What is the advantages of call by reference?

Whenever you pass a value to a function, that value must be copied. If the value is large or complex, this can hinder performance, particularly when the function does not need to alter the value. To improve performance, we use the pass by reference semantic. Rather than passing the value itself, we pass the address of the value. That is, the address is copied, not the value. The function can then refer to the value by dereferencing the address.

Ideally, pass by reference should only be used when the function does not need to alter the value. This is achieved by declaring the function's formal argument a pointer to constant type. This makes it clear to the caller the value of the actual argument will not be altered by the function.

In some cases, we do require the function to alter the value. In these cases the argument is regarded as being an output parameter because it allows the function to return another value besides the return value. Typically, the caller will allocate some memory for the function to use (perhaps initialising it with a value), and then pass the address of that memory to the function. The function's formal argument is therefore declared a pointer to non-constant type, making it clear to the caller that the function will modify the value being pointed.