considerations before beginning a project with irrlicht

[torchLight]

Are there any limitations to the input set up (keys and mice) with irrlicht.

Also does using irrlicht cause poorer performance compared to going straight to opengl and directX. I really dont want to as the sdk's for that are much harsher than irrlicht but if it will mean better long run potentail then I need to at least wrap my head around it. Both openGL and directx have sdks but they look geared toward large teams not indie.

Really what I want to know is will irrlicht satisfy the needs of an indie student or will I need to work with opengl and directx in the long run. My goal is an rts with centered around economy and city building rather than make a base and blow crap up.

Does irrlicht work directly with drivers or does it have another sdk like nvidia opengl sdk behind it. I mean is it standing on its own or is it a wrapper for opengl, directx, ect...

Am I limited by irrlicht as an indie in the long run or will it boost productivity and leave performance alone (or make it better)?

Does combining sdks and libraries slow performance or just compile time? Can I have a hundred different middlewares and be good or will that cause disfunction, that assumes that I do the impossible and keep track of a hundred middlewares and mind that its hypothetical. Basicly do I need to limit myself to one or two or can I take a single feature from some middle ware and not be punished performance wise for it.

Can I use nvidia opengl sdk and irrlicht if say they have a feature for hdr that irrlicht lacks (not saying that, just hypothetically). For instance take the model viewer demo and put in the nvidia opengl hdr. Will two renderer sdks conflict too much and leave me a monster project just to put one nvidia feature into an app, will it mean writing nvidia into irrlicht or just coding in the lines and calls to nvidia for that feature and nothing else. Like the impressive glow from nvida to a game using irrlicht as the main engine.

I think thats all thats got me confused. I think they are some important things to look at though and I'm hopping some ppl already know whats what saving me trial an error.

[radiant]

Well.... you would certanly have a faster application if you run directly over openGL or DX, but.... of course you would also have a faster app if you program in asambly instead of c++... Irrlicht its a graphic engine and as such... it eases the development of graphic apps... 3D and 2D.... at what cost..u ask... small enough to prefer it over dicrect GPU manipulataion...

Irrlicht can use DX8, Dx9 and OpenGL and 2 software drivers....

So.... unless in your project procesor cycles are so important that u cant waste a single one(wich u are already doing since you are using non-computer lenguage to program -> binary) you are fine with irrlicht cause its really easy to learn and get an app up and runnig(depends on the app of course)

Hope it helps

[CuteAlien]

You can still work with DirectX or OpenGL if you have to and as Irrlicht is OpenSource you can even modify it to your liking if you need that. But when you work with an engine you work at a higher level. With OpenGL/DirectX you do basically work with the lowest 3D elements. Like polygons, lights, bitmaps. With a 3D Engine like Irrlicht you work more with stuff like models, meshes, images. You can still access the low-level stuff, but in most cases you won't do that.

Irrlicht is set on top of other libraries. So it will use OpenGL or DirectX (or both). It can also render with Software, but that will be slower and has less features.

Combining SDK's does usually not slow you down. It could do so when you reach the memory limits of your computer, but that's usually never caused by libraries but by data (like textures). You should not worry about that.

Different libraries can not necessarly easily be combined. So if you have one library offering HDR and another does the rendering then they will probably not work together. But due to opensource you can always extend irrlicht and sometimes it's possible to write interfaces to combine different libraries. As Irrlicht can for example use OpenGL you can add all OpenGL features to it as long as you use the OpenGL renderer . The opengl features certainly won't work for DirectX, so that would be an example of libraries where you can't take the features of one library and put it in the other.

Doing an RTS is certainly no trivial job. You need to have good programming experience to finish such a project. Though the first simcity (was is realtime? Well - it had city's... I remember that ) was done by one programmer - so it isn't impossible.