rob: pan is same as swipe

Rob: zoom is scale

Patrick: pan and swipe may need differentiation, but we can discuss further

ACTION: Patrick to draft a rough initial outline of gesture spec idea as separate markdown file in PE repo

ACTION: PLH to make tweak to charter to try and disambiguate in/out of scope more precisely

Follow-up from UI Events meeting https://docs.google.com/document/d/1iRFgqtReyomoCwZFEdRUdoW1ManD5t656CVUGJm7Xos/

Patrick: just as reminder what the concrete actions are from yesterday's UI events meeting

Patrick: if i remember correctly, philippe you said you'll do a PR...

<smaug> zcorpan is fixing UIEvents build errors w3c/uievents#408

PLH: yes on my todo. already done triage/labelling issues, will transfer over

Patrick: if i recall correctly, our charter wording is already saying that we'll look at relationship between PE and mouse, pointer lock, etc. we're perhaps a bit cheeky that "yes, we'll define relationship with mouse events by basically stealing mouse event definition for ourselves", but think it's still ok

Olli: do we want to mention wheel events? they extend mouse events, but perhaps we do need to be explicit

Patrick: philippe want me to propose a tweak to charter to mention wheel?

PLH: either you do it, or i do it. whoever does it first...

PLH: looking at the charter again, where it mentions "clarifying the relationship with other interaction events..." ... do we maybe even want to rename Pointer Events to something more generic?

Patrick: Pointer UI Events ;)

PLH: i'll see if we can tweak the definition further, where it talks about "models"

WPTs

Patrick: we had an action "Rob to look at adding help/meta to WPTs to disambiguate level 3 and level 4 ones"

PLH: do we actually need it?

Patrick: i thought we wanted/needed it to disambiguate better which tests we need for 3 specifically

mustaq: developers probably just care about latest version...

PLH: if it was just to make W3C process work smoother ... just send me a one-time list of tests manually, and we can then forget it as soon as transition request has been approved. if it's not useful for you beyond that, let's not do the meta/help stuff

Olli: yeah when implementing i don't care about version...

PLH: let's not then, and if Rob or whoever can just provide a manual list of what's relevant for 3, then let's just go with that

PLH: if faster, provide a list of what's just tentative/for 4, so by exclusion we can pinpoint what's just for 3

ACTION: provide list for PLH to work out what's just Level 3 specific for transition request

<mustaq> https://wpt.fyi/results/pointerevents?label=experimental&label=master&aligned

Mustaq: there's some tentative and non-standard ones ...

PLH: i can produce a list, an implementation report, and see if i can find any gaps for the purpose of level 3 transition

PLH: is it ok if i add the draft implementation report to the level 3 branch

Patrick: yeah go fo it

PLH: does it publish the branch to github.io?

Patrick: no think only main gets pushed to github.io

Olli: those experimental/non-standard wpts ... think we should get rid of them ...

Olli: some things are from internet explorer times...

Olli: maybe rename them to something ... still nice to have around?

Mustaq: maybe splitting tests ... standard and non-standard

mustag: ... tentative ...

<mustaq> w3c/pointerevents#167

mustaq: does the above look reasonable?

Olli: somehow renaming, or separate directory...

Olli: does any other spec test this sort of non-standard stuff

Olli: "to be removed features"

Olli: couldn't find any webcompat issues in bugzilla relating to these. one i found was very old...

ACTION: mustaq to explore what to do with the non-standard tests

mustaq: meanwhile philippe ignore non-standard tests

PLH: ok!

<plh> https://www.w3.org/2021/04/wpt-fyi-snapshot.html?filter=pointerevents

PLH: found the tool that does snapshots. may need some tweaking

PLH: i'll look into it, some undefined ones

Olli: mustaq was the related target ... slot ... failing in chromium?

Olli: was it part of this or pointer events tests?

mustaq: yeah we need proper new test for that

Olli: ... after target remove from slot ... browsers aren't happy with that

<mustaq> https://wpt.fyi/results/pointerevents/pointerevent_after_target_removed_from_slot.html%3Fmouse?label=experimental&label=master&aligned

Olli: ... propagation part is defined in DOM spec. issues are not about PE itself

mustaq: can you file an issue and cc me for that?

PLH: once we update PE spec and UI events spec, we're going to have to move tests in WPT...but we can wait for 4

mustaq: do we have to move them, or can we just point to the mouse event spec folder?

PLH: once we get to implementation report for 4, we'll have a tricky time trying to define what tests PE 4 will cover

mustaq: worried about links for mouse/ui events tests breaking

PLH: something to consider ... which is easier for developers

Patrick: next meeting 14 Januar 2026. have a good break/holiday everybody