Up until today I’d been planning to add some support for real numbers, or at least scalable floating-point-like numbers that approximate real numbers. That’s no longer on the to-do list.

Real numbers aren’t actually representable with the discrete mathematics available within a digital system. The numbers that are representable are all equally representable as rational numbers. In terms of what can actually be represented, rationals are, in fact, a bigger set.

What this does mean is that rationals now gain much more significance than I’d originally expected. It’s possible that I may want to build a more efficient form of scaled natural number instead. It also suggests that there should be functions to create rationals from floating point types, or to generate fixed-size floating point values from rationals. Similarly, though, there need to be functions to generate fixed-size integer values from the natural and integer number types.