c program for binary to decimal conversion

Question

Please help me to sort out the error in the following program.I am not getting the correct output for the decimal part. Thanks in advance

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

main()
{
    int p,i,j,t,x;
    float n,sum,d;

    printf("enter the binary number");
    scanf("%f",&n);

    p=(int)n;
    d=n-p;

    i=0;
    while(p>0)
    {
        t=p%10;
        sum=sum+t*pow(2,i);
        p=p/10;
        i=i+1;
    }

    i=1;
    while(d>0.0)
    {
        d=d*10;
        x=(int)d;
        sum=sum+x*pow(2,-i);
        d=d-x;
        i=i+1;
    }

    printf("decimal value is %f",sum);
    getch();
}

Answer 1

You're problem is that you're using float when you should be using int and you haven't initialized any of your variables. Change float to int and initialize your variables to 0 and your code will work.

The reason your code isn't working is because sum isn't initialized. When you use sum = sum + t * pow (2, i), you are adding an arbitrary number to the result in the first go around. sum could be anything before it's initialized (ie -1241232).

Even though the only variables that really need to be initialized are sum, d, and i, it's a good practice to initialize them anyway (better safe than sorry). In fact, initialized variables are a very common cause of errors. Initialize everything.

The floating point datatype is also causing problems in you code, in the second loop. When the user inputs a floating point number, like 110.1101, it gets approximated to another number (ie 110.110098) - this has to do with the limitation of the floating point datatype. In your second loop, with the input approximated to 110.1100998, the problem is caused in this statement:

sum = sum + x * pow (2.0f, -i);

Because x isn't 1 or 0 throughout the whole loop (sometimes 9 or 8, etc). What's more, the loop doesn't just iterate 4 times (1101), it does many more. This is because of this statement:

d = d * 10;

When d gets multiplied by 10, it goes from 0.100998 to about 1.00979 - not 1.00998 like you'd expect. These two discrepancies are what's causing your results to be erroneous when you input values more precise than a single digit.

Finally, here's some working code:

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

int main (){

    int p = 0, i = 0, j = 0, t = 0, x = 0;
    float n = 0.0f, sum = 0.0f, d = 0.0f;

    printf("enter the binary number: ");
    scanf("%f", &n);

    p = n;
    d = n - p;

    while (p > 0){ // never use the float datatype for iteration
        t = p % 10;
        sum = sum + t * pow (2.0f,i);
        p = p / 10;
        i = i + 1;
    }

    i = 1;

    while (d > 0.00001){
        d = d * 10;
        x = d;
        sum = sum + x * pow (2.0f, -i);
        d = d - x;
        i = i + 1;
    }

    printf ("decimal value is %f\n", sum);
    getchar ();
    return 0;
}

Answer 2

I am assuming you are expecting input of the form [01.]* with the constraint that there is only one decimal point. You are reading this input like a regular floating point number, but interpreting its decimal digits as binary digits and computing the decimal value.

main() should be declared int main (). getch() isn't a standard C call. You can use getchar().

You are probably getting crazy output because sum is not initialized to 0.

When you compute your fractional terms, you have to keep in mind that floating point representation is binary, and not all decimal numbers have an exact representation. So you need to perform rounding. Also, you are checking for d>0 as your stopping condition for your second loop, but floating point comparison is trickier than that. You can read this answer for a complete explanation as to why it is tricky, but you should basically do something like this:

    float err=.000001;
    while(d>err)
    {
        d=d*10;
        err=err*10;
        x=(int)(d+.5);
        /*...*/

The floating point complexity of your code can be reduced by reading your input as a string, and parsing that instead.

    char *n;
    char p[INPUT_MAX];
    char d[INPUT_MAX];
    char input[INPUT_MAX];
    float sum, p_sum = 0, d_sum = 0;
    int i;
    printf("enter the binary number: ");
    if (fgets(input, sizeof(input), stdin) == 0) {
        puts("no input!");
        exit(0);
    }
    n = strchr(input, '.');
    if (sscanf(input, "%[01]", p) == 1) {
        for (i = 0; p[i] != '\0'; ++i) {
            p_sum = 2*p_sum + (p[i] - '0');
        }
    }
    if (sscanf(n, ".%[01]", d) == 1) {
        for (i = strlen(d); i > 0; --i) {
            d_sum = (d_sum + (d[i-1] - '0'))/2;
        }
    }
    sum = p_sum + d_sum;
    printf("decimal value is %f\n",sum);