Add a proposal for an "occlusion extents" function

[myaaaaaaaaa]

#2148 converted into a proposal as recommended by @TomOlson . Apologies in advance for any awkwardness.

Includes potential methods to emulate this function on older hardware.

See #2148 for some GIFs that may make this feature easier to visualize.

[CLAassistant]

All committers have signed the CLA.

[TomOlson]

@myaaaaaaaaa Thanks very much! The WG has been discussing next steps - to be honest, it is unusual for us to get a detailed proposal for a new hardware feature that looks like it might be a genuinely good idea, so we did not have an established process for handling it.

We have decided to treat this proposal the same way we treat internal proposals from working group members. That is, the PR will stay open and all members (and in this case, GitHub visitors as well) will be invited to discuss and investigate. This may result in discussion on this PR, the parent issue (#2148), or Khronos internal issues. If one or more GPU vendors decide to implement, they will propose an extension specification, and when that is accepted into the spec, this proposal document (or a later version of it) will be merged along with the spec language. You will of course get credit as original contributor on the extension, if you can give us a name to list you by.

I have to tell you that the time frame for this is long even if the idea works brilliantly. The lead time to prove that a new feature is useful, integrate it into the GPU architecture, and build and test the silicon implementation can easily be 4-5 years. So this is unlikely to land in hardware earlier than 2028. It is frustrating, but it is reality.

If by any chance you will be at SIGGRAPH next week, I hope you will come and meet us at the "Vulkan, Forging Ahead" BOF on Wednesday.