Meeting minutes
<ada> https://github.com/immersive-web/webxr/issues/674
Ada: WebXR #674: https://github.com/immersive-web/webxr/issues/674
webxr#674 Should requestReferenceSpace() wait until getViewerPose() won't be null?
Brandon: requestReferenceSpace not return until it is ready
… Maybe minor benefits or penalities for this behavior, especially if the method never returns
… Check with group to see if there are any additional or removed concerns
… <crickets>
Brandon: Discussion with Manish. Probably OK to wait on response resolution.
… It this takes more than a few seconds, then there may be trouble. Don't feel that it necessary to standardize
… across platforms.
Alex: Interesting places where the detailed interactions differ between platforms. Only time where this may cause
… problems is when it takes a while to establish local space. May be an issue in AR systems where it takes
… a while to construct the physical local space. Virtual local space should establish quickly.
… may cause apps to have extra code to deal with "dead time" while the local pose gets established
Manish: Return with null pose is useful as long as it is understood null == don't know as oppose to something
… else more sinister. Reasonable to expects apps to handle it.
Alex: agrees.
Manish: Feels wierd if BaseSpace can create a nulll pose after returning a real pose. But may be a spec language issue
Ada: Asks Brandon if sufficient info now?
Brandon: Most overtalked by dog, but appears to agree.
Ada: Next Item - https://github.com/immersive-web/webxr/issues/743
webxr#743 Allow sensitive UI to make use of the visible-blurred state
Manish: Trusted UI take over screen (everything else is hidden). No input to non-trusted UI ==> null poses for everything.
… Visible blurred has input hidden, but pose is not. No interaction with user.
<artem> +100500
<Manishearth> https://github.com/immersive-web/webxr/pull/1034
Manish: UA should use hidden or blurred depending on context. E.g., hidden for passwords
Brandon: The text provides an optional state that a UA may choose to use.
Diana: Derrives from experiments on trusted UI. Freeze scene in a user-obvious manner
Artem: Already implemented. Found its necessary to throttle to ~1/2 frame-rate. UX not god at other values
Alex: agrees. Hololens has implemented something similar to interact with wrist device
Brandon: <something>
Rik: Concerned - implementations have eye-trackers need to not have access to data while trusted UI is displayed
Brandon: No poses ==> no tracking.
… At Google, force page to over-render and only supply quantized input data
Diana: 1 uses cases: text input, button clicking. Text almost certainly trusted; button unknown
Ada: OK to continue in issue
Brandon: Would like to know if anyone is opposed to the concept.
webxr#980 Add XRInputSourceEvent::{screenX, screenY}
<srickets>
https://github.com/immersive-web/webxr/pull/980
Manish: Need to differentiate world vs. screen coordinates for rendering. Really applies to hand-held AR devices
Alex: Concern that the core has abstracted across all devices. To get screen-space UI, have been using DOM overlay
… When an item is rendered in XR space, it may be fine for AR, but can cause display problems for VR
… Splitting content to some that requires certain types of devices for rendering
Manish: Issue applys polyfill for DOM Overlay. It also allows other device+accessories to work. E.g., put game controller controls on display.
Piotr: Adds another (3rd) means to obtain this info. There is already game pad and DOM Overlay. Don't like that it
… is only valid for screen-space interfaces. This is a special case set of data and may need it's own handling
Brandon: Asks Piotr about getting the data through the game pad as (non-standard) Chrome?
Piotr: Not exactly. Standard or non-standard game pads do provide that info
Brandon: Piotr's method delivers a blob of data. This means is more web-standards like.
… Wiuthout this issue, there may be very poor individually developed solutions (screen-size transparent PNG to capture
… touch point).
Manish: Gamepad is not a standard way for exposing screen space coordinates. Would like something in spec...
… rather than individual one-off means that is different for each device
Piotr: Still concerned about adding data that is limited to certain display devices onto a an event that applies everywhere
Alex_Cooper: Agrees with Piotr.
Alex_Turner: Wondering what it means to add 2 parameters to the Core and not in a specific module?
Manish: OK to make changes to spec. Must make sense to be in Core. This one could belong in AR module; however,
… Layers may use this as well -- though Layers may implement this differently.
Ada: Is this a breaking change?
Brandon: This is strictly additive.
… Using dictionaries to determine where to put changes. Adding parameters to the event requires that the associated
… dictionary be updated. In this case, this dictionary is in Core; hence the update to Core.
AOB
Ada: Closes discussion in call on issue
… Additional topics?
<dom> YAY!
<artem> hooray!
Brandon: Notes that the Charter has been accepted for two more years.
<atsushi> https://www.w3.org/2020/05/immersive-Web-wg-charter.html
Brandon: Thanks the Chairs for continuing to push the Charter through the process to keep the group active.
Ada: Super close to PCR milestone. Take a look at the HELP issues. See agenda (https://github.com/immersive-web/administrivia/blob/master/meetings/wg/2020-05-12-Immersive_Web_Working_Group_Teleconference-agenda.md)
… for a listing of the items that can use some work.
Ada: Closes meeting