Ok does this make sense?:

#!/usr/bin/perl
use warnings;
use strict;

use XML::Simple;
use Carp;
use File::Find;
use Data::Dumper;

my %hash;

my $default_dir = "C:/Main/work";
if (@ARGV == 1) 
{  
   $default_dir = shift @ARGV
)
elsif (@ARGV >1)
{
   die "too many command line args\n";
};

find (\&process_file_name, $default_dir);

# find() will start at the $default_dir
# and call process_file_name for each file or
# directory that it finds at or "beneath" the
# $default_dir

sub process_file_name
{
   my $current_file = $File::Find::Name;
   #$File::Find::name is the complete pathname to the file.
   return unless (-f $current_file and
                  $current_file =~ /_service\.xml$/);
                  
   # we are at a data file (not a directory)
   # and this file name ends in _service.xml
   
   process_the_file ($current_file);
}

sub process_the_file
{
   my $file_path = shift;
   ...open the $file_path
   ...figure out what you want from that file...
   ...$hash{$file_path} = "some data"; 
              # this is how the data is "returned"
              # %hash has "global scope"
              # add an entry to %hash instead of
              # returning a value from this sub
   ...close the file
}
[download]