#15 Page not founds clutter index

[Simon Wilkinson]

Some badly configured sites return "Page not found" pages
with a HTTP response code of 200 (OK). This leads the
indexer to index them, and rapidly build up a load of
duff pages if the MD5 hashes don't catch them (if the
page is generated dynamically, and includes differing
content depending on the URL/Referrer, the hash will
be different).

So, we should have a post-filter which can deal with
this in some way.