Ward, Greg2015-11-122015-11-1220031017-4656https://doi.org/10.2312/egt.20031099This tutorial develops the thesis that the real world is best represented by real numbers, which are approximated by floating point values in the computer. As the floating point unit (FPU) continues to accelerate, outpacing the arithmetic logic unit (ALU), it makes more and more sense to do our graphics calculations entirely with real numbers. We even see the mainstream changing in this direction with the introduction of floating point frame buffers in a new generation of graphics cards, such as NVidia’s GeForce FX.