Why is there a need to define the domain of the divisor in dividing functions?​