Divinding floating points by zero

[Leandro Melo]

Hi!
Little doubt.
Kathy's book says that when divided by zero, a floating point may become infinity or -infinity depending on the sign of the zero.
I checked it out and verified that, at least with sun's jvm, this result is actually due to the sign of the number beeing divided, the sign of the zero doesn't really matter.
Any comments?

[shandilya popuru]

hi leandro

thats right the sign of the numerator is taken into consideration rather than the sign of zero

[ankur rathi]

yes , you are right ... I have also tried ...

[Barry Gaunt]

The following example shows that the sign of the 0.0 certainly has an effect on the sign of the infinite quotent:

public class DividingByZero
{
public DividingByZero()
{
}
 
public static void main(String[] args)
{
    System.out.println(1.0/(-0.0));
    System.out.println(1.0/(0.0));
    System.out.println(-1.0/(-0.0));
    System.out.println(-1.0/(0.0));
 
}
}
  

Output:

-Infinity
Infinity
Infinity
-Infinity
  

[Barry Gaunt]

This code:

public class MinusZero
{
 

public static void main(String[] args)
{    
    System.out.println(-0);
    System.out.println((float)-0);
 
}
}  

and its output:

0
0.0
  

might give you an idea what's happening in your tests.

[ankur rathi]

public class DividingByZero
{
public DividingByZero()
{
}
 
public static void main(String[] args)
{
    System.out.println(1.0/(-0));
    System.out.println(1.0/0);
    System.out.println(-1.0/(-0));
    System.out.println(-1.0/0);
 
}
}

output :
Infinite
Infinite
-Infinite
-Infinite

[ankur rathi]

Are you mean :
there is no difference b/w 0 & -0 .

[ankur rathi]

I have checked :

1] there is no difference b/w 0 & -0 .
2] but this two 0.0 & -0.0 are different .

[Mark Spritzler]

The actually is a difference between -0.0 and 0.0, try sorting it and see what order they come out in.

Mark

[Mark Spritzler]

"Sandy"-
Welcome to the JavaRanch! Please adjust your displayed name to meet the

JavaRanch Naming Policy.

You can change it

here.

You used to be in compliance. Why did you change your name?

Thanks! and welcome to the JavaRanch!

Mark
[ January 14, 2005: Message edited by: Mark Spritzler ]

[Leandro Melo]

Allright, you pointed the singalling diferences between 0.0 and 0 and -0.0 and 0. But, what is the reason behind it?

[Jeroen Wenting]

Originally posted by rathi ji:
Are you mean :

I think he's actually quite nice

[Tony Morris]

Allright, you pointed the singalling diferences between 0.0 and 0 and -0.0 and 0. But, what is the reason behind it?

The definitive source that you appear to searching for is called "IEEE754".
This defines how floating-point data types are represented, and how they are operated on, etc.

[Edwin Dalorzo]

Please download:
http://www.cs.berkeley.edu/~wkahan/ieee754status/IEEE754.PDF