today ode is a standard for making physics in games, it was used in many comercial games like xpand rally or bloadrine and i was thinking about making it available to applets as jogl using joglAppletLauncher, have anyone thought about it. So we would be able making unsigned applets with jode.
are you talking about JOODE or OdeJava?
JOODE is pure Java, so there should be no problem.
you are right there is a pure java physics imlpementation, sorry my mistake , but in joode project web site there are not any files:
https://joode.dev.java.net/servlets/ProjectDocumentList?folderID=0&expandFolder=0&folderID=0
please someone update in above site that joode is available here;
https://sourceforge.net/projects/joode/
thanx arne for your time, joode it was what i looked for
No File Packages - sourceforge have not got any files for this project or i am wrong .?
but this link works:
http://homepages.inf.ed.ac.uk/s0570397/joode-src.tar.gz
but is it the latest joode?
JOODE is a very new project, we’re porting ODE to java - until now there’s only cvs.
The up-to-date cvs lies at sourceforge, because java.net took so long to approve the project.
several types of joints, Sphere-Sphere and Sphere-Box collisions and SimpleSpace have been implemented yet.
So probably not yet usable
EDIT: dunno about that tar - we’ve not created it - I’ll check
have you got any links to demos of joode use? appleats or programs?
it is old?
so where to find the latest joode?
you would have to download the cvs from
https://sourceforge.net/projects/joode/
there are test-packages with test-classes.
you’ll need a cvs client to download the cvs-repository
there are no builds yet - interesting that there was already one “unofficial one” in the net 
EDIT: Ahh it was probably toms own webspace
thanx arne you were very halpfull, it is very intresting project, you are making great job with joode, thanx
Sorry … yeah its confusing. Sourceforge is the current host of the project but we plan to migrate … probably when version .5 is ready.
At the moment the rigid body side of things is working, including unpowered unlimited joints. However we only have code for sphere collisions so that limits the scope somewhat. I am working on proper collisions at the moment.
it is one of the most for me intresting project i have found recently, but you could make some version in one file (sources and binaries) not splited to several files so people could have easier examine your project, now only way is to use that cvs program, well it is just an idea, but there is much of free space in the net…
Yeah don’t worry. We will get it all packaged up once we have a version worth using. That time is fast approaching …
i have made recently some investigation in internet about computer physics and found many articles worth reading. Many people claims that the future of physics are physics graphics cards like AGEIA (old novodex), video demos are realy impresive, but i think that the future of physics are not seperate cards (who would like to spend addtional 400$ or maybe even more?) but additional extensions for future graphic cards that give developers an ability to program physics in gpu.
Even today graphic cards can also be used to program physics, check out this article:
http://www.shaderx2.com/shaderx.PDF - it shows some simple example of falling cubs and uses the easiest way to use of LCP alghoritm, what is the most intresting that unoptimized alghoritm is really fast - about 2x times faster than in cpu. More intresting info about for what can be used gpu please go here: www.gpgpu.org, and especialy here: http://graphics.stanford.edu/~mhouston/public_talks/R520-mhouston.pdf. Well generally gpus are about 10x times faster in such operations like memory usage and mathematics operations and it is nothing suprising that such company like havoc announced their support for gpus.
But what is the most intresting ati said that new generation of their cards will support many advanced physics faetures. The latest card x1800 have very intresting demo of water waves and droplets on glass. For more information please go here: http://techreport.com/ja.zz?comments=8887 .
I am talkimg about it becouse joode project intrested me very much. Probably the majority of joode users will use 3d garphics api like jogl / lwjgl so maybe some the most important functions of joode could be implemented in graphic hardware like in that paper: http://www.shaderx2.com/shaderx.PDF . Today the biggest problem for physics is to make a good Lcp solver, using gpu it can be done preaty fast, so maybe joode in future could implement such support for graphic hardware processor usage. Many peolpe in this forum talked about joode support for AGEIA etc, i think it is not worth to even think about it. Implemeting gpu physics in joode will make true crossplatform hardware accelerated physiccs engine and who knows maybe faster then odeJava if the most computinal expensive alghoritm would be implemented in gpu. It could be RAW POWER!!!
