hplus0603 said:
Are they (Second Life) still using their server-animation approach, where every object and every joint needs to be driven from the server every frame?

Second Life doesn't control character joints from the server. The server sends updates with position, orientation, velocity, angular velocity, and acceleration, and the client moves the object accordingly. Character animations are entirely client side; the server just tells a character to run a .BVH file, which the client does blindly, without any attention to collisions.

It's also not frame oriented. Server and client are not synchronized at all. Simulation frame rate is 45Hz, while refresh frame rate is up to the client. Which usually goes compute-bound making OpenGL draw calls from a single thread.

The result is that movement control is rather loose and soggy by modern game standards. Roughly the right things happen, but not with the precision one expects from a game today.

Sounds a bit similar to what I'm working on.

Metaverse, by Luca Grabacr. At the end, you'll see where this was made.

Nagle said:
That's a problem. Once you've built it, what do you do in there? The metaverse itself is indifferent to you, like a big city.

hplus0603 said:
I had become disillusioned with the chase for the metaverse.

Which is the core problem. How do we make the metaverse fun? I do not know how to answer that. Nor, as far as I can tell, does anyone else.

Current ideas in the press are not very good. Mostly, they're Make Money Fast schemes. In-game purchases of really overpriced NFTs, branding, etc. All the worst parts of middle school, in VR. Here's a piece of writing on that: https://blog.bitmex.com/rock-paper-scissors-says-go/​​​​ That is a depressing but necessary read.

It's noteworthy that the people making NFT noise all have bad or nonexistent 3D worlds, and the people with good 3D worlds don't mention NFTs much.

There are people who want to hold meetings in there. That's Facebook Workroom and Breakroom. OK, that's fine, but it doesn't need a whole big-world virtual world with real-time building. NVidia Omniverse is impressive, but it's basically doing for 3D programs what shared text editors do for text. That's a different kind of application. Good work, though.

I can sort of figure out how to get the performance of something as general as Second Life up to acceptable game levels. (Basically, they just need a technology refresh, with some extra server side stuff to do in semi real time the optimizations level builders in Unity and Unreal do during offline level building.) Roblox is taking something that already has good performance and moving it towards photorealism. All this is hard, but do-able.

But the classic question of the new Second Life user remains: “What do I do now?”

swiftcoder said: If they can pick up all that existing user engagement and transplant it into their nascent AR/VR worlds… their metaverse might just be sticky enough.

That's clearly what Facebook has in mind. So far, they've struck out twice, with Facebook Spaces and Facebook Horizon. Three times, if you count the Oculus acquisition, which remains a niche market at Facebook scale.

Virtual worlds just don't fit on phone screens. Social network engagement depends on “always on”, which requires a mobile device. You can be in the real world, or the virtual world, but not both at once. So immersive social worlds will be a place you go to for a while, like a club or bar, not full-time communication, like messaging.

Facebook, I have the horrible feeling, will end up offering some AR thing like Hyperreality. (Look up the video.) That's a well made dystopian video of artificial reality, heavy on the ads. That's totally Facebook's niche. That will probably come out as soon as, as John Carmack says, AR goggles get down to swim goggle size.

What does that leave as the niche for the virtual world/metaverse side? Virtual worlds seem to be an escape from the real world, like games, not an adjunct to it, like social networks. Not clear how this plays out.