Owen - Have you ever seen our work with Flatland (from UNM)...
This is the only link I have right now (pretty old)... http://www.ahpcc.unm.edu/homunculus/ Flatland is a VR/Simulation framework developed by Tom Caudell & Co... which we adopted for doing not only visualization but coupled simulations.... We added a layer (Flux) for data flow management and Tom has his own variant (E-Loom) which he uses for Neural Nets and similar. It is implemented in C/OpenGL and has a simple C-binding... allowing dynamic loading of multiple applications... you register a handful of callbacks that get executed at various appropriate times (first startup, shutdown, before any drawing happens in the render-loop, during the render loop, for providing realistic shadows, etc. It is not well documented but there are a number of us (LANL, Maui SCC, Uniformed Medical Services, etc...) who have invested a bit of time in adopting and extending it. Among other behaviour-driven systems, we have a network/graph/tree/heirarchy layout system under development. - Steve On Oct 19, 2006, at 10:37 AM, Owen Densmore wrote: > This may seem a bit odd, but bear with me! > > We've been using Processing.org's great graphics system: the > libraries and the nifty IDE and tools for managing > "sketchbooks" (projects), building web pages/applets, and even > building applications for Windows, Mac and Linux. We've had success > (i.e. got paid for!) two projects using Processing: a Stadium model, > and a generalized Data Visualization system. Quite nice. > > So now we've got two interesting environments for modeling: NetLogo, > our old friend which keeps getting better, and Processing which seems > great for what I'd call "wire frame" modeling. We've also got high > end rendering experience with Blender: we can both use it to build > Processing meshes for our models and can render agent motion inside > blender, using NetLogo and Processing output. > > BUT we're missing a "sweet spot" in the middle: a fairly realistic 3D > environment that can do realtime modeling .. i.e. animation via > behavior. We also want to have some notion of "physics" .. i.e. > things bouncing off walls or agents colliding. (Blender thus far has > not worked, but we're still poking.) > > This prompting me to look into Java graphics and game books, one of > which is Killer Game Programming in Java. The book has a website > which includes a LOT of material that is not in his book: > http://fivedots.coe.psu.ac.th/~ad/jg/ > This let to getting in touch with the author, asking for pointers to > "game engines", see email attached. > > Now to the question to FRIAM: Has anyone found a good environment > for agent based modeling with "game-like" 3D realism and with modest > libraries for collision detection, scene graphs and so on? > > -- Owen > > Owen Densmore http://backspaces.net > > Begin forwarded message: >> From: "Dr. Andrew Davison" <[EMAIL PROTECTED]> >> Date: September 4, 2006 10:56:45 PM MDT >> To: Owen Densmore <[EMAIL PROTECTED]> >> Subject: Re: Processing.org >> >> >> Owen, >> >> Thanks for the pointer to processing.org. It's a very nice system. >> >>> We're also looking for Java "game engines" to make our work >>> simpler. Do you have any pointers? >> >> Here's a snippet from an article I'm writing: >> >> Xith3D (http://xith.org) uses the same basic scene graph structure >> as Java 3D, but can also directly call OpenGL operations. Since the >> high-level APIs of Xith3D and Java 3D are so similar, porting Java >> 3D code over to Xith3D is fairly straightforward. There are >> versions of Xith3D that run on top of JOGL and LWJGL. >> >> jME Graphics Engine (jMonkey Engine, http:// >> www.mojomonkeycoding.com/) was inspired by the scene graph engine >> described in 3D Game Engine Design by David H. Eberly (http:// >> www.magic-software.com/Books.html). jME is built on top of LWJGL. >> >> JAVA is DOOMED (http://javaisdoomed.sourceforge.net) includes >> loaders for Quake 2 MD2 and 3D Studio Max 3DS files . The >> implementation uses JOGL, and the distribution includes Escape, a >> Doom-like game. >> >> Aviatrix3D (http://aviatrix3d.j3d.org/) is a retained-mode Java >> scene graph API above JOGL. Its tool set is aimed at data >> visualization rather than gaming, and supports CAVEs, domes, and HMDs. >> >> JView (http://www.rl.af.mil/tech/programs/JVIEW/) is another >> visualization API, supporting both 2D and 3D graphics, developed by >> the U.S. Air Force Research Lab. GL4Java, an older low-level Java >> API for OpenGL, was used to build it. >> >> Espresso3D (http://www.espresso3d.com/), a games-oriented library, >> includes OpenAL audio, sprites, collision detection, input, and >> rendering support. It's built using LWJGL. >> >> - Andrew >> > > > ============================================================ > FRIAM Applied Complexity Group listserv > Meets Fridays 9a-11:30 at cafe at St. John's College > lectures, archives, unsubscribe, maps at http://www.friam.org ============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org
