[pmayrgundter]

Not sure about midpoint being well-defined on anything we can subtract.. Z and R are infinite.. there are a lot of values in there that don't compute.

To vary your point here, the axioms for twos complement and IEEE floating-point aren't well known or observed.

[ChadNauseam]

There are countably infinite turing machines and there is one for every element in Z. But there are uncountably infinite real numbers, so we’re out of luck for almost all of them.