How often are physics forumlas used in 3D programming?

[Bullet Pulse]

Like the title says, how often are physics formulas used in 3D programming?
For example, would you more commonly use an iterative solution find an object's location over time, or a formula that calculates it at time x?

[Arce]

I spent a good 20 mins typing a nice, detailed response showing how you could achieve the same thing either way or use the two interchangeably, and stating that it really just depends.

But, since it took so long to type, when I hit sumbit, the motherfucker made me relogin, thereby clearing the entire form. GODDDDDAAAMANIT.


So, somebody else take this one, please?

[K-Bal]

If you have interacting objects I don't see how you could solve this with a non-iterative approach.

[Arce]

I'll type ya a response tomorrow, I hope.

Sorry man.

[Falco Girgis]

If you have interacting objects I don't see how you could solve this with a non-iterative approach.

Exactly.

You'll find that 99.99% of actual "formulas" are useless in game programming, because they are all a function of time. In a video game, every frame is another timestep, so the only place that these formulas even really make sense is going from one frame to another.

The math is still the same and the concepts are still extremely important--even more important since you can't blindly use a formula you're given to determine a value.

[qpHalcy0n]

In 3D programming you'll see them quite frequently. However, you often will not see them in "textbook form". The methods you learn about in the classroom are exact methods which are often very computationally expensive. You're more likely to see the same formulas but widdled down under "assumptions" that reduce the complexity of the formulae. For example, under the assumption that delta time remains constant or some other action, direction or force remains constant.

Interestingly, modern graphics hardware implement partial derivative functions which will give you average rate of change over X and over Y in a given domain natively. They're somewhat expensive but they're pretty damn handy in terms of predicting from motion buffers.

[davidthefat]

There is an article in the Game Developer Magazine about physics

[Falco Girgis]

If you have interacting objects I don't see how you could solve this with a non-iterative approach.

Exactly.

You'll find that 99.99% of actual "formulas" are useless in game programming, because they are all a function of time. In a video game, every frame is another timestep, so the only place that these formulas even really make sense is going from one frame to another.

The math is still the same and the concepts are still extremely important--even more important since you can't blindly use a formula you're given to determine a value.

I would actually like to partially retract this statement. What I said wasn't completely true. 99.99% of the time, functions of TIME are not used. General equations that are instantaneous such as the law of conservation of momentum are used constantly (albeit in different forms than you see in your textbook as qp has said).