Resolution upgrades are sure to come in the form of LCOS micro displays. Our current viewing system should work with 4K displays now as they are Flash based. I guess we would be grossly oversampling for OculusRift if we did not have zoom. I can understand why you need some sort of damping like stabilization at higher zoom. That is like my cameras. I could see where a click on a subject could function as a "target designator" for image stabilization. Your "virtual tripod is a good anology and I guess that I would add that it could also perform in ways that are better than a real "fluid head".
Relative motion of an image that is different than one's real motion can be disorienting if it does not have consistant game physics. I bet that lag would also be a consideration along with overshoot. I suspect that it might be possible to use rate of change to predict anticipated positions and display those in time for the head to arrive at the correct location.
If I were creating images only for the Oculus I would use a simpler camera system and create more panoramas by moving from location to location. Alternately, I can use stereo videos to connect locations or even make panoramic stereo movies. Another method that I currently use is to integrate my panoramas with multiple walk around viewpoints to create object models.
Examples of the photo generated models are posted here;
http://www.holographics3d.com/hg/galleries/main/models/3/turqgalsmall.pdfhttp://www.holographics3d.com/calffalls.pdfClick and drag the object to rotate. Hold control with click and drag to move sideways. Use the mouse wheel to zoom.
I am seeking a viewing system that could allow us to view 3D models in stereoscopic 3D that works with all platforms, especially 3D TVs. We learned years ago that posting in a format that converts to the desired viewing format on the fly is necessary. Otherwise, one stops taking photos and spends all their time converting, posting, and linking.
John Toeppen