5 ms·
When I was at Google working on VR there, I was always jealous of Oculus's vertical integration that let them move so much faster than Google's VR team. Working
by sxp 2y ago
When I was at Google working on VR there, I was always jealous of Oculus's vertical integration that let them move so much faster than Google's VR team. Working with partner hardware was the biggest time sink in terms of utility gained per hour spent for Google's VR plans and it's going to be a giant drag for Meta. Because of business contracts, there were firewalls in the source code which meant fixing bugs that spanned the stack took forever and there was a bunch of back-and-forth between companies about whose responsibility it is. This is very different from the standard culture of Google & Meta where a dev can (relatively) easily access the source code for any part of the stack and fix the bug without going through tons of red tape.
On top of this, hardware partners have very different goals from the core OS developer (i.e, Google or Meta) who want to grow the ecosystem. So the hardware vendors add various features to their hardware and the OS devs need to add hacks to work around it. In theory, the OS developer writes a "compatibility" doc and requires hardware makers to follow it. In practice, the hardware maker fails to do so and the OS maker has to put in software hacks to work around hardware bugs. And then engineers on the OS team have to waste time chasing these bugs which affect a tiny portion of users but are high priority due to business contracts.
Some of those business contracts made no sense from a VR ecosystem perspective, but Google went through with them for other reasons such as preventing a key phone maker from jumping ship to another OS or company. That's probably what's happening here. Meta doesn't want third-party vendors to build VR solutions on top of AVP or Google's XR OS, so they're offering crumbs to distract other companies.
- supernovae 2y agoI always felt Google did what Google does best by shooting off its own foot by controlling who has access to the google store with an iron fist. Since Meta can't win over google, i guess the next best thing to do is make sure meta store can be on other hardware since G isn't allowing that.
- loulouxiv 2y agoSince you say that you were working on VR at Google, did you work on Cardboard Camera or know someone that did ? Do you think there is any chance that the app woukd be open sourced some day ?
- sxp 2y agoI did work on Cardboard. The SDK was open sourced to some extent at https://github.com/googlevr/cardboard https://github.com/googlevr/cardboard . But I doubt Cardboard Camera will be open sourced since everyone has moved on. I was disappointed when I could no longer run the app since it's 32-bit only.
- loulouxiv 2y agoIt is working flawlessly on my current phone despite it having a aarch64 CPU, I guess that my OS must have some kind of multilib setup to keep compatibility with 32 bits apps. I am really anxious of not being able to continue capturing stereopanoramas whenever I will have to replace my trusty Huawei P30... Is there any way of knowing in advance if a given phone would be able to run Cardboard Camera ? I have built a collection of 200+ panoramas and really enjoy capturing some more when I am visiting new places. I got even more into this since I hacked myself some workflow to convert them and finally viewed them in all their glory in my Quest 2. The stereoscopic effect makes them really immersive ! I want to make the collection public when I will have taken the time to make a more practical/pretty listing page.
- raxxorraxor 2y agoI think this VR generation ends with "nobody gets the cookie". It is an experience to try it and you could do some fun things with it. But none of the VR vendors was focused on user demands. Valve probably had the most user orientation, but their devices had still a huge cost barrier. I had a Rift S (or still have) and I looked at how I could interface it for dabbling. Wasn't all bad, but I still lost interest immediately because of artificial vendor lock-in.