Have the doc here: https://docs.google.com/document/d/1BgyOo1pxr6f8xHA8o7w9-o5vlCSuWtyJg9HPbEgffco/edit?usp=sharing, pls take a look at it if you are interested.

FYI. Since no more comments on design doc. I started working on it now.

There's a chrome extension that seems to meet your needs: https://chrome.google.com/webstore/detail/webvr-api-emulation/gbdnpaebafagioggnhkacnaaahpiefil

This might be an easier way to get your needs met (without adding a lot of complexity to the code base) - would this work for you?

Yes, this extension could work for testing. But I think this proposal could support the situation that display vr content in web page without HMDs(e.g. https://docs.google.com/presentation/d/1FYbOKq_CyUjoCztLPvzk49oRc9P6zsjGDgyo1aTrXHw/edit#slide=id.p)

Yes, we want VR content (e.g, WebVR) to be available to all clients regardless of whether the user has an HMD (or wants to put it on). This is tracked by  issue 672630 , and that depends on some WebVR spec work.

However, it wouldn't make sense to expose the VR browser to users not in an HMD. Yet it would be useful for Chromium developers working on the VR browser and WebVR (faking capabilities not exposed to users without HMDs). That is what this issue tracks.

Supporting either use case (and desktop HMDs) will require some refactoring, which could include significant changes to interfaces and other code. This is tracked, at least partially, in  issue 672654 . The new design will consider the first case and should consider the second. Given these upcoming changes, it probably doesn't make sense to start implementing anything quite yet.

Thx for your explain the difference between  issue 672630  and this issue. After reading, I still think that virtual device could be an option to solve both case with some flag control. Since there will come a new design soon(tracked in  issue 672654 ), such implementation which based on old architecture will make no sense.
If I misunderstand anything, pls correct me.
P.S. one more question, you said that  issue 672630  depends on some WebVR spec work,
could you offer links of such issues? thx.

There isn't yet a spec issue (the entire topic is difficult to describe), but we'll file one soon and update  issue 672630 .

Although we have a UI test app now, there is still work to be done to enable this in the browser and for general use, including WebVR.

Removing VR-Desktop, as this isn't desktop specific, and not required to ship desktop support.

Removing Internals>VR component and remapping to Internals>XR