4 ms·
I've lived through this with aframe.io and as part of the original WebVR team. It was one of the top frameworks with a large community, but content creators at
by ngokevin 5y ago
I've lived through this with aframe.io and as part of the original WebVR team. It was one of the top frameworks with a large community, but content creators at a nascent stage can't shift the balance without control of the browsers or spec
- prox 5y agoCan you elaborate? You mean Content creators need control of the browser (development) ?
- ngokevin 5y agoSome control over the direction of the spec to avoid situations like taking years to rewrite a working shipped spec, removing controller support, removing haptics, adding unneeded APIs, focusing on Cardboard and 3DOF, removing or not supporting link traversal, fuzzing headset tracking data. Or to be able to get useful features such as access to the headset camera or a way to bind textures without giant framedrops If there was to be a "standardized/decentralized" "metaverse", things would have to move much faster. Think of all the features a "metaverse" might want, and each item would have to go through spec gatekeepers that don't even use/develop VR
- prox 5y agoYeah that seems like a problem. In the early days of computing, you had consortiums competing against eachother. That might be helpful in speeding things up. I wonder if it is possible to build a Verse on something like Godot, which is open already and has VR tools.
- ngokevin 5y agoYou can build verses with anything, but something standardized and decentralized like the Web? Competition back then may have been possible, but it's daunting now when you have winner-takes-all corporations with Google and the Web, and Mark Zuckerberg with VR. You could do it with Godot, but distribution is also important. I built a VR browser before that browsed standardized VR worlds, but control of hardware also felt important. Perhaps there was a long-term path to success there.