search

w3c / deviceorientation

W3C Device Orientation spec
https://www.w3.org/TR/orientation-event/
Other
49 stars 32 forks source link

Remove the oncompassneedscalibration event #107

Closed reillyeon closed 1 year ago

reillyeon commented 1 year ago

There are no implementations of this event.

Fixed #38.

Preview | Diff

anssiko commented 1 year ago

@diekus per https://caniuse.com/deviceorientation "compassneedscalibration event is supported in IE11 only for compatible devices with Windows 8.1+"

Can you or someone else from Microsoft confirm whether this caniuse.com information is accurate? Are you fine with the proposed removal?

marcoscaceres commented 1 year ago

Given that IE is discontinued, and Windows 8.1 support ended on January 10, 2023, it’s probably pretty safe to merge this.

diekus commented 1 year ago

Hola @anssiko , out of curiosity, what is the concern here with the event? There doesn't seem to be any implementation in Chromium nor other vendors.

I can confirm Chromium Edge does not implement this event.

marcoscaceres commented 1 year ago

The concern is basically the one you stated: no one implements it and so it could be removed from the spec. Otherwise, developers may try to rely on it etc.

Removing it also simplifies the spec quite a bit.

anssiko commented 1 year ago

Thanks @diekus! You were the only one implementing this feature so wanted to check with you.

@marcoscaceres I infer you're supportive of this removal from WebKit's point of view, correct?

marcoscaceres commented 1 year ago

That is correct. It would be great to remove it 🙏

anssiko commented 1 year ago

As a further evidence, we installed https://github.com/w3c/deviceorientation/pull/71 ~4 years ago to give a heads-up of a possible removal. I personally consider that an adequate time.

We are in fact in a similar situation with ondeviceorientationabsolute.

marcoscaceres commented 1 year ago

Cool. Can we merge this?

reillyeon commented 1 year ago

ondeviceorientationabsolute is at least supported by Chromium-based browsers. According to our metrics usage is significantly lower than ondeviceorientation however as long as these APIs are used passively (without the site requesting permission) I don't trust these metrics to reflect how critical the API is to site functionality. Once we complete work on Chromium issue 947112 and have the ability to measure how often this feature is used with explicit user consent we will be able to make a more informed decision about whether the "absolute" version of these sensor events is important to developers.