On point #3, this is exactly why I've been building Primrose (https://www.primrosevr.com). Primrose is meant to be a framework for web developers to be able to build applications that live inside of a virtual environment. Microsoft calls this "Holographic Applications" (I don't exactly like that term, but I suppose I will get over it). The point is to divorce the need to do 3D graphics and geometry from building the application, to have those as the baseline and provide on top of that a shared set of UI metaphors. There is a really easy analogy to draw with standard, 2D GUI systems: you shouldn't have to write an HTML rendering engine before making a Web application, you shouldn't have to work in a game framework to make VR-enabled applications.

Please check out Primrose. I know it has some rough edges here and there, but I've already used it to make some interesting things (a client of mine was recently featured on Bloomberg.com for http://rex.legend3d.com). I know people are wary about "single-contributor" projects, but I've already been building Primrose for 2 years now, it's not going anywhere, and I'm open to bringing collaborators on, just nobody has really stepped up (and I've been so focused on working on VR projects for clients that I've not really had the time to proselytize).

Somebody is going to bring up A-Frame: I think A-Frame is a really nice system, I just think its design goals don't really match what I think is important. A-Frame wants to be the entity-component system for WebVR. That's great. But I don't think that meshes well with "get web developers on board". I don't want developers to have to think about what sort of motion controller component to use in their system. I actually want the system to be more restricted, less open-ended than A-Frame. Also, Primrose came out long before A-Frame, so I'm still married to Primrose for as long as I can be productive in it.

There are some limitations where I haven't quite reached my goal of making VR accessible to web developers, but that is more an issue of limited number of man-months. You can build useful applications with Primrose today. But I have a very clear goal in mind and if it's something you agree with, I would appreciate the help.