In MYTHOS I have opted to expose only the underlying Allegro transforms and create ways to manage them better rather than create a proper camera system built-in like what is necessary in most 3D engines. I do not want to complicate any perspective tricks that might rely on transforms, since changes to the transform the GPU is working with are relative to the previous transform. Without some two-step BS to control the order camera transforms are applied to each entity it would quickly become a headache for a small few developers.