I think I just sent it to you. I also put it up on the program archives. Prisms is the actual engine, while Cubes is a program that called the engine as a subprogram; all Prisms does is draw the prism. In cubes, you define size of the prism to be drawn and you get to move it around in a virtual virtual 3d world. (There isn't an actual virtual world, it only gives the appearance of one) H= height, W=width, D=depth, and S=solid(0 for a transparent prism, 1 for a solid one.)
If you so desired to use the engine yourself, here are four very important tidbits to remember.
1. Include this at the begging of the program:
94→Xmax
0→Ymin
62→Ymax
AxesOff
LabelOff
ExprOff
GridOff
FnOff
Degree
Float
2. "send" the engine L₂ in this format:
{x coordinate, y coordinate, width, height, depth, solid}
3. Store the X and Y coordinates of the vanishing point into V and W, respectively.
4. Remember that the engine currently uses the variables L₁, L₂ , S, T, N, O, I, J, V, W, and Theta
PS I know that the prisms begin to bow out if the depth is more than about 20. This is because it doesn't use true one-point perspective. It uses a little short-cut to make it faster; kind of like drawing two squares and connecting them to draw an ice cube, as opposed to taking every corner of it out to the vanishing point.
Also, do you think that I should post my lights-out game? It is text based, but uses that short assembly program for inverted text for highlighting. I'm not sure if it's worth posting.