Re: quadratic planes

It looks like a good solution to my problem.
Thanks


Jon Watte (ODE) a écrit :
>
> If you want parametric solutions (rather than a mesh approximation),
> your best bet is probably to calculate the intersections yourself, and
> generate contact joints based on your calculations. Sphere contained in
> cylinder is fairly simple to solve analytically -- but if you built an
> "inverted cylinder" geom, you'd have to handle rays, boxes, trimeshes,
> and a few other kinds of geoms, too.
>
> There is nothing in ODE that forces you to use dCollide. You can create
> collision joints from data that you calculate yourself.
>
> Cheers,
>
>           / h+
>
>
> Benoit Forestier wrote:
> > Hello,
> > firstly, i would like to apologize for my poor english.
> > I would like to simulate  the motion (and collisions) of spheres 
> within an
> > empty cylinder.
> > from what i saw in the user guide, there is no geom type i can use to
> > perform that kind of calculation.
> > has anybody heard about any quadratic plane geom class (i.e. a plane 
> with
> > equation of the form ax^2 + by^2 + cz^2 + dxy + exz + fyz + gx + hy 
> + iz + j
> > = 0)?
> > should i  build my own class?
> > is there any other way to model a cylindrical vessel?
> > Thanks
> >  
>