The expected behavior of the WebVR APIs when there is an "HMD" is clear, but there are a variety of other scenarios where the behavior should be clearly defined and implemented in Chrome.

Other scenarios include:
* Magic window - Moving a device, such as a phone, to explore an environment but without a headset and thus not using a binocular rendering.
  - It appears that such devices should expose a VRDisplay that reports canPresent=false.
  - This makes sense on a phone and tablet, but what about Chromebooks, laptops, and other devices that don't run a "mobile" OS?
* "Desktop" operating systems - Is there anything the WebVR APIs can or should provide the application? See the previous item.
* "VR controllers" without a headset

It may also be useful to specify how implementations should behave and applications should react if the user does not activate an HMD. For example, a user may not have their Daydream View or Cardboard readily available even though the browser might have reported it as an available display.

In addition, we should think about recommendations for authors that want to provide a rich VR experience across clients. (This may involve handling keyboard, mouse, touch, pointer events, etc.)

For more discussion on some of these issues, see the thread starting at https://mail.mozilla.org/pipermail/web-vr-discuss/2016-November/001929.html and in particular https://mail.mozilla.org/pipermail/web-vr-discuss/2016-November/001937.html.