The documentation for Math::BigFloat says that zero is stored as 0E1 - presumably for the same reason as DBI returns 0E0. However, although the exponent() function does return 1, it still evaluates to false in a boolean context:

use Math::BigFloat;

my $x = Math::BigFloat->bzero();
my $y = '0e+1';

print $/, "x is " , $x->bsstr();
print $/, "y is " , $y;

print $/;
print $/, "x + 0 is " , $x + 0;
print $/, "y + 0 is " , $y + 0;

print $/;
print $/, "x is " , ( $x ) ? "true" : "false";
print $/, "y is " , ( $y ) ? "true" : "false";

print $/;
print $/, "x eq 0e+1 is ", ( $x eq '0e+1' )  ? "true" : "false";
print $/, "y eq 0e+1 is ", ( $y eq '0e+1' )  ? "true" : "false";
print $/, "x eq y is ", ( $x eq $y )  ? "true" : "false";


C:\iakobski\perl>test1.pl

x is 0e+1
y is 0e+1

x + 0 is 0
y + 0 is 0

x is false
y is true

x eq 0e+1 is false
y eq 0e+1 is true
x eq y is false
[download]

My questions: 1. Why? 2. Anyone know what this comment means in the docs:

"A zero is represented and returned as 0E1, not 0E0 (after Knuth)."