Ok. I've hacked together a little GL program (attached), my goal is to be able to "fly" around the "world" I've created, freely. Right now I'm stuck moving in tracks. I have only a basic understanding of the Matrix math included ("borrowed" it from a program of George Foot's).

I "think" I can get the movement I want by including the "vel" vector into the matrix equations somehow, but I'm not exactly sure how.

to build the program,
in unix: ./cbuild.c
in mingw: gcc cbuild.c -o cbuild.exe && cbuild.exe

It requires the latest AllegroGL CVS, and Allegro 4.2+.

edit, hmm, I realize some here may be a little lazy, so I'll post some of the relavant code:
Camera Class:

Keyboard/Movement:

Now, it could be just me, but theres got to be a way to include the "pos+vel" into the matrix calculations in Camera::setup.

edit2, 5+ hours and nothing? Thanks guys.

Is this camera rotation or camera movement?

[edit]
Perhaps you could use sin/cos when you rotate the camera.

I have to go out now but perhaps drawing what you want to happen then connecting the patterns would help you form an idea of what you need to do.
Though do the example using only 2 dimensions first.

^ Disregard that if it doesn't help you at all.

Thanks Archon. Its both really.

I want to properly move through a 3d environment with all 360 degrees of freedom.

I've attempted to modify my camera update to the following, but it seems I've got something totally wrong about the gl matrices:

What exactly am I doing wrong? I want the camera to be able to fly around as if it were, say a space ship, or a plane, in and arround the cubes and whatnot.

edit, KittyCat has given me one fix:

  GLfloat mat2[16] = {
    1, 0, 0, 0,
    0, 1, 0, 0,
    0, 0, 1, 0,
    pos.x, pos.y, pos.z, 1
  };

It helps a little, but now the arrow keys cause the camera's position to slowly rotate around its pos.

edit2, I've commended the pos.Normalize(); and its going better.

edit3, Now, the camera doesn't move in the direction of the direction vectors...

I usually store both rotation and position together in a struct (a class actually), and each as a vector, where the rotation vector indicates axis (the vector's direction) and angle (its length). This "position" struct can be used both for objects and for cameras. To activate it as a drawing offset (as an object position), do a glRotate() first, then a glTranslate(). To use as a camera, you do the translate first and with opposite signs, then the rotate, also with opposite signs.
To make the camera move, you'll need the front vector, which is a bit trickier using this approach. You can use the matrix sample code that comes with the gl documentation, or rip code from allegro, to convert the rotation angle into a matrix. Then all you need to do is use the appropriate row (or column? can't seem to remember) as front vector.
You won't need to know a lot more about matrix math than this, but it helps to think of a matrix as a local coordinate system with 4 vectors (x, y, z, w), that describe 3 axes and an origin (w). Multiplying a vector by a matrix means taking a local vector and calculating its position in absolute (real-world) space. Matrix multiplication means nesting several of these coordinate systems, and the resulting matrix can be used to transform directly from the innermost coordinate system to world-space.