Are you saying you'd like to have specific int32, int64, float32, float64 types?
It has the advantage that it is possible to give a reasonable type to most functions without a rewrite (even if the types would be terribly long to accommodate the weak underlying type system)