error: cannot convert ‘float (*)(int)’ to ‘float’ - c++

Question

My program converts temperature from the Fahrenheit scale to the Celcius scale and finally to absolute value scale.
#include <iostream>
#include <cmath>
#include <iomanip>
using namespace std;
int farh;
float cels(int a)
{
float c;
const int m0 = 32;
const float m1 = 0.5555;
c=(a-m0)/m1;
return c;
}
float ab(float a)
{
const float m2 = 273.15;
float d;
d=a-m2;
return d;
}
int main() {
const int WIDTH = 16;
cout << setiosflags ( ios :: left );
cout << setw(WIDTH) << "Fahrenheit" << setw(WIDTH) << "Celcius" << setw(WIDTH) << "Absolute Value" << '\n';
cout.setf(ios::fixed);
cout.precision(2);
for (farh = 0 ; farh <= 300 ; farh = farh + 20) {
cout.width(16);
cout << farh << cels(farh) << ab(cels) << "\n";
}
return 0;
}
The Compile time error message I receive is:
d26.cc: In function ‘int main()’:
d26.cc:38:40: error: cannot convert ‘float (*)(int)’ to ‘float’ for argument ‘1’ to ‘float ab(float)’
cout << farh << cels(farh) << ab(cels) << "\n";
Why am I receiving this error?

Answer 1

ab takes a float and returns a float:
float ab(float a)
But cels isn't a float, it's a function:
float cels(int a)
You probably meant
ab(cels(farh))
Or to take a temporary:
float cur_cels = cels(farh);
cout << farh << cur_cels << ab(cur_cels) << "\n";
Side-note, ab should probably be named kelvin.

Answer 2

Actually you have passed a funtion pointer to ab. if your intent is to pass the function ( clearly not! ) you can use the following syntax :
float ab(float(*callback)(int),int pass) {
callback(pass); /* It calls the function indirectly with <pass> */
}
Its great for menu creation for example if you have two options:
1. Fahrenheit to Cell
2. Fahrenheit to Kelvin it will be useful
You can search callback functions in C with google.