> I have a special hatred of floating point numbers. I wish and advocate for a better standard. Like what? Packing the real number line in 32 (or 64) bits is

> I have a special hatred of floating point numbers. I wish and advocate for a better standard. Like what? Packing the real number line in 32 (or 64) bits is an incredible trick. The roughly scale-invariant distribution of represented values means you can do math without regard to choice of units, which would be impossible with any fixed point scheme.