Good solid games using JME?


I'm new in this forum, but I've been lurking for few months now.

I'm thinking of using JME for my next project but sadly most of the projects I've seen and played are not very good regarding stability and compatibility of the engine. I mean I can run all of the engine tests that come with the source, they run great but most of the indie projects out there using JME are quirky and not very professional. I come from a C++/OGL/DX background and of course I've seen tons of good solid indie games using those languages and APIs but not much from Java/JME, excepting LWJGL as I've seen a lot of good commercial games using it.

Does anyone know where can I find a list of successful projects using JME that you can play, (I've already checked the screenshot section) ?

Thank you in advance.

There’s a bunch…the majority of commercial games using jME are still in development, but have you seen Bang Howdy?

I just played the game, it looks nice but it ran somewhat slow, judging by the graphics I think the game doesn't use any heavy pixel shader stuff so I think it should run much faster than that. Maybe it's because I was playing it directly from the web ? does the download version is faster than the one embedded on the site ?

Also I'm wondering now if the are any demos doing stress tests with animated characters (skinning) ? I'm working on a simple game but I need several of my chars on screen at once, about 20 to 30, I was wondering if jME could handle that kind of load ? I'm planning to use the collada format, but I know jME internally works with a custom format.

I have seen and coded a few demos displaying several characters in C++ with other engines and I could achieve some pretty good framerates with around 30 characters on screen (just a simple plane with characters on top)

Opinions ?

1 Like

fwiw, at work we've done skinning tests and had interactive frame rates with up to 70 or so skinned models (Collada) in a complex scene.  As for Bang!, they were the early adopters with the technology, and while they have done a great job with it, the technology has, and will, continue to improve beyond what they had access to.

1 Like

by "work" I assume you are referring to a heavily modified version of jME used for commercial purposes ? if not then it would be great to have some of those test on the CVS sources.

Another little question, are there any good java bindings for some of the open source physics engines out there ? I saw jME physics but the website has not been updated in a while and I haven't seen any good demos, I played some of the demo games in the forums using JME physics but it seems it is not stable at all. It would be great to have more options for physics engines, one that comes to my mind is the newton physics engine, I've used with C++ and it's way more stable than ODE.

1 Like

Renanse and Mojomonkey work for ncSoft and the majority of what they work on is directly from jME.  They also get to contribute back the majority of their work (apart from proprietary features which is understandable).  I don't want to speak out of turn though, Renanse can explain better, but I just wanted to say that the jME community gets many benefits from them, and it's not a "watered down" version of jME.

jME-Physics is designed to support additional implementations so if you would be interested in a Newton implementation all you have to do is connect it. ;)  I've personally not been too thrilled with ODE, but it works…most of the time. :)  There was also an endeavor a while back to create a PhysX implementation but I don't know if anything ever came of it.

I think if you're trying to make a decision if jME is stable enough to use in game development I would say the fact that large game companies are developing games with it, JavaOne used it for their Flying Dukes contest, and it definitely has more community support than any other Java 3D engine.  If, however, you're trying to decide if Java is more powerful than C++ for game development it's not as easy of a pick. Java is one sense always going to be at least slightly slower than the most optimized C++ code by nature. However, since we rarely see C++ code that is highly optimized there is lots of room for Java to be faster. :wink:

1 Like

dude, it's fixed already! get that cvs thing working and subscribe to the cvs mailing list! :stuck_out_tongue:

1 Like
datx said:

by "work" I assume you are referring to a heavily modified version of jME used for commercial purposes ? if not then it would be great to have some of those test on the CVS sources.

By work, I mean using jME, straight-up, for commercial purposes.  We do toolsets and addons (like special UI, terrain stuff, etc.) that stays proprietary, but anything that is updated/fixed/added to the core packages goes back to CVS.

