[ODE] Worst Case Scenario
Graham Fyffe
gfyffe at gmail.com
Mon Feb 7 20:31:48 MST 2005
Seems to me one thing ODE is sorely missing is a dBodySaveState(some
struct pointer) and dBodyRestoreState(some struct pointer).
- Graham Fyffe
On Tue, 08 Feb 2005 09:17:57 +1100, Geoff Carlton <gcarlton at iinet.net.au> wrote:
> I do exactly that, but then its a crazy micromachines type game where
> the objects die when crushed and when they fall off anyway.
>
> In general, I found exploding almost always happens due to conflicting
> contacts. Particularly the edges of triangles would catch the bodies
> and with a few pointing in different directions the object would trip
> up, which looks odd even if it doesn't explode. In my case I stored all
> contacts for the object and then sorted through them, averaging or
> eliminating some of them (in my case, I had lots of separate trimesh
> geoms, so the bits of code inside ODE to sort through duplicates on a
> geom-geom basis didn't help much).
>
> You could always store a duplicate set of {pos,lvel,avel} and poke them
> into any body that acts up.. Might work, who knows? :-)
>
> Geoff
>
> mucki wrote:
>
> > Hi!
> >
> > I am using ODE in our 3d engine for physics, and I have already
> > added some code to improve overall stability (like linear and angular
> > dampening to improve auto-disable or contact count limits). But still
> > it might happen that ODE simulations explode. The question I have is:
> > What do you do in your games when ODE explodes ? When only secondary
> > actors are affected, then it is possible to just remove them from the
> > scene, or fake something, but what if the main actor explodes ? You
> > cannot remove a race car from the game, just because the collision
> > with an obstacle resulted in explosion :)
> >
> > mucki
> > _______________________________________________
> > ODE mailing list
> > ODE at q12.org
> > http://q12.org/mailman/listinfo/ode
> >
> >
>
> _______________________________________________
> ODE mailing list
> ODE at q12.org
> http://q12.org/mailman/listinfo/ode
>
More information about the ODE
mailing list