3D solid objects are not defined as points in Cad-Cam models. They are defined using 3D surfaces which fully envelope a volume. The 3D surfaces are defined by mathmatical geometric statements. A simple example is a plane surface, bounded by a polyline.
You can define 3D point data as the intersection of 3D lines and either 3D surfaces or 3D solids. In this manner you can create a grid or pattern of intersecting lines, calculate the intersecting points and then save just the point set as a seperate file.
The problem with this method, unless you have a surface like that created with a surface scanning probe, is that you need higher definiiton in certain areas to more accurately define the details of the model. A radiused or chamfered edge of a tiny hole is a good example of an area of the surface that needs more closely spaced points, as opposed to a wide flat area with sharp edges and no penetrations.
Some programs can create triangular approximations of 3D objects. Usually an interpolation value can be set so that you will get enough triangles to accurately approximate the small details. These triangles are saved as an .stl file, and with a little calculation you could convert an .stl file into a point cloud by calculating the center of each triangle. The problem with this method is that the data set becomes very large when a small interpolation value is set. There are some programs around that will analyze an .stl file and create non-uniform triangle sizes based on the curvature of the surfaces, rather than the initial geometry. This can reduce the file size, but still yields a massive( copious in iges) data set for fine detail of a large model.
Fred Smith - IMService