Tom Honermann:
> What is the motivation? What problem(s) would this solve?

A code that doesn't depend on a particular size of a byte will be easier
to write and easier to reason about.

C++20 requires integers to be two's complement because there is no
relevant hardware with other representations.

Why not to restrict values of CHAR_BIT to values that are relevant too?

I'm working on a proposal for sane integer types that introduces new
context-sensitive keywords in the form intN. If we restrict CHAR_BIT
now, then the wording and reasoning for new types becomes much easier.

https://github.com/Lyberta/cpp-sane-integers