Ask HN: What makes for a great VR OS?

I was reading Ars' preview of Google I/O (http://arstechnica.com/gadgets/2015/05/google-tracker-io-2015-edition-android-m-chromecast-2-and-lots-more/) which mentioned that they are preparing a version of Android suitable for VR. While I am experienced in combined hardware/software platforms and consumer electronics, my knowledge of operating systems and how to evaluate them is pretty weak. It's an area that's totally blind to me, and I couldn't help but think this is Google's equivalent of Microsoft's initial versions of Windows for phones, which were ill-suited to the form factor and usage model.

Beyond just predictions for whether Android or Oculus or whomever wins this space, I was curious what criteria you would use to evaluate the suitability of an OS platform, specifically for a VR application. What technical underpinnings are necessary? What features do you think will really differentiate a platform? If there's anyone writing about this in an approachable way, I'd be interested. I know Carmack speaks about it a great deal, and I've read a lot of Abrash's thoughts. They all seem to range from reasonable to brilliant, but I don't have enough background to really offer a thorough criticism.