[ODE] Geist3D, ODE and Open-source

[Jordi]

It's windows only software?



Saturday 08 April 2006 05:48、jstier at cs.uvic.ca さんは書きました:
> Hi Everyone,
>
> For the past five years I have been developing a virtual reality authoring
> tool and engine (www.geist3d.org) as part of my research. I have also used
> to ODE since 2002. Just recently I have first released Geist3D in a
> news-feed at opengl.org. I was surprised by the traffic this has produced.
> The internet is bigger than I thought. In this light I have decided to
> Open-source Geist3D as well. Now I have two problems; how to properly
> integrate ODE and under what license to release it.
>
> Since I have implemented my own scene graph there is a lot of overlapping
> data. For each data member in the ODE engine there exists the same member
> in my own scene tree nodes. The problem is that I also have Python bindings
> and Python requires that you pass to the interpreter a struct containing
> the members in order to ensure fast access. In ODE all of the members are
> hidden and they cannot be presented in the way that python expects them.
> So, I am considering to remove the outer layer of ODE and replace/merge the
> structs i.e. dxBody, dxJoint with my own node types. This way, they will
> only have one memory address, ensuring good performance and consistency.
> There is a similar problem with the bounding boxes, because they are
> important for occlusion culling which, as far as I know, ODE does not do.
> Now, how can I combine ODE and Geist3D so that any further development
> benefits both sides. Not to mention all of the discussions about Bullet and
> GJK... If I do this, how much support can I expect from the ODE community
> and how can we combine our efforts.
>
> The other question I have is what open source license to use. I know that
> ODE is LGPL, i.e. do what you want! Geist3D, however is an authoring tool
> with a GUI and a runtime engine. Does this license make sense here ? What
> are the implications and what does it mean.
>
> Anyways, if you have some time and a good graphics card have a look at
> Geist3D. It was intended as a mechatronics simulation tool. Basically you
> have a virtual environment which reflects the effect of mechanical
> actuators -- obviously by leveraging ODE to a large extent. Using the
> sensor telemetry from the environment you can then develop a control system
> using Petri Nets and Python, which in turn changes the virtual environment.
> Basically, the simulation models the same feedback loop that you will find
> with between a real mechatronic system and its environment... at least to
> some extent.
>
> The installation of Geist3D includes a bunch of interesting models. When
> you go to the website you will find Geist3D and Geist3D-programmable under
> the download section. The former is more intended to show that GLSL shading
> language features, but there is a ragdoll demo, a domino and a box stack.
> The latter contains the really interesting stuff as far as rigid body
> dynamics goes. There is a small factory and a landscape with autonomous,
> four wheeled, vehicles.
>
> Cheers
> Jochen
>
>
> P.S. I have already outstripped my bandwidth with the ISP. Their billing
> system is down and I cannot purchase any more. But they guaranteed that the
> site will remain up an running. If you cannot get to www.geist3d.org let me
> know and I will repost this message again :)
>
> P.S.S. The code is not there yet. I just yesterday decided to open-source
> and there is soooo much cleanup to do!
>
>
> ----------------------------------------------------------------
> This message was sent using IMP, the Internet Messaging Program.
> _______________________________________________
> ODE mailing list
> ODE at q12.org
> http://q12.org/mailman/listinfo/ode