What software methodology and group mechanics decisions worked out well, and which ones didn ?
Why? Coming up with a class level architecture UML diagram proved helpful in understanding the basic relations between the classes and what was inherited from what. Keeping the game object-oriented and creating new files for each class made the files manageable, but possibly confused other group members as to how all parts were interacting as classes were being added often. Our group also encountered difficulties with OpenGL and creating dynamic models after the networking was fully loaded. Because of this difficulty we had to redo some code to preload all of the game elements and create texture objects/display lists/vertex buffers up front that could be shared globally. Global data can sometimes be considered bad OO but in the case of a game you will probably be using globals or singletons quite a bit. Deciding to have one person on graphics, one on user interfacing (both input and output), and one on integration was definitely helpful and needed. We ended up having one person on full
