Updating Variables Within Loop

[Vasia]

Hey everyone,

To preface, I have very little programming knowledge and I'm only just beginning with the very basics.
I'm supposed to be using the fibonacci sequence to estimate the golden ratio. So, I managed to create a function (which I believe works as I've tested it) and everything seems to be fine... EXCEPT I can't get the a & b variable to update at the end of my while loop. I've copied all my code below. I really can't figure out what's wrong (though I'm sure it's something super obvious) and I would really appreciate any help.

THanks
--> formatting is a little weird for comments after I copied the code her... sorry

#include <stdio.h>

#include <math.h>


int fibonacci(int N) {    //N is the term of the Fibonacci sequence starting at the 0th term


    int i = 0;            //loop index

    if (N == 0 || N == 1) {
        return 1;
    }

int previousN = 1;            //last integer in fibonacci sequence preceeding new value
int previousPreviousN = 1;    //last integer preceeding the previousN in fibonacci sequence
int newValue = 0;             //current value of fibonnaci sequence at term N

    for (i = 0; i < (N-1); i++) {
        newValue = previousN + previousPreviousN;
        previousPreviousN = previousN;
        previousN = newValue;
    }
    return newValue;
}


int main(void) {

constdouble truePhi = (1 + sqrt(5))/2; //the true value of phi used in error calculations
    int i = 0;                          // loop index
    int a = 2;                          // a = n+1 term of Fibonacci
    int b = 1;                          // b = n term of Fibonacci
    double previousApprox = 1.0;        // previous approximation of phi
    double newApprox = 0.0;             // new approximation of phi
double relError;                    // relative error between the computed approximation and the true value of phi

    while (fabs(previousApprox - newApprox) >= 0.001) {
        printf("%d  ", i);
printf("Phi approximation = %lf  ", newApprox);

        relError = fabs((newApprox-truePhi)/truePhi);
        printf("Error = %lf\n", relError);

        previousApprox = newApprox;
        newApprox = fibonacci(a)/fibonacci(b);

        a++;
        b++;
        i++;
    }
return0;
 }

[Salem]

When posting code, make sure you do 'copy as text / paste as text' where the options are available. Copying font tags from your IDE messes up.

> newApprox = fibonacci(a)/fibonacci(b);
You should make your fibonacci return double, so you don't get massive rounding errors with int/int truncation.

#include <stdio.h>
#include <math.h>


double fibonacci(int N) {    //N is the term of the Fibonacci sequence starting at the 0th term
    int i = 0;            //loop index

    if (N == 0 || N == 1) {
        return 1;
    }

    int previousN = 1;            //last integer in fibonacci sequence preceeding new value
    int previousPreviousN = 1;    //last integer preceeding the previousN in fibonacci sequence
    int newValue = 0;             //current value of fibonnaci sequence at term N

    for (i = 0; i < (N-1); i++) {
        newValue = previousN + previousPreviousN;
        previousPreviousN = previousN;
        previousN = newValue;
    }
    return newValue;
}


int main(void) {
    const double truePhi = (1 + sqrt(5))/2; //the true value of phi used in error calculations
    int i = 0;                          // loop index
    int a = 2;                          // a = n+1 term of Fibonacci
    int b = 1;                          // b = n term of Fibonacci
    double previousApprox = 1.0;        // previous approximation of phi
    double newApprox = 0.0;             // new approximation of phi
    double relError;                    // relative error between the computed approximation and the true value of phi

    while (fabs(previousApprox - newApprox) >= 0.001) {
        printf("%d  ", i);
        printf("Phi approximation = %lf  ", newApprox);

        relError = fabs((newApprox-truePhi)/truePhi);
        printf("Error = %lf\n", relError);

        previousApprox = newApprox;
        newApprox = fibonacci(a)/fibonacci(b);

        a++;
        b++;
        i++;
    }

    return 0;
 }
 


$ gcc -Wall -Wextra -g foo.c -lm
$ ./a.out 
0  Phi approximation = 0.000000  Error = 1.000000
1  Phi approximation = 2.000000  Error = 0.236068
2  Phi approximation = 1.500000  Error = 0.072949
3  Phi approximation = 1.666667  Error = 0.030057
4  Phi approximation = 1.600000  Error = 0.011146
5  Phi approximation = 1.625000  Error = 0.004305
6  Phi approximation = 1.615385  Error = 0.001637
7  Phi approximation = 1.619048  Error = 0.000626
8  Phi approximation = 1.617647  Error = 0.000239
$

[john.c]

I was going to mention Salem's point. Although I would've cast (one of) the return values instead of changing fibonacci's return type. I suggest making fibonacci use long ints to maximize it's range.

Also, you only need one of the variables i, a, and b, since it's always true that a == i + 2 and b == i + 1.