I've been looking into the Nutch search engine for a little while, and I've finally gotten around to implementing the beginnings of a site search. What I've found interesting is that for a given search result, you can get the information about why it is where it is in your own SERP, and will kick out the anchor text that you've given your own page. It's nothing groundbreaking, but it is interesting to take a look at the information. Have a look at a search results page: http://www.stevekallestad.com/searc...g=en&hitsPerPage=10&hitsPerSite=0&clustering= And then the explanation for a given result: http://www.stevekallestad.com/search/explain.jsp?idx=0&id=94&query=prototype&lang=en I haven't really got a true understanding of all the scoring information, but to the FAQ entry on the explain page is at this link. And the anchor text I gave that result from my own site (I've only set this up for a site search) http://www.stevekallestad.com/search/anchors.jsp?idx=0&id=94 You can see that the Movable Type templates give pages a lot of links, but a lot of the links have terrible anchor text in the form of dates. To check out your own site, you'd have to install nutch and have it crawl your site. You can actually do that from your own local windows machine fairly easily by following the tutorial. It requires Java, Tomcat, and cygwin - but all three are actually pretty easy to install. It's easier, though if you have a linux box or a Mac handy.