[ODE] executive toy

[Michael Lacher]

Nguyen Binh wrote:

> Hi Michael,
> 
>    Haven't try so far but what is your problem?
>    I'm afraid that we need a special case to correctly model the
>    energy lost after each collision.
> 
>    It's better if you could put a demo online...
> 
I wouldn't mind if there is no energy lost (it might well go on 
forever), but shouldn't the bounce param of the contact handle this anyway ?

The problem is that the balls seem to tend to "stick together". That is: 
some or all of the balls move as groups, as if they were sticky, 
although they shouldn't.

A second problem i experienced is, that with dWorldQuickStep, the balls 
behave very erratically. even with only a single ball (that is: one 
ball, one hinge joint to connect it to the static environment) and 
nothing else, the ball starts jittering and sometimes explodes ...

mucki