[ODE] the laws of timestep

[David Cobb]

I think the common solution to C is to run smaller several stepworlds 
with a fixed smaller timestep (you should always have a fixed timestep!) 
and then let the remainder time roll over to the next frame.  If this is 
unclear let me know and I'll go verbose.

vincent ngai wrote:

>Hi,
>
>I'm wondering if there are any hard and fast rules to govern time-stepping.
>Right now I'm using delta time every frame with 3 cases to consider.
>
>a) MIN_DELTA<CUR_DELTA<MAX_DELTA
>b) CUR_DELTA<MIN_DELTA
>c)CUR_DELTA>MAX_DELTA
>
>
>My question is actually for the case C, what is a good way of resolving it.
>Currently I'm running several stepworld with smaller deltas if I find
>that the delta time is too large. However this could again increase
>the delta time for the next frame etc. Should I just run stepworld
>with the smaller deltas and let the time taken up for that to vanish?
>
>
>
>
>  
>