I think even with these operations, it's more intuitive to a programmer to just try all the bit patterns using f32::from_bits on integers especially if they aren't already deeply familiar with how the IEEE representation works, which seems like a pre-requisite for this experiment.