[ODE] Setting orientation - Once before and once after attaching joints

[skjold@cistron.nl]

Actually, I suddenly have the strangest sense that my problem is just a silly mistake of mine. I forgot for a moment that rotation happens in body-coordinates (as opposed to world-coordinates). What I did was place one body to the left and one to the right, attach a slider joint, and then rotate both bodies around the vertical axis to give them some initial orientation (and position) on the ground plane. But what I didn't realize is that the left body would still be left of the right body after these rotations, which violated the constraint of that joint (a slider). I haven't gotten around to fix and test this, but it seems I already solved my problem.

About drawing the joints, that seems relatively easy, at least for illustrative purposes (as opposed to aesthetic purposes). But the catch is, the only way for some joints to determine their location/orientation (insofar as they have any), is indeed by inferring it from the bodies attached to them. And my problem was precisely with the location/orientation of those bodies. What I mean is that, when the constraints are violated to such an extent that the bodies are hurled into the void, which was the core of my problem, then drawing those joints in a way that makes sense becomes tricky too.


> > 
> > A small question about setting the initial rotation of an articulated structure.
> 
> You know, it would be helpful in the examples to draw the joint location &
> orientation, or even better have a mode where you can toggle it on and
> off.  I suspect we'd get a lot fewer questions of this type.  At the
> moment you need to infer the joint location & orientation from the body's
> behaviour, which can be tricky.
> 
> Just a thought,
> Martin
>