algorithm

[beginner at programming]

I am using jean meeus's astronomical algorithm's formula :

int = whole number.

if (month > 2) leave month and year unchanged
if (month == 1 or month == 2) then year - 1 and month + 12

a = INT (year / 100)
b = 2 - a + INT ( a / 4)

Julian day = INT (365.25 * ( year + 4716)) + INT (30.6001 * ( month + 1 )) + day + b - 1524.5


For example, October 4.81, 1957;
or y = 1957 m = 10 d = 1957

because m > 2 we leave y and m unchanged

so

a = INT (1957 / 100) = INT (19.57) = 19

b = 2 - 19 + INT (19 / 4) = 2 - 19 + 4 = -13

jd = INT (365.25 * 6637) + INT (30.6001 * 11) + 4.81 - 13 - 1524.5

jd = 2436116.31

#include <iostream>
#include <iomanip>
#include <cmath>


using std::cout; 
using std::cin;
using std::setprecision;
using std::endl;

/* GLOBAL VARIABLES */
double y = 1957;
double m = 10;
double d = 4.81;

double a;
double b;

double jd;
int main ()
{
    if (m > 2) {
        y = y;
        m = m;
    }
        
        else if (m == 1 || m == 2) {
            y--;
            m+=12;
        }
    
    
    a = floor (y / 100.0);
    b = 2 - a + floor (a / 4);
    jd = floor (365.25 * (y + 4716)) + floor (30.001 * ( m + 1)) + d + b - 1524.5;
    
    cout << setprecision (16) << jd << endl;
    cin.get();
    return 0;
} 

So why is my output:

2436110.36

it should really be:

2436116.31

as posted above

I did this right. I obtained a copy of the book that this source is in and
I double checked it more than once and I did it right.

[cartpauj]

Computers do the best they can but you're definitely suffering from some rounding errors by the computer.

[beginner at programming]

how do i fix that
BTW this algorithm is meant for computers which is weird.

Should I use something else than floor in cmath

[PanGalactic]

What is this calculating?

I would think round() is what you want.

[beginner at programming]

What is this calculating?

Jean Meeus's algorithm for converting Gregorian calander date to a Julian date


What header is round in? I have never heard of it. And I am Wanting to round down to the point to where there is no decimal. only a whole number

[PanGalactic]

Your formula is off. 30.6001 is the constant, not 30.001. And floor() is correct in this case.

[beginner at programming]

thanks for pointing that out.

I was about to use like 80 variables, split the formula into its parentheses go all modf (not sure if that would work)

[PanGalactic]

By the way, round() is part of C99, so it's in math.h (cmath) for most compilers these days even though it's not part of the C++ standard (yet).

[Duthomhas]

But C99 is not part of C++.

Here are some basic Rounding Algorithms
http://www.cplusplus.com/forum/articles/3638/