HikeR
Почитав форум, насчёт колёс, понял что у них трудности. Они не могут найти другого решения. А насчёт неуправляемой Celesta Bollide, был ответ: Yes, it veers at 100 mph. If we could wave a magic wand and fix it we would.
Есть также пост о физики, ещё от одного разработчика:
First is that everything in the physics engine is based on physics theories from textbooks and papers. For each and every physics phenomenon that we simulate, an immense amount of studying and research has been done so as to simulate it as close to physics theory as possible (given the processing limitations).
This means that the physics simulation does not come from our opinion or "taste" of how things should behave, but from real physics equations.
Due to above, it is very difficult for us to make physics "mistakes". Most of the time, we know that something does not follow the proper physics equations, because things tend to explode (if the system's energy is not conserved correctly for example), or behave in crazy ways (cars sliding all around by themselves).
Concerning gravity, our gravity acceleration is at -9.81 m/s^2 . It is a little bit bigger than the real life measurement of earth's gravity acceleration at -9.80665 m/s^2 . But other than the 0.04% difference, which you are not able to see in game, the number is correct.
Concerning inertia, if our inertia calculations weren't correct the whole thing would explode or it would behave in *very* strange ways.
Concerning friction. A normal car (not a dragster) is able to pull up to 1G turns. If you go to the wheel debug screen, and look at the G-meter in there you can confirm it. The main problem we had with the previous wheels (before the 0. 3.03 update), was that they didn't had correct weight distribution resulting in a lot bigger gyroscopic forces (which tend to oppose turning). Fixing this mistake improved the driving behavior.
There might be other mistakes that we are unaware of, and when we find them we'll fix them. My guess is that the remaining mistakes are not so serious. Because if they were, they would also have serious/obvious side effects which means that we would have found them a long time ago.
We also have erroneous behavior (which we know about) that we haven't yet found a way to simulate adequately enough.
A good example of that is the parking brakes. Parking brakes are extremely hard to simulate correctly (without faking them). The "jitter" that you see in BeamNG' parking brakes is due to us not wanting to "fake" their simulation. Nevertheless, even if our parking brake simulation is not perfect, it can still hold a vehicle (which took an immense amount of work to achieve in a physically correct way).
Изменено: Craftworth, 29 Сентябрь 2013 - 19:36