I can see a case for decimal floating point types, but fixed point adds nothing beyond what can be performed directly via arithmetic on the underlying integers. It's basically a convenience feature.