[ODE] Auto Disable Improvement patch #1586738

Francisco Leon projectileman at yahoo.com
Thu Nov 2 09:46:51 MST 2006


--- Kyle Hubert <khubert at gmail.com> wrote:
> What happens when a body does have a contact joint
> and it is still
> following a parabolic path? For instance, a box
> slowly tipping over on
> one corner typically auto disables while reaching
> it's zenith because
> it's linear and angular velocity is small. If a user
> suspects that
> they can increase the global disable velocities with
> this patch, this
> problem would become more prevalent. No?
> 
> -Kyle
> _______________________________________________
> ODE mailing list
> ODE at q12.org
> http://q12.org/mailman/listinfo/ode

No, may you've misunderstood how the auto-disable
system works. 

Objects can move following a parabolic path.
Auto-disable system activates when the velocity of  an
object remains as zero lenght during some time.
 
In some times game designers apply greedy strategies
for reducing processing work and increase the fluidity
of the animations. One of them is to force non-dynamic
bodies (Like decorative objects, rocks, statues or any
enviroment artifacts etc.) to sleep more quickly.

But happens that the old velocity-autodisable
criterion lacks when a non-dynamic object is throwed
vertically with an straight path. When the object
reachs the top altitude, it freezes (stops).

For avoid this on common simulations with gravity,
decorative objects shouldn't sleep until they touch
the floor or something in the ground. Is a trick, no
science behind this.

On outer-space enviroments, is stupid to apply the
auto-disable criterion, because in the space objects
are  constantly moving all the time.

Machines and animated characters need to disactivate
the autodisable system. 

"Technocracy Rules with Supremacy"
Visit http://gimpact.sourceforge.net


 
____________________________________________________________________________________
Cheap Talk? Check out Yahoo! Messenger's low PC-to-Phone call rates 
(http://voice.yahoo.com)



More information about the ODE mailing list