[ODE] orientation in ODe

[Hannes Norda]

Well actually I am a bit confused. I have read about ODE not having 
"handedness" and I do not convert my data when I build my geoms. I build 
trimesh geoms with data directly from my vertex/indexbuffers and 
everything simulates just fine. Gravity in my world is (0, -gravity, 0).
What I DO convert is matrices (as written in 
http://q12.org/cgi-bin/wiki.pl?FrequentlyAskedQuestions) I get from 
geoms for rendering. I supposed ODE computed normals internally for 
collisions etc, but I thought these would match the data I sent in, in 
my coordinatesystem...

Well, anyway, I mainly wanted to check if I collided with a face 
pointing up or down. Putting the collision-normals z-component in the 
y-component of the vector and dotting that with the negative nomalized 
gravity vector gave me correct results...

I guess I am just doing something wrong, I didnt look too much into it 
yesterday and now it's pre-morningcoffee-time for me here, so I might 
just be confused here... Thanks for the reply!

Hannes Norda
Custom Red Software Studios
norda at customred.com



Jon Watte wrote:

>If it points in the Z direction, then it's because something collided with something such that a Z component was generated.
>
>If you don't convert your data (except for gravity) when putting it in, then you don't need to convert it when taking it out. If you DO convert it when putting it in, you have to convert it when taking it out -- except there's so many cases, you're likely to get it wrong somewhere. Thus, I recommend not converting, neither when putting data in nor taking it out.
>
>Just think of ODE as working in your coordinate system, and put in data using your coordinate system; you'll get data out in your coordinate system, as well!
>
>Cheers,
>
>			/ h+
>
>
>-----Original Message-----
>From: ode-bounces at q12.org [mailto:ode-bounces at q12.org]On Behalf Of
>Hannes Norda
>Sent: Monday, March 29, 2004 1:45 AM
>To: ode at q12.org
>Subject: [ODE] orientation in ODe
>
>
>Hi!
>Quick question I hope...
>
>I know that orientation in ODE really doesn't matter, since simulations 
>behave as expected depending on how you set the gravity... But, I 
>noticed that when retrieving the collision normal from dContactGeom the 
>z component points up. (I use DirectX-style orientation btw.) So I guess 
>I need to convert the normal to my space. Now I'm just curious, where 
>does x and y components point??
>
>/Hannes
>
>
>
>_______________________________________________
>ODE mailing list
>ODE at q12.org
>http://q12.org/mailman/listinfo/ode
>
>
>
>  
>