Subroutines

Question 1

Can you rewrite the following subroutine as a function?

subroutine sub(y, x, a, b, c)
    intrinsic none
    real :: x, y, a, b, c

    y = a*x**2 + b*x + c
end subroutine

No, there are too many arguments. [No, the number of arguments as such is not really relevant.]
Yes, the function signature would be real function sub(x, a, b, c). [Indeed, only the argument y is modified.] [x]
Yes, the signature would be real function sub(y, x, a, b, c) [Although technically okay, it would be poor style since the intent is to return the value of y as the result of the function.] [x]
Yes, the function signature would be function sub(x, a, b, c) result(y). [Yes, this is a good signature for the function.] [x]

Can you rewrite the following subroutine as a function?

subroutine sub(n)
    intrinsic none
    integer :: n, i

    do i = 1, n
        print *, n**2
    end do
end subroutine sub

Yes, the signature would be logical function sub(n) and it would always return .true.. [Indeed, it is not really elegant, but it would be correct.] [x]
Yes, the signature of the function would be void function sub(n). [No, there is no void type in Fortran.]
Yes, the signature would be integer function sub(n) and it would always return 1. [Indeed, it is not really elegant, but it would be correct.] [x]
Yes, the signature of the function would be void function sub(n). [No, there is no void type in Fortran.]

Consider the following subroutine. When you call it with arguments 3 and 5, what is the result?

subroutine sub(a, b)
    implicit none
    integer :: a, b
    a = 3*b
end subroutine sub

The result will be that a has the value 15. [No, there is no variable a passed as an argument to the subroutine.]
You will get a runtime error. [Indeed, this will typically generate a segmentation fault.] [x]
You will get a compilation error. [No, this compiles just fine.]
The result will be that a has the value 3. [No, there is no variable a passed to the subroutine.]