👀 Canvas place element is very interesting for potentially DOM content in plane layers.

[AdaRoseCannon]

https://github.com/WICG/canvas-place-element

It's super early days but could be a really interesting solution for WebXR.

[cabanier]

Is there an explainer that talks how they solved the privacy issues?

[AdaRoseCannon]

Not that I have seen, but I haven't had a good look through their breakout session minutes: https://docs.google.com/document/d/1cZfeeJ3wnPxikYJiW4P637hO6SNOeV5sQy2kEm5w0J8/edit#heading=h.xsf51u7rrtz

It would be good to do a shared meeting with them to let them know the privacy issues that we have encountered with similar efforts.

It does seem though that they are looking to solve the privacy issues on a case-by-case basis, which would've been too challenging to do within the Immersive Web.

[cabanier]

yes, we should have a joint meeting.
I agree that this would be a gamechanger for WebGL/WebGPU based experiences such as WebXR

[AdaRoseCannon]

/agenda Arrange a joint meeting with these participants and Immersive Web.

[m-blix]

Are there WebXR-specific privacy considerations?

How would WebXR input work with this? This seems like one challenging part to figure out. Potentially it could work in an ~automatic way from a user perspective if use as a layer as we have the layer position, parameters, etc along with the user input parameters.

It would be interesting if could figure out input on a non-layer as well.

[AdaRoseCannon]

I think it would have to work with the input's targetRaySpace and where they intersect the planes and select events so it's largely opaque to the DOM that it is being used in WebXR but also we don't want the Website generating automated clicks on ads and such.