Re: [pointerevents] Coordinates of a `pointercancel` event (#463)

Thanks @patrickhlauke for creating the PR reflecting our last meeting's decision, this highlights a questionable difference between our decision vs what we do for a touch `pointerup`.  Let me elaborate.

A touch `pointerup` is very similar to a touch `pointercancel` because each of them represents a moment the touch-point has become effectively non-existent for the page.  From this perspective, if the `pointercancel` event carries all attributes of the last seen `pointermove`, the same should hold for a touch `pointerup` too because an up means the touch-contact is gone. 
 This would be in conflict with both:
- the current spec (which explicitly mentions `pointerup` pressure to be zero, and we [decided similarly](https://github.com/w3c/pointerevents/issues/225) for few other attributes of `pointerup`), and
- current implementations (a very fast swipe on Firefox Android logically shows an updated coordinates in `pointerup` vs the last `pointermove`).

Now an additional point, which is perhaps more important here: looking back at the main motivation behind our decision, I think our comparison with `touchcancel` coordinates was not precise because that spec is silent about the coordinates of the `TouchEvent` itself!  The [`changedTouches`](https://w3c.github.io/touch-events/#dom-touchevent-changedtouches) specifies the last known location of individual `TouchPoint` in the "group", and says nothing about the location of the "group" (i.e. the group "called" `ToucheEvent`).  In particular, whenever there are two `changedTouches`, the coordinates of the `TouchEvent` remains logically inconclusive!

---

Because of both of these, it now seems to me that a `pointercancel` should not be required to "copy" last event's data.  I am now in fact biased to go as far as suggesting the simplest requirement that "this event carries no data other than `pointerId`", for sake of being more readable and concise for the developers who won't lose anything with such a change.

Thoughts?

-- 
GitHub Notification of comment by mustaqahmed
Please view or discuss this issue at https://github.com/w3c/pointerevents/issues/463#issuecomment-1431792215 using your GitHub account


-- 
Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config

Received on Wednesday, 15 February 2023 18:05:35 UTC