Release calc version 2.10.3t5.45

help/abs

@@ -15,14 +15,18 @@ TYPES
     eps		ignored if x is real, nonzero real for complex x,
 		defaults to epsilon().
 
-    return	real
+    return	non-negative real
 
 DESCRIPTION
-    If x is real, returns x if x is positive or zero, -x if x is negative.
+    If x is real, returns the absolute value of x, i.e. x if x >= 0,
+    -x if x < 0.
 
-    For complex x, returns the multiple of eps nearest or next to nearest
-    to the absolute value of x.  The result usually has error less in
-    absolute value than abs(eps), but should not exceed 0.75 * abs(eps).
+    For complex x with zero real part, returns the absolute value of im(x).
+
+    For other complex x, returns the multiple of eps nearest to the absolute
+    value of x, or in the case of two equally near nearest values, the
+    the nearest even multiple of eps.  In particular, with eps = 10^-n,
+    the result will be the absolute value correct to n decimal places.
 
 EXAMPLE
     > print abs(3.4), abs(-3.4)
@@ -35,7 +39,7 @@ LIMITS
     none
 
 LIBRARY
-    none
+    NUMBER *qqabs(NUMBER *x)
 
 SEE ALSO
     cmp, epsilon, hypot, norm, near, obj