Since the beginning this project has been a huge challenge in every aspect and our team, and especially I, have learned a lot in the process of making this possible.
The project is basically a huge physics simulation where you can control almost every aspect; we are using the unity game development tool and at the beginning the physx engine to power the entire simulation but while pushing the limits of what we could do with the engine we came to realize that we would eventually need more control over the simulation.
Among many physics engines available bullet seems to be the perfect option to take, it’s open source, it has a promising future and we would be able to make changes to the code if needed. The problem was that it represented a huge risk, the game was already running with unity physx, we already had the community playing with a pre-alpha build and the team was not in condition to suffer any more delays or disappointments.
The idea of using bullet was way too interesting to let it go, so I started preparing a simple plugin and playing around with simple stuff such as cubes and gravity, I found it to be really cool having cubes falling without using unity rigidbodies or colliders, I continued working on the plugin until it reached a point where as a team, we took the decision to actually put it into the game and make it our top priority as the only way to make a real progress in the development of the game.
The game uses simple building blocks to allow you making robots, most of the parts are simple shapes like cubes, cylinders and spheres. Those shapes correspond to bot parts like continuums, wheels or extensions, among others, in our game. Other parts where a little bit trickier to build.
The elbow, motor and piston, provide the basic movement to a robot. For all these parts we took advantage of the generic 6dof constraint and motors.
The piston was planned since the beginning, but we couldn't make it work stable enough using the unity physics, making it a proof of the advantages of using bullet. This powerful constraint, motors and some pids empowers the bots in our game.
In our game connecting parts is the most basic thing you can do, and all these connections are made, once again, using the generic 6dof constraint. It’s really cool to make a bot and think about the number of constraints that are actually working at the same time.
We see a promising future to bullet physics and we are looking forward bullet 3.x, since, from now on, our game future is deeply tied to the development of the bullet physics engine.
Finally, we have managed to release a new pre-alpha build with the bullet physics working and we’ve also launched a campaing in kickstarter to make this game what it deserves to be, you can help us on: http://www.kickstarter.com/projects/roz ... x?ref=live
Thanks to Mr. Coumans and the bullet community for all you help.
Regards Emilio Parker.
http://rawbots.net