[ODE] Auto Disable Improvement patch #1586738

[John Miles]

True, if you're concerned about the relatively-rare case of objects thrown
straight up in the air.  Otherwise the historical velocities won't cancel.

The heuristic of not autodisabling anything with no joints could be the best
way to deal with apogee problems.  I agree that average velocity
calculations aren't as advantageous there.

-- john

> -----Original Message-----
> From: ode-bounces at q12.org [mailto:ode-bounces at q12.org]On Behalf Of
> Jaroslav Sinecky
> Sent: Thursday, November 02, 2006 2:19 AM
> To: Christoph Beyer
> Cc: ode at q12.org
> Subject: Re: [ODE] Auto Disable Improvement patch #1586738
>
>
> You're right, it's not exactly the same but it's very similar in effect.
> You'll have to tune the velocity threshold and the number of steps used
> for calculating the average value for each particular case. Just
> consider our simple case of object thrown up in the air ... you have
> exactly the same problem as before, for certain number of steps the
> average velocity will be very close to zero, so for a number of steps
> relatively low in respect to your threshold value, the object will get
> autodisabled in the air.
> Well ... thinking about it now, the average velocity solution seems
> actually much worse for this case because for any number of steps
> chosen, there will always be a moment when the average velocity will be
> zero even if the object is going down with some high instantaneous
> velocity again. So anyway you would have to employ it together with some
> current criteria of momentaneous velocity or step/time counting.
> The same as any object undergoing some periodical movement will have
> average velocity zero all the time :-)
>
> Jaroslav
>
> Christoph Beyer wrote:
> > *Jaroslav Sinecky wrote:*
> >
> >> I would say that average velocity will not help, we actually
> already do that
>