[ODE] ODE questions

[Jon Watte (ODE)]

Yes, ODE has improved since 0.5. However, it was perfectly possible to 
create stable solutions in ODE 0.5.

Something that most people who haven't done physics programming won't 
appreciate, is that physics programming in any library is mostly about 
mitigating instability, and tuning parameters to get the desired 
behavior in spite of the inherent inaccuracies present in discrete 
physical simulation. Getting a stable simulation on any library (ODE, 
Bullet, Havok, PhysX, ...) is not a matter of just "connect and go" -- 
you have to apply care, tuning, and experience to make it behave well.

Cheers,

          / h+


Enrico Zschemisch wrote:
> Hi,
>
> we are currently searching for a physics engine and want to choose now between Bullet and ODE. Unfortunately one of our programmers has experienced problems when applying constraints to joints, there were instabilities. This was with version 0.5, so I am eager to know if this has changed with the current version?
>
>
> Thanks, Enrico
>