[PD] pmpd bug?

[cyrille henry]

marius schebella a écrit :
> hi cyrille,
> I was playing around with some collisions and found a strange
> behaviour which looks buggy to me.
that look buggy to me.
but the bug is not in pmpd, neither in your patch, it's just the pmpd 
physics that is an aproximation of reality. but this can be tix easilly.
in fact some parametters are not well adjusted.

> I was shooting a sqare against a
> circle and at some points the rebound went to strange directions.
it's not square : the masses have no size. better to see them as small 
sphere.
> I tried to make a debug patch to show this behaviour
> (circle_testpmp.pd) you will also need the other abstraction to run
> it.
> If someone else wants to look at it, please do!
> thanks for any help, or corrections. (btw. I also found this strange
> behaviour with rebounds from lines, so maybe it is not a bug, but I am
> just doin something wrong).

in fact, you send a big force to the masses when it interact with the 
green circle. and you also have a very big rigidity of the same green 
circle.
this made the small masse to interact with the circle for only 1 
iteration, but somtime 2. it receive lot's more force when interacting 2 
time better than 1 time.
so the movement was diferents.
the problem comme from the time discretitation that is only an aproximation.

i remove the normal force, and reduce normal rigidity in order to allow 
a "better" interaction.

look like working great now.

cyrille

> marius.
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> PD-list at iem.at mailing list
> UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: circle_pmpdtest.pd
URL: <http://lists.puredata.info/pipermail/pd-list/attachments/20061115/4a43c2a9/attachment.txt>