While I don''t have time for an in-depth explanation. The basic idea behind saving your octree is to make all structures/classes with the ability to save themselves. Thus if you ever use the structures in another area of your program, you do not have to rewrite the code to save them, all that then remains is to write a function in your class that goes through all the structures/member variables and calls their save function, passing in a file pointer (or a filename) specified upon calling the function. That''s how I go about it anyways, I''m not sure if it the best way, but it works for me. The way I worked on it was to write out the basic function, write all necessary data to file, free the class data, and test your save function by calling the load function and seeing if all goes to plan. If it does not, you can step through your code and check if any important data has been missed out in your save routines.

Some things to watch out for are ensuring your node structure is preserved when saving and reloading. Any NULL pointers in the node structure should be written to file as a delimiting character, otherwise the load routine will not pick up on the fact that a portion of the tree has been completely loaded (no character will be written to file when saving), and will continue to load nodes from other parts of the tree because it would never be told to recreate the null pointer. Also, when storing pointers to data, perhaps pointers to vertices, the memory location of the verticees will obviously change the next time the vertices are recreated, so instead you need to store the location of the vertices in your structure, or a unique ID or something similar.

Hope that helps,

Cheers,

Steve