4 ms·

You can have low-bandwidth realism with a heavy client, for example you can enumerate all people on earth with 33 bits. But to me, video games with fancy CGI i

You can have low-bandwidth realism with a heavy client, for example you can enumerate all people on earth with 33 bits. But to me, video games with fancy CGI is like a chessboard with ornate gold pieces. Impressive, but in the end it's still just chess.

bullen 5y ago
My 3D MMO engine can now have 2000 non-instanced animated characters each with an interchangeable item equipped at 60 FPS without a world on a 1050Ti (80W GPU). My target is to keep that number above 1000 when the world + physics is in. On Jetson Nano (1/2 Nintendo Switch at 5W GPU) that number is 300 and I hope 150 on release! The server can handle 10.000 action players (no spraying/headshot but mario/zelda mechanics/physics) at an avererage of 50 players in view on a large AWS machine. Nobody has ever played an action MMO (with fun gameplay and physics) that plays like mario but scales to 10.000 players yet. When they do, they won't play anything else ever again (or do anything else in meatspace) is my prediction. You are spot on; gameplay is more important than surface.
- ikrenji 5y ago
  you can have a good game that plays well and you can have a good game that looks pretty. but the truly great games both play well and look pretty...
  - bullen 5y ago
    Yes, but my game will have real modeled characters with silky smooth AAA animations, and still scale! Roblox and Minecraft could only build the future by spending less time on fidelity, hence the awkward animations and "ugly" characters.
- arminiusreturns 5y ago
  Are you doing your own engine or what scaffolding are you using? I'm assuming something that is pretty close to end vulkan? This is a topic close to my heart, lots of questions about how to scale, where to put the seams (node based loading has issues, have to be careful about loading screens, etc), how to keep player state backend access super fast once attributes bloat, how to keep low-lag player interaction when you want to force as close to a pure server-dominant state decision layer. I think the tech itself is there just about, but there are some real technical limitations to why it hasn't been done (well) before. I'd love to hear how you are addressing these things. As for the conversation about the matrix and the metaverse, my conclusion a long time ago was that the real matrix/metaverse will be FOSS. Thats why all these big companies (Meta, Epic, etc) dipping their toe into this admittedly juicy arena are going to fail in their vision. (also one of the reasons why I still have hope for a hacker culture of freedom in the future) be warned though: with all this censorship they are coming for your game worlds next
  - bullen 5y ago
    I stopped at OpenGL (ES) 3 that has VAO (the last meaningful feature, internal improvements omitted, those can come later behind the scenes), Vulkan only scales the embarrasingly parallel stuff which does not impact gameplay. (Mostly visual, zero motion-to-photon improvements have been made and again the meaningful practical internal improvements can be added retroactively in the OpenGL drivers.) The server is allready open-source (under mixed algpl/mit/commercial license): http://fuse.rupy.se/about.html http://fuse.rupy.se/about.html The last engine has to be modifyable and resellable, so the ownership model should probably avoid corporate models seen the legal/bureaucratic problems that brings. Right now I'm keeping all rights for myself and my son will probably inherit those. I'm still experimenting with the licence pricing that you can find on my github sponsors page. The b2c product should cost ~$10/year and gives you the right to mod and redistribute for free independently of me, something both Minecraft and Roblox lack. Unfortunately the client will initially only be partially open-source; keeping the core engine .exe closed, and the game specific .dll open-source as an implementation example. Also the cl.exe compiler is not redistributable, so I'm looking at solutions (use another compiler for the game .dll) for that until people move to linux: https://developercommunity.visualstudio.com/t/allow-redistribution-of-the-clexe-compiler/975508 https://developercommunity.visualstudio.com/t/allow-redistri... The server is allready prooven in all technical aspects with 350.000 b2c customers, and the client is months away from PoC: http://talk.binarytask.com/task?id=5959519327505901449 http://talk.binarytask.com/task?id=5959519327505901449 They can't censor TCP/IP, or atleast not in a way I can't work around with my own encryption. The internet removes that possibility permanently, it's binary: either you have global connections and then the world is open or you don't and then we're all screwed back to the stone age.

