Quote:Original post by jbizzler
I'm just starting to look into the world of OpenGL. A couple of things confuse me.

1. How does the versioning work. It's not like with Direct3D where I download an SDK version. Right now, I'm just using the libraries in the Windows SDK. I know my graphics card supports OpenGL 2.0, but how do I tap into its features and such. Is it just a series of specifications, or do I need a new SDK?

2. I know the Direct3D world is getting rid of the fixed-function pipeline. Is the OpenGL community doing the same? I've noticed in all the tutorials I find, there's the PIXELFORMATDESCRIPTOR, with what looks like a lot of fixed-function settings. Is there a clean way to work around this, or should I just forget all those values? Is there any reason I should use it, and if I shouldn't, what are my guidelines for not doing so.

3. I've noticed the WGL stuff only works in WindowProc. Why is this? Is there a workaround? What if I set up an HDC and HGLRC, make them current, but never use them because my app also supports Direct3D. Would this hurt anything?

Thanks for reading my post. I hope you can answer my questions.

1) You can just use extensions.

2) AFAIK OpenGL:ES 2.0 doesn't support the old fixed function pipeline.

A little more info on point #2:

Both OpenGL|ES and "Desktop" OpenGL are moving away from the fixed-function pipeline. There are future plans to provide a "lean and mean" version of OpenGL which will basically mirror the hardware implimentation of 3D hardware, with the old fixed-function model moving into an optional software layer. Other examples of things that are to be moved include Quad and Polygon primitives, for instance.


Basically, OpenGL is moving to resemble modern hardware more closely; this is what comprises the "Lean and Mean" version. Legacy features and interfaces will become optional and be implimented in software, such as drivers, which may or may not make use of the programmable pipeline (quads/n-sided primitives could be triangulated using a shader program for instance, but its not required).


Khronos.org is the new home of OpenGL, read the last issue of their Pipeline newsletter for more information.

Quote:Original post by jbizzler
2. I know the Direct3D world is getting rid of the fixed-function pipeline. Is the OpenGL community doing the same? I've noticed in all the tutorials I find, there's the PIXELFORMATDESCRIPTOR, with what looks like a lot of fixed-function settings. Is there a clean way to work around this, or should I just forget all those values? Is there any reason I should use it, and if I shouldn't, what are my guidelines for not doing so.

As people above said, OpenGL is moving away from the fixed function pipeline. However, PIXELFORMATDESCRIPTOR has nothing to do with OpenGL, nor with the fixed function pipeline. It's simply a Win32 structure used to define the pixel format of a rendering window.

Quote:Original post by jbizzler
3. I've noticed the WGL stuff only works in WindowProc. Why is this?

It isn't. WGL functions work fine outside of a WinProc.

Quote:Original post by jbizzler
What if I set up an HDC and HGLRC, make them current, but never use them because my app also supports Direct3D. Would this hurt anything?

That would be very bad. Either use OGL, or use D3D. Do not partially initialize one, and then continue with the other. This will seriously mess things up. OGL and D3D are mutually exclusive within the same application. They cannot be used simultaneously.

1.) BAD NEWS: Unfortunately, because of the old "spec only" mentality, the current way versioning works in the API is based on the platform's implementation. For Windows, Microsoft has kept it lagging behind at OGL v1.2 for ages. And the only way to access functions of newer versions is through a slightly messy process of identifying whether the extension is supported or not by searching through a large string of available extensions on your platform. Then, manually loading all the functions associated with it via the clunky "wglGetProcAddress()". Before doing that, you'd also have to have downloaded the "glext.h" header file from opengl.org and included it in your project.

GOOD NEWS: The GLEW and GLEE libraries make this pitifully easy to work around. GLEE just loads every available extension, while GLEW lets you pick and choose the extensions you want, and/or the version of OGL you want via macros. Fortunately, the newer, more involved mentality of OpenGL promises to make the process less troublesome in future versions.

2.) The new "lean and mean" version of OGL, code-named Longs Peak, will be removing all the fixed-function capabilities and layering them on top of the core functionality.

As a clarification, the PIXELFORMATDESCRIPTOR, and the wgl* functions are part of how you configure Windows to receive and display drawing commands from OGL. It's somewhat akin to setting up the caps in D3D. Similarly, you're also setting up HDC (device context) and HGLRC (rendering context), which may look a little different, but is the same idea.

For a better idea of all this, check out NeHe's tutorials @ http://nehe.gamedev.net/ .

Some info on using GL 2.0 here, how to use it, etc.
http://www.opengl.org/wiki/index.php/Main_Page