[ODE] Machine to fast and a possible bug?

[Gerhard W. Gruber]

On Thu, 16 Dec 2004 09:53:36 +1100, Geoff Carlton <gcarlton at iinet.net.au>
wrote:

>I'm not sure about the bugs, but as for the sphere stacking, I noticed 
>that just yesterday myself, and my immediate reaction was, cool!

:) Yeah. It looks pretty. Especially when you stack a lot of them and then
introduce an error and they tumble down. I love this test application. I can
just play with it and watch the objects tumbling around. Well, seems I'm
easily entertained. :)

>When you look at a stack of spheres that are mathematically positioned 
>on atop the other, in the absense of wind, imperfections in the spheres 
>themselves, etc, they would actually stack like that.  I'm impressed 
>that it works for ode, I expected slight jitters to cause them to fall 
>over.  Certainly last time I stacked boxes on a trimesh, they jittered 
>and fell over, which was somewhat disappointing.

For boxes this would be annyoing. I don't mind if ODE can do this, just wanted
to verify that this is no bug.

> Or perhaps if ode is too accurate, introduce an ode patch that adds random noise to each contact? ;-)

I don't really think this is neccessary. If I'm aware of this I can easily
circumvent it in application code, and if the behaviour is correct, which it
is, then I don't think that ODE should be held responsible for it. After all,
ODE is general and "fixing" this might not be desired in all situations.

-- 
Gerhard Gruber

Für jedes menschliche Problem gibt es immer eine einfache Lösung:
Klar, einleuchtend und falsch. (Henry Louis Mencken)