Being stuck with walking around the supermarket (baah, humbug!), I tried to devote my trains of thought to Stargella. For me, it is kind of a difficult project; much more complex then most of the things I’ve worked on.
In thinking about how to structure the rendering code, I started to think more on the issue of dealing with the data sets to be rendered, which basically means sprites (2d), models (3d), and rendering the level (2d/3d), and I want as much of the code to be shared between several game projects, which makes baking the lasagna a bit involved. Since my current focus is on the 2D oriented, Stargella, I find it wiser to concentrate on sprites then three dimensional models. At first I was thinking, it would be ideal to represent animations for sprites, as a simple image stack; but I don’t want to go through all that for unanimated sprites, which would still require a singular stack frame. The fix to that is simple, only use stacks for holding the animation details! Then I was thinking, gee, wouldn’t it be cool, if we could attach some type of script or shader program to be run, so a level of progmmability could be used in implementing special effects. For example, using program to stake out the centre of a starship, split it into segments and have them drift away from one another, while interleaving it with an explosion-like image or particle effect. Simple enough to do from C I’m sure, but I would much rather have it be apart of the game data, then the game code. I’m not sure if that would be doable from mating with GLSL files, or if it would mean binding a script with the sprite. Either way, as far as general game scripting goes, it should have some level of access to the graphic stuff as a matter of completeness.
For now, I think I’ll concentrate on resource handling. It should be easy enough, just create dynamic arrays as needed (e.g. of loaded textures), and each time a new texture is loaded, assign it an unique identifier that indexes into the array; …
Honestly, I don’t have a problem with programming in C, other then it can be time consuming at times o/. But, because I have to learn OpenGL along the way, it is, shall we say a freaking lot of docs to parse.
While I’m more familiar with Direct3D then OpenGL, even if you can still get a good bang for you buck out of a DirectX SDK without having to stoop to C++, I’m not willing to write a DirectX backend for my game. Obviously, I want to avoid using glThis and glThat all over the game code, but ahem, the issue of shaders enter the picture. Short of making a definitive choice at compile time, that would have to live for the projects entitiy, the only way to support runtime choice, would also require writing both HLSL (DirectX) and GLSL (OpenGL) code, and that’s a headache I won’t put up with! As much as I favour the idea of nVidia’s Cg, it is not available on enough platforms, and implementing it where needed, to much trouble.
If people other then me want to play the game, they can buy a graphics card made in the 21st century and install its drivers >_>.
—- this post has been interrupted so many damn rat fucking times by family, that it is unfinished and will likely always remain so. This is one of many reasons why I fucking hate my life.