How to convert a negative value to positive

[Angus Comber]

Hello

Sorry this really is a newbie question!

I am doing some floating point arithmetic and calculating the time
difference between two dates. The date values being comparied are actually
double values. I was going to compare the dates by calculating the
difference betwen the two numbers. But sometimes you get a -ve and
sometimes a +ve value. I want to disregard a small variance eg of up to two
minutes but what is best way to compare values.

ie calculate difference and then convert any -ve values to +ve then check if
(diff > x)

or some other way?

AC

[Gordon Burditt]

>Sorry this really is a newbie question!

I am doing some floating point arithmetic and calculating the time
difference between two dates. The date values being comparied are actually
double values. I was going to compare the dates by calculating the
difference betwen the two numbers. But sometimes you get a -ve and
sometimes a +ve value.
The unary minus operator will convert a negative value to a positive one.
I want to disregard a small variance eg of up to two
minutes but what is best way to compare values.

ie calculate difference and then convert any -ve values to +ve then check if
(diff > x)

You may want to use fabs(), in code similar to this:
#include <math.h>
double x1;
double x2;
double delta;

... put values in x1, x2, and delta ...;
if (fabs(x1-x2) > delta) { ... too far apart ... ; }
else { ... pretty close ... ; }

Gordon L. Burditt

[glen herrmannsfeldt]

Angus Comber wrote:

I am doing some floating point arithmetic and calculating the time
difference between two dates. The date values being comparied are actually
double values. I was going to compare the dates by calculating the
difference betwen the two numbers. But sometimes you get a -ve and
sometimes a +ve value. I want to disregard a small variance eg of up to two
minutes but what is best way to compare values.

Personally, storing dates as floating point seems
very strange to me. Yes, time is continuous(*), but not the
way it is usually measured, and not the way you want it
measured.

(*) I believe it is still unknown if quantum mechanics quantizes
time or not.

-- glen