It depends on how you intend to use it. For example, in digital signal processing (audio), most processing is done using uint32 values. It would make no sense t

It depends on how you intend to use it. For example, in digital signal processing (audio), most processing is done using uint32 values. It would make no sense to receive the value as a float, only to convert it back to int before using it. Case in point: most software volume controls allow to specify the output volume (or rather, control the attenuation) using the full uint16 range. In this case, the entire volume range is expressed from 0x0000 to 0xFFFF. But in practice, this is just an amplification (multiplier) value between 0 and 1.