Just want to point out that the algorithm as given is incorrect. n is the number of elements in the array, so the first line should read:

function reverse_in_place(a[0 .. (n - 1)])
[download]

Alternatively, you would need to change the arithmetic:

function reverse_in_place(a[0 .. n])
    for i from 0 to floor( ((n + 1) / 2) - 1 )
        tmp  := a[i]
        a[i] := a[n - i]
        a[n - i] := tmp
[download]

Update 1: Implementing the algorithm in Perl is straightforward, provided you are allowed to use the built-in functions split and join (scalar isn’t really needed in the sub):

use strict;
use warnings;

my $string = 'abcdefghi';

print 'Original string: ', $string,                   "\n";
print 'Using function:  ', reverse_in_place($string), "\n";
print 'Using reverse:   ', scalar reverse  ($string), "\n";    # For c
+omparison

sub reverse_in_place
{
    my ($string) = @_;
    my  @array   = split //, $string;
    my  $n       = scalar @array;

    for (0 .. $n / 2 - 1)
    {
        my $tmp             = $array[$_];
        $array[$_]          = $array[$n - $_ - 1];
        $array[$n - $_ - 1] = $tmp;
    }

    return join('', @array);
}
[download]

Update 2: Fixed out-by-one error in the for loop. Thanks to j0se for pointing it out.

Hope that helps,