... "If the number is exactly equal, or equal+1 or equal-1".

Your problem statement seems to assume integer operations: "If the number is exactly equal, or exactly equal+1 or exactly equal-1." The tests given by ikegami here are not quite equivalent to your OPed test if the numbers are not integers: (update: ikegami already assumed integers in his reply, but the following examples are still pertinent.) | If the numbers are not integers:

c:\@Work\Perl\monks>perl -wMstrict -le
"my $x = 2;
 my $y = 2.00000000000000123;
 print qq{$x $y};
 ;;
 if ($x == $y or $x == $y-1 or $x == $y+1) { print 'in exact range'; }
 if ($y-1 <= $x && $x <= $y+1)             { print 'in <= range';    }
 if (abs($y-$x) <= 1)                      { print 'in abs range';   }
"
2 2
in <= range
in abs range
[download]

Remember that Perl numbers are essentially floating-point doubles. This may introduce some twists and pitfalls. (Update: Certain combinations of operations may produce a number like 2.00000000000000123 that may appear, in certain circumstances, to be an integer, but isn't. See What Every Computer Scientist Should Know About Floating-Point Arithmetic.) One way to force integer operations is with the integer pragma:

c:\@Work\Perl\monks>perl -wMstrict -le
"use integer;
 ;;
 my $x = 2;
 my $y = 2.00000000000000123;
 print qq{$x $y};
 ;;
 if ($x == $y or $x == $y-1 or $x == $y+1) { print 'in exact range'; }
 if ($y-1 <= $x && $x <= $y+1)             { print 'in <= range';    }
 if (abs($y-$x) <= 1)                      { print 'in abs range';   }
"
2 2
in exact range
in <= range
in abs range
[download]

Update: Changed both code examples to use the floating-point number 2.00000000000000123 (vice 2.5) for $y.