for ( my $i = scalar(@$results) - 1; $i >= 0; $i-- ){
 for ( my $j = 1; $j <= $i; $j++){
  if ( $$results[$j - 1]{'hits'} < $$results[$j]{'hits'} ){
      my $temp = $$results[$j - 1];
      $$results[$j - 1] = $$results[$j];
      $$results[$j] = $temp;
   }
 }
}
[download]

This is the first thing that stands out. I am guessing that you want the ones with the most 'hits' first correct? Why not just use a simple sort. For example, I think this has the same functionality as the above code without me trying to figure out exactly what it was I was trying to accomplish a couple of months down the road (not that commenting it wouldn't help in this respect.)

@$results = sort{ $b->{'hits'} <=> $a->{'hits'} } @$results;
[download]

Even if you don't want to take it this far you can exchange these couple lines

my $temp = $$results[$j - 1];
$$results[$j - 1] = $$results[$j];
$$results[$j] = $temp;
[download]

with

($$results[$j - 1], $$results[$j]) = ($$results[$j],$$results[$j - 1])
[download]

which is similar to one of the first things i remember seeing in the camel (the swapping without a temporary variable thing.

Though there might be more things this just stuck out at me as more a C type thing than a Perl type thing

-enlil