Hi there,
I just wrote a small blog post about my experience in working with Java for game programming. It is mostly things I stumbled upon while working on Caromble! for the last few years.
Check it out: http://www.goo.gl/GD7nm
Grts!
Peter
Blogpost about Java for Game Programming
Author
Great post. One thing that you might want to address on it though is the ease with which Java games can be modified through bytecode modification using libraries like ASM. It certainly is not a deal breaker, and at the end of the day someone who wants to work hard enough will achieve the same results with games written in compiled languages, but might be worth just putting it out there for people who don't know about it.
You first start worrying about someone hacking your bytecode and then you end up writing your games in some obscure OO framework for plain C.
EDIT: One thing, how exactly do you propose to handle FloatBuffers ? Some of them are kinda easy to reuse (ie, matrices, vectors) but the "heaviest" ones, meshes and things like that, aren't kinda what you'd easily set up for another object to use it.
Author
You first start worrying about someone hacking your bytecode and then you end up writing your games in some obscure OO framework for plain C.
EDIT: One thing, how exactly do you propose to handle FloatBuffers ? Some of them are kinda easy to reuse (ie, matrices, vectors) but the "heaviest" ones, meshes and things like that, aren't kinda what you'd easily set up for another object to use it.
Well there are always cases were you do not want the source easily accessible (things like protecting against piracy). But if you care about that, you're going to have to look for obfuscation tools (just like with almost any other language).
About the larger float buffers: most of the time you do not have to worry too much about them. Meshes are usually pretty constant during the duration of a level/map and only change during the map reload where the System.gc() usually takes care of the old ones. That is of course, if you do skinning/texture-animations etc in the shaders.
If you do edit meshes/textures/etc a lot on the CPU side of things, make sure you edit the buffers in place and only allocate a new buffer when it needs to grow (or maybe when it you need shrink it a lot).
I can also imagine that is some rare cases it could be useful to have some sort of buffer pool where you return the buffer that is the same size or the one that is the least amount bigger than the data you want to push to OpenGL.
Grts!
Peter
All right. Now more on updating data and such... I have a barebones renderer right now, barely able to draw a heightmap, some soft shading, and a few more things.
I have everything that needs to be sent to the GPU duplicated. Once in a regular float[] array and another in a FloatBuffer. So before sending it to the GPU, I update it (ie, grab float[] data, write it on the FloatBuffer) and make my draw call.
Now, this is done with a dirty flag. If you manipulate the array in some way, it gets "dirty" so the update call makes the copy to the buffer, otherwise it does nothing. Actually my math stuff is awfully coupled with OpenGL stuff since matrices and vectors contain both an array with the data and a lazily initializated buffer.
Anyway, point is, duplicate data. Do you write directly on the Float/Int/etc Buffers always or you have a "frontend" in the way of a regular array and then update it on a buffer?
I understand that Buffer objects have backing arrays already, but the direct buffers needed for OpenGL interop aren't guaranteed to have those.
Author
Well, we only use the buffers, there really is no need to have a separate array. The only downside is that you're not allowed to change the buffers during rendering (if your game is multi-threaded). If your mesh never changes, you can even forget about the buffer once you've created the VBO (but it is usually easier to have the buffer in memory as well).
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement
