This problem has probably already been answered before, but I''ve been having trouble with the search page for this forum. The basic problem is this: I have a point and click interface where I need to be able to do 2 things. Select objects from 3D space using an orthographic cursor location, and to be able to select points on the level data using similar restraints (walk-to commands). I''m aware that gluUnproject will give me a solution of sorts, but that''s not an option for me for the following reasons: I don''t use a depth buffer for the level data, as I store all that information in a BSP tree to avoid lagging my program with a z buffer, plus I have surfaces that use blending for transparency and reflection. gluUnproject doesn''t give the correct 3D space coords. X and Y coords are fine, but it tends to have a signifigant error factor for the Z coord, which gets bigger the farther along your viewing fustrum you are. So I''ve decided to use a Ray-Poly intersection method but I''ve run into a problem. How do I take a 2D orthographic screen coord and divine the correct ray from it? Another consideration is that the scene has camera rotations and transformations to further complicate things. Also my Ray-Poly algorithm takes a Ray-Plane intersection and checks to see if the point of intersection is within the bounds of the poly (always a convex triangle). Is there a better way? Thanks ----------------------------- Mice are an excellent source of mice.

It seems like I always end up posting answers to my own question 5 minutes later

Anyhow here's a solution I found from another forum (over at openGL.org) that I found easy to understand and straightforward. I don't know if it can be optimized or not, but it's a start:

http://www.opengl.org/discussion_boards/ubb/Forum3/HTML/000601.html

-----------------------------
Mice are an excellent source of mice.

Edited by - Jo on October 7, 2000 4:40:36 PM

check out MrGamemaker for an excellent tutorial!

Edited by - Raptor on October 8, 2000 2:14:43 PM