transforming from screen to object coordinates

[Dave Zirl]

This may sound quite simple, but we are having problems none-the-less.
We want to pick a point on the screen(x,y) and convert them into
object(world) coordinates(X,Y,Z).  The problem seems to be in the multiple
layers of transformations that the screen coordinates must undergo
before they become object coordinates.  According to the graphics library
program guide the following transformations are involved(in either 
direction (obect <-->screen))

obect coords --> model/view matrix --> eye coords  --> projection matrix -->
clip coords  --> divide by w normalized coords --> viewport system -->
window coords --> window offset --> screen coords -->pixel

However, the problem lies in converting pixel values returned from 
MOUSEX and MOUSEY into the eye coordinate system and subsequently
going from the eye to object coordinate system.

If anyone has done any similar programming their assistance on this problem 
would be greatly appreciated.

I can be reached at dmz at rutile.rutgers.edu

Thanks in advance

David Zirl
Rutgers University
Department of Ceramics
PO Box 909
Piscataway, NJ 08855-0909

201-932-3083

dmz at rutile.rutgers.edu