Could it be possible in joode, becouse most of i users will use jogl / lwjgl so it will maybe not so hard to connect it with gpu. What do you think?
And who knows maybe some future cards will have some hardware support for Lcp computation, joode could take advantage of it.
mmh to combine it with jogl or lwjgl for using the gpu for physics is a nice idea. We would even be able to keep it pure java But because we want to stay independent of graphics renderers we can make it as a speed-up that can be “plugged in”, when it is possible. There are still people (like me), who aren’t able to use pixel shaders.
Because you seem so interested about this project, I uploaded the latest cvs to my own webspace. (Tom, do you know how I can upload files for joode?)
So here it is (It contains also some changes I made, that are not yet on the cvs, because they are either unnecessary or don’t work, but they should not matter to you, because everything, that works on the cvs, works here, too):
http://www.wi-bw.tfh-wildau.de/~_amueller/JOODE.zip
(It’s zip, not because I’m a windows-guru, but because more people are able to read zip, then tar.gz)
Arne
THANX THANX THANX for this file!!! ;D ;D ;D
of course joode should be implemented absolutly as pure java (becouse it is probably first such project of rigid body / constraints physics imlpementation, except odejava), but gpu support could be made as extension for users who uses some specyfic 3d librarys, anyway again thanx for that file, i think that many peolpe will be intrested in it, ok now i will check what you have made in joode…
… you have made quite much job, i am happy that you choosed float instead of doubles , becouse doubles are only faster in the latest cpu’s like pentium 4. Generaly FPU unit of any proccesors such amd / pentium can handle any format like float, double to even long double (80 bit) but transfering data using fld / fst especially for older processors is much faster using floats than double. Of course on the latest 64 bit processors this situaton is reversed. I am happy that project is on so advanced level , now i am waiting for oficial release of joode so there could be made a competition of spead betwean joode and odeJava and i think that you are maikng instead of ode also an implementation of pure java spead / memory usage benchmarck, becouse physics is computionaly very expensive, it is like TWO in ONE like that hair shampoo 
Hello,
Arne: What the problem with uploading CVS? You have access so it should be relatively straight forward.
The GPU thing sounds great but complicated. I shall not worry too much about it yet.
[quote]i think that the future of physics are not seperate cards
[/quote]
I dunno. I think it sounded absurd when 3D accelerator cards came out. I expect physics card eventually will become necissary. We are still in infancy though at the moment. I give it maybe 2 years.
Sorry guys … I broke my glasses last night so I can’t see today. I am typing 2 inches away from the screen… I will talk more when I get a new pair…
i did not said that physics acelerators will not be important in game industry, i said that not many people will buy them as seperated cards, people would like to spend additional dolars to buy good 3d graphics cards with physics capabilities, 3d vendors like ati and nvidia are competing all the time, so to achive new customers they will add support for physics. Why? becouse it will be cheapier than creating seperated product (less money spend on marketing etc) and probable much faster becouse everything will be in one card ( fast vram for graphics and physics without using agp/pci-express etc if we use seperated cards).
at ageia.com we ca find this statment : "The AGEIA PhysX processor was chosen as one of 100 products or technologies across 12 categories that symbolize a significant step forward within its category " and this is true, but if it will became a standard decide customers, of course developers, huge companies has also huge influence, but customers will decide if it will be a standard. I think like a customer, i would like to buy 3d/physics card as one product than buing again a very expensive card. Well we can argue, but as you say we will see how it will be in 1 or 2 years time,
probably asus will make first cards with ageia, as ceo of ageia said it will cost about 300$ to end user…
[quote]Arne: What the problem with uploading CVS? You have access so it should be relatively straight forward.
[/quote]
No, I have no problems uploading to cvs, but I thought, because I’ve now made a build, I could now also simply upload it to the File-section at Sourceforge (even if we’re not yet ready for a build, it would be good for people like kylix999, who want to look at it if it is at a more visible place than my webspace 
). Is this done also by cvs? I never did such a thing, so that’s why I asked - maybe also more out of interest, it was just the occasion