Difference in realization of shapes

[nitramdurcek]

Hi :)

I'm writing my Bachelor thesis and I'd like to explain difference between Pure, Convex and Random shapes. I understand when to use which one but In fact, I'm interested in some inner realization of these shapes - why is it recommended to use Pure ones etc. I guess it has something to do with communication with physics engines or something like that. I just dont want to write there totally brief info without explanation, when to use which .

Thx a lot, you rock Vrep !

[coppelia]

Hello,

There is the difference between pure shapes, convex shapes and random shapes, because physics engines can handle all of them in a more or less efficient/precise way (handle --> calculation of contact points and collision response):

• pure shapes. Those are primitives such as rectangles, cylinders or spheres. Since they are simple, it is possible to perform very fast and reliable calculations on them.
• convex shapes. Those are also quite simple, and there exist algorithms that can handle convex shapes better/faster than random shapes.
• random shapes. Since they are random, it is difficult to handle them efficiently. Also, most of the time, random shapes will not be that stable as pure or convex shapes.

To illustrate this, you have to know that a physics engine, in order to correctly calculate collision points and contact forces often needs to know the penetration depth between two shapes. But what is the penetration depth between two random shapes? Typically, what is the penetration depth between two flat triangles? Unless an algorithm is working with time coherency, there is no answer to that.

Cheers

[nitramdurcek]

Actually, second part of the answer helped a lot. I just didnt want to write about shapes usage with no clue whats happening behind. What you wrote is quite enough.

Thx a lot !