Yet another look at it...

#!/usr/bin/perl

use warnings;
use strict;

use LWP::Simple;
use HTML::Parser;

# 'print' URL to save a smidge of bandwidth
my $url = "http://perlmonks.org/?displaytype=print;node_id=31832";
if (my $page = get($url)) {
    my $parser = HTML::Parser->new( api_version => 3 );
    $parser->handler( start => sub {
            return if shift ne 'title';
            my $self = shift;
            $self->handler( text => sub {
                    my $title = shift;
                    # Only ready for 4-function math with integers
                    if ($title =~ m#((\d+) (\+) (\d+))#) {
                        print "$1 = ", eval("$2 $3 $4"), "\n";
                    }
                }, 'dtext'
            );
            $self->handler( end => sub {
                    shift->eof if shift eq 'title' ;
                }, 'tagname,self'
            );
        },
        'tagname,self'
    );

    $parser->parse($page);
}
else {
    print "Unable to fetch '$url' for some reason.\n";
}
[download]

I chose to leave the database layer out of it this time.