How do I draw a rotating prism in OpenGL?

I'开发者_高级运维m trying to draw a simple crystal that rotates on its axis. I can get the shape right easily enough by drawing a pyramid and then drawing it again upside down, but I've got two problems.

First off, even though I draw everything in the same color, two of the faces come out a different color as the other two.

Second, it's placing a "bottom" on each pyramid that's visible through the translucent walls of the crystal, which ruins the effect. Is there any way to get rid of it?

Here's the code I'm using to set up and draw the GL scene. There's a lot more OpenGL code than this, of course, but this is the relevant part.

procedure Initialize;
begin
  glShadeModel(GL_SMOOTH);
  glClearColor(0.0, 0.0, 0.0, 0.5);
  glClearDepth(1.0);
  glEnable(GL_DEPTH_TEST);
  glDepthFunc(GL_LEQUAL);
  glEnable(GL_BLEND);
  glBlendFunc(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA);
  glHint(GL_PERSPECTIVE_CORRECTION_HINT,GL_NICEST);
end;

procedure Draw; //gets called in a loop
begin
  glClear(GL_COLOR_BUFFER_BIT or GL_DEPTH_BUFFER_BIT);
  glLoadIdentity();
  glTranslatef(-1.5,-0.5,-6.0);
  glRotatef(rotation,0.0,1.0,0.0);
  glBegin(GL_TRIANGLE_FAN);
    glColor4f(0, 0, 1, 0.2);
    glVertex3f(0, 3.4, 0);
    glVertex3f(-1, 0, -1);
    glVertex3f(-1, 0, 1);
    glVertex3f(1, 0, 1);
    glVertex3f(1, 0, -1);
    glVertex3f(-1, 0, -1);
  glEnd;
  glBegin(GL_TRIANGLE_FAN);
    glVertex3f(0, -3.4, 0);
    glVertex3f(-1, 0, -1);
    glVertex3f(-1, 0, 1);
    glVertex3f(1, 0, 1);
    glVertex3f(1, 0, -1);
    glVertex3f(-1, 0, -1);
  glEnd;
  rotation := rotation + 0.02;
end;

Anyone know what I'm doing wrong and how to fix it?

I'm trying to draw a simple crystal

Stop. Crystals are translucent, and the moment you start drawing translucent objects, you can basically discard any notion of the effect being "simple". Rendering a true prism (which refracts different wavelengths of light differently) is something that requires raytracing of some form to get right. And there are many ray tracers that can't even get it right, since they only trace R, G and B wavelengths, whereas you need to trace many wavelengths to approximate the refraction and light splitting pattern of a prism.

The best you're going to get is on a rasterizer like OpenGL some level of fakery.

I can't explain what's going on with the faces, but the problem with seeing through to the other polygons is simple: you're not using backface culling. Unless you want to see the back faces of transparent objects, you need to make sure that backface culling is active.