At 6/22/10 11:25 AM, Glaiel-Gamer wrote:
I tried and failed 4 times on writing a physics engine before deciding to use Box2D (c++), but the knowledge gained through those failures is invaluable.
Whole lot you learned if you then chose Box2d and thought it's good!
At 6/21/10 10:54 PM, Toast wrote:
i find that often times it's annoying to program physics on flash, because the actual physics are very simple, but it becomes inelegant and ugly when you code it in flash, especially when angles are involved.
Physics mostly deals with simple illustrative situations. When programming you're faced with full complexity and you have to deal with it such that you don't become hindered by bad code.
If you're working on a game, if it's applicable you should use physics engine. You may be good in physics class or you may believe you have potential to build your own engine, but in truth at that point you don't really know the first thing about physics simulation, so you have no more reason to try to make your own engine than anyone else.
If you're looking to learn about physics simulation, starting blank is *still* reinventing the wheel. As you'd work, you'd encounter various issues to address and then refactor your design of the engine to better reflect the task as you come to better understand it.
It's the saying that simplicity is the result of effort. In the case of physics engines, the resulting design is fairly simple, but the road to figuring it out is long.
And this road has already been treaded for you. So treading it again is a waste of time, you're just ignoring a map that's already drawn.
If you want to learn about simulation, just inspect the current engines. The best way would be to create your own by adapting from other ones. You'll assimilate all the ideas that have shaped them.