use strict;
use GDBM_File ;

my %hash;

tie %hash, 'GDBM_File', 'words.dbm', &GDBM_WRCREAT, 0640;

my $dict_file = '/usr/share/dict/longlist';

open DICT, "<$dict_file" or die "open($dict_file): $!";

my @words = ();
my $prefix = '';

while(<DICT>) {
  chomp;
  $_ = lc($_);
  if(substr($_, 0, 3) ne $prefix) {
    if(@words) {
      $hash{$prefix} = join(' ', @words);
      @words = ();
    }
    $prefix = substr($_, 0, 3);
  }
  push @words, $_;
}

untie %hash ;

##</code><code>##

use strict;
use Benchmark;
use GDBM_File ;

my $dbm_file = 'words.dbm';

my %hash;
$/ = undef;

my $data = <DATA>;
my @words = map {lc($_)} ($data =~ /(\w+)/g);
print scalar(@words), "\n";

timethis(1, 'one_tie()');
timethis(1, 'many_ties()');

sub one_tie {
  my($found, $not_found) = (0, 0);
  tie %hash, 'GDBM_File', $dbm_file, &GDBM_WRCREAT, 0640;
  foreach (@words) {
    my $prefix = substr($_, 0, 3);
    if(exists($hash{$prefix})  and  $hash{$prefix} =~ /\b$_\b/) {
      $found++;
    }
    else {
      $not_found++;
    }
  }
  untie %hash ;
  print "Found: $found\nNot found: $not_found\n";
}


sub many_ties {
  my($found, $not_found) = (0, 0);
  foreach (@words) {
    tie %hash, 'GDBM_File', $dbm_file, &GDBM_WRCREAT, 0640;
    my $prefix = substr($_, 0, 3);
    if(exists($hash{$prefix})  and  $hash{$prefix} =~ /\b$_\b/) {
      $found++;
    }
    else {
      $not_found++;
    }
    untie %hash ;
  }
  print "Found: $found\nNot found: $not_found\n";
}


__DATA__

insert output of 'perldoc perltoot' here :-)