Curious though, it sounds like you are big into C++ development... you mention your experiences there quite a bit.  Why not do your project in C++?  Personally I'm an advocate of doing things with the best tools that you are the most comfortable using.  For me that is jME, but of course for many others outside the Java community it would be something else.  Java game development is still breaking out of its shell in many ways, so if you do use jME there will be some rough edges we are still working on.  That said, I anticipate development work over the next 10-12 months to shave off many of those rough edges.
1 Like


About physics engines, of all the games using open source engines only the ones using Newton seem to have the stability required for going commercial. I played with PhysX and it's amazing but the reason is that isn't open source and you have to pay quite a bit for a commercial license, so my list would be:





Though last thing I heard was that the people from AGEIA are now allowing anyone to use PhysX for commercial purposes without paying any license which is great because PhysX is rock solid, but I'm not sure if using PhysX through java would break any of the eula stuff.

In my own experience ODE is unstable, and last time I checked it didn't support convex hulls, and uses OPCODE for collisions which might be good or bad if you know the lib and are willing to modify it for your own purposes (interpenetrations).

Bullet is one of the new engines that looks promising, it's quite stable and I think they have integrated the engine with blender so now blender supports physics through Bullet.

deus_ex_machina, renanse, I'm not really comparing Java with C++, but I wanted to know if this engine was capable of being used for something commercial. I've gone through a lot of engines, Ogre is great but is just a graphics engine so you need some time to integrate it with input+network+physics libs,  you basically have to make your own efficient framework with ogre to make a full game, and one of the problems with Ogre is that it has a LOT of dependencies by itself, meaning you end up with a lot of files and DLLs taking quite a bit of space for a small game (currently it doesn't support collada).

Irrlicht is great in that it'is very easy to work with but some of its parts are way out-dated like the animation system, without mentioning that some of the features listed on the site are broken, like loaders (collada is one of them), etc. So it's not really usable unless you are willing to put a good amount of work to "finish" the engine.

Crystal space, last time I checked was a total mess, not well designed at all which translates to a steep learning curve.

Riot in Emerald city looks like a cool game and I'll check it out when you have something playable :slight_smile:

But I'm reviewing jME because I want to use it mostly for small 3d games, sort of "casual but not quite", and I've heard that coding in Java is faster than C/C++ .I'm not sure about that because I've been coding in Java for years and it feels just like C++ with heavy OO to me :slight_smile:

It's incredible that of all the engines out there you would think that there's a good solution that you can use without messing with the actual engine source, but no, for 2d games it's different there are some very good libs that you can use to make a full game without messing with the source, I guess I'll have to pick one and bite the bullet if that means spending weeks integrating several libs together or modifying the engine's source. I'm checking jME first though :slight_smile:

1 Like

Oh, do not take my rumbling about language speeds personal, I'm just playing smart-ass :smiley: And I looooooove building languages and compilers, so that's an area I actually know something about, au-contraire to JME :wink:

About changing source, there's an easier policy: you just come here and drink everyone's blood with the straw until they fix the thing you want. Notice how everyone is touchy about zOrder? That's my achievement :smiley:

Ok, enough kidding: JME is really, really good. "Riot" is in development for about 8 or 9 months now, and one single problem we had was zOrdering. Of course, there are some features that we would like to work our way (for example, InputHandlers), but if you bend slightly to the engine's rules, you get accustomed to it quite well and quite fast.

I can only warmly recommend it. (I expect to be paid by developers for this drama-talk)


I was just kidding about zOrder. Get used to it as I very rarely speak seriously. :frowning:

I think zOrder is going to become a rude word here :smiley: "Hi, I'm noob, can someone explain to me how to use z-****r?"

1 Like

datx, an area we can definitely use more help in is physics implementations.  The problem we have is very few of us have the C++ knowledge (and those that do don't have the time) to make the connection between something like Newton of PhysX and Java.  If that is something you'd be interested / willing to do you'd be a welcome asset to this community. :wink:

…not to say you wouldn't be welcome anyway…but for the purposes of this conversation and trying to convince you to help in that specific aspect of development I'll go ahead and make that assertion.  :stuck_out_tongue:

1 Like

I'm surprised no one has mentioned Rover Run. I enjoyed it.

1 Like

That is a good one…the graphics are impressive, the physics work well, and played very smooth on my machine. :slight_smile:

1 Like

though not released, lots of people have played our HockeyChallenge beta with great feedback(shown and played at gdc too). our other demo Spirits was very stable and smooth as well, with pretty high-end features…

1 Like