bullen 5y ago

My 3D MMO engine can now have 2000 non-instanced animated characters each with an interchangeable item equipped at 60 FPS without a world on a 1050Ti (80W GPU). My target is to keep that number above 1000 when the world + physics is in. On Jetson Nano (1/2 Nintendo Switch at 5W GPU) that number is 300 and I hope 150 on release! The server can handle 10.000 action players (no spraying/headshot but mario/zelda mechanics/physics) at an avererage of 50 players in view on a large AWS machine. Nobody has ever played an action MMO (with fun gameplay and physics) that plays like mario but scales to 10.000 players yet. When they do, they won't play anything else ever again (or do anything else in meatspace) is my prediction. You are spot on; gameplay is more important than surface.

ikrenji 5y ago
you can have a good game that plays well and you can have a good game that looks pretty. but the truly great games both play well and look pretty...
- bullen 5y ago
  Yes, but my game will have real modeled characters with silky smooth AAA animations, and still scale! Roblox and Minecraft could only build the future by spending less time on fidelity, hence the awkward animations and "ugly" characters.
arminiusreturns 5y ago
Are you doing your own engine or what scaffolding are you using? I'm assuming something that is pretty close to end vulkan? This is a topic close to my heart, lots of questions about how to scale, where to put the seams (node based loading has issues, have to be careful about loading screens, etc), how to keep player state backend access super fast once attributes bloat, how to keep low-lag player interaction when you want to force as close to a pure server-dominant state decision layer. I think the tech itself is there just about, but there are some real technical limitations to why it hasn't been done (well) before. I'd love to hear how you are addressing these things. As for the conversation about the matrix and the metaverse, my conclusion a long time ago was that the real matrix/metaverse will be FOSS. Thats why all these big companies (Meta, Epic, etc) dipping their toe into this admittedly juicy arena are going to fail in their vision. (also one of the reasons why I still have hope for a hacker culture of freedom in the future) be warned though: with all this censorship they are coming for your game worlds next
- bullen 5y ago
  I stopped at OpenGL (ES) 3 that has VAO (the last meaningful feature, internal improvements omitted, those can come later behind the scenes), Vulkan only scales the embarrasingly parallel stuff which does not impact gameplay. (Mostly visual, zero motion-to-photon improvements have been made and again the meaningful practical internal improvements can be added retroactively in the OpenGL drivers.) The server is allready open-source (under mixed algpl/mit/commercial license): http://fuse.rupy.se/about.html http://fuse.rupy.se/about.html The last engine has to be modifyable and resellable, so the ownership model should probably avoid corporate models seen the legal/bureaucratic problems that brings. Right now I'm keeping all rights for myself and my son will probably inherit those. I'm still experimenting with the licence pricing that you can find on my github sponsors page. The b2c product should cost ~$10/year and gives you the right to mod and redistribute for free independently of me, something both Minecraft and Roblox lack. Unfortunately the client will initially only be partially open-source; keeping the core engine .exe closed, and the game specific .dll open-source as an implementation example. Also the cl.exe compiler is not redistributable, so I'm looking at solutions (use another compiler for the game .dll) for that until people move to linux: https://developercommunity.visualstudio.com/t/allow-redistribution-of-the-clexe-compiler/975508 https://developercommunity.visualstudio.com/t/allow-redistri... The server is allready prooven in all technical aspects with 350.000 b2c customers, and the client is months away from PoC: http://talk.binarytask.com/task?id=5959519327505901449 http://talk.binarytask.com/task?id=5959519327505901449 They can't censor TCP/IP, or atleast not in a way I can't work around with my own encryption. The internet removes that possibility permanently, it's binary: either you have global connections and then the world is open or you don't and then we're all screwed back to the stone age.