strange behavior

[Carlos M. Pascal]

Clips 6.3, Win7 64biti

The expression (= 0.4 (+ 0.2 0.2)) is true,
but (= 0.4 (- 0.6 0.2)) is false. Why?

[Gary Riley]

0.1 and its multiples are irrational in binary format so equality comparisons to these values can yield unexpected results. You'll get similar results in the C language underlying CLIPS:

#include <stdio.h>

int main()
  {
   printf("0.4 == (0.2 + 0.2) is %d\n", (int) (0.4 == (0.2 + 0.2)));
   printf("0.4 == (0.6 - 0.2) is %d\n", (int) (0.4 == (0.6 - 0.2)));
   return 0;
  }

And in other languages such as Java:

public class Test
  {
   public static void main(String args[])
     {  
      System.out.println("0.4 == (0.2 + 0.2) is " + (0.4 == (0.2 + 0.2)));
      System.out.println("0.4 == (0.6 - 0.2) is " + (0.4 == (0.6 - 0.2)));
     }  
  }

Many languages support arbitrary precision decimal numbers (such as the BigDecimal class in Java) to address this issue.

Some additional information can be found here: http://stackoverflow.com/questions/1089018/why-cant-decimal-numbers-be-represented-exactly-in-binary

[Carlos M. Pascal]

Thank you, Gary Riley!

I have used an unorthodox solution

(eq 0.4a (sym-cat (- 0.6 0.2) a))