[alberto-m]

> one could make the case for true being all bits set

Historical note: this was the case in QBasic, where true was defined as -1.

[Joker_vD]

There, apparently, were quite a number of ISAs where checking the sign bit was more convenient (performant?) than checking (in)equality with zero.

[jabl]

Some Fortran compilers also did this. MS Powerstation Fortran at least, IIRC.