Deterministic physic and Bullet in T3D 1.1
by Sei · in Torque 3D Professional · 12/02/2009 (6:36 pm) · 3 replies
So I know GG is cooking up Bullet in 1.1. I have some questions about it.
1. Is Bullet officially supported or a sample implementation?
2. How many Bullet features are ported into 1.1?
3. Is it networked?
1. Is Bullet officially supported or a sample implementation?
2. How many Bullet features are ported into 1.1?
3. Is it networked?
About the author
#2
Cant wait for 1.1 :D
12/04/2009 (11:26 pm)
Sounds good. I have evaluated many game engines so far none provide satisfying networking/physic solution. If you can pull it off(networked physic), I am sure it will be a big selling point for many people.Cant wait for 1.1 :D
Associate Tom Spilman
Sickhead Games
Still working on it, but you'll at least have:
1. Static collision from terrain, TSStatics, etc.
2. Single rigid bodies via a new PhysicsShape object which works with both PhysX and Bullet.
3. A player controller.
I'm still getting the basics worked out, so i haven't looked at what Bullet can do with soft bodies, cloth, and its multicore support.
I'll try to get as much done as possible in the time we have left for 1.1. You can bet on seeing more in 1.2 and possibly Havok support as well.
Thats always the gotcha when you talk about physics... networking. So i can assure you this... statics and the player will be networked.
I'm still in the process of figuring out the networking support in PhysicsShape. I'm looking at a couple of options:
Server side only where the object is simulated on the server and the client only gets positional updates.
Client side only where the object only exists and is simulated on the client. These objects would be passive and will never block or support the player. Think smaller debris.
And possibly a mode where the object is simulated on the server and client with some corrections if the client side object gets too far out of place to the server.
I make no promises of what i'll get working... this is an extremely hard problem that no one has solved.