You are right, and thanks for pointing out. My original analysis took the assumption that both s/// and split iterate through the sentence with the same performance, however that was wrong, and split() is much slower:

use strict;
use warnings;
use Benchmark qw (:all);

my $txt = "a" x 100;

sub seperate {
    split //, $txt;
}

sub replace {
    $txt =~ s/a/b/g;
}

my $result = timethese(100000, {'seperate' => \&seperate, 'replace' =>
+ \&replace});
[download]

This gives:

Benchmark: timing 10000 iterations of replace, seperate...
   replace:  0 wallclock secs ( 0.00 usr +  0.00 sys =  0.00 CPU)
            (warning: too few iterations for a reliable count)
  seperate:  2 wallclock secs ( 1.20 usr +  0.00 sys =  1.20 CPU) @ 83
+05.65/s (n
=10000)

C:\Perl\bin>perl -w math1.pl
Benchmark: timing 100000 iterations of replace, seperate...
   replace:  1 wallclock secs ( 0.03 usr +  0.00 sys =  0.03 CPU) @ 31
+25000.00/s
 (n=100000)
            (warning: too few iterations for a reliable count)
  seperate: 16 wallclock secs (12.50 usr +  0.00 sys = 12.50 CPU) @ 80
+00.00/s (n
=100000)
[download]