If we're talking about N and Z (which makes sense since most programs are grounded in a mathematical model with some physical constraints added on top), then pl

If we're talking about N and Z (which makes sense since most programs are grounded in a mathematical model with some physical constraints added on top), then please acknowledge that N doesn't have subtraction defined, which illustrates why it's not very useful for most code. You could also argue that subtraction for unsigned numbers is only partially defined, just like for signed numbers. But that would be missing the point, since most numbers in practical programs are small, and for these numbers the signed range is indeed much more useful since subtraction is defined for signed integers (e.g. i32) for small numbers. This is not the case for unsigned, where 3u-4u is usually not what you want.