Search Operation Guide
  
  

The search engine uses a search query specified by the person searching. This query can contain Boolean logic and location operators (as AND, OR, and NEAR) to narrow the search parameters. The query is run against four separate sections of a document. These sections are:

  1. Document Contents: This is the actual text contained in a given document. It includes absolutely everything in the document (ie. <head> and <body> content).
  2. Document Title: This is the text contained in the <title></title> tag of an html document.
  3. Image Alt Tags: This is the text contained in the alt= part of an image tag.  You should include an Alt tag if you would like your graphical image to be included in a search.  For example:
    < src="ballou.gif" alt="Ballou Hall, Tufts most historic building" border="0">
  4. Meta Keywords: This is the descriptive text contained in a meta keywords tag (which needs to be enclosed within the head tag) of an html document. For example, "This is a description of the content of my webstie" would be in the Meta Keywords list for the following HTML document:
    <html>
    <head>
    <title>My website</title>
    <Meta name="keywords" contents="This is a description of the content of my website">
    <head/>
    <body>
    </body>
    </html>

When the query matches against a given section of a document, points are awarded to the document. The number of points awarded is a function of the number of times the query matches in a given section vs. the size of a given section. For example, if a query was searching for the word "expedition", the following paragraph:

Our dangerous expedition.

would receive more points than this paragraph:

This is the complete story of our dangerous expedition.

Even though the word "expedition" only occurs once in each paragraph, the first paragraph is smaller than the second paragraph. As a result, the matching of the first paragraph is proportionally higher for the first paragraph. This proportional matching is done for each of the four sections of a document and the results are averaged together. The average number of points for a document is used to create the document's rank.

The final output of the search results is sorted by descending rank (i.e. the higher the rank, the better the position in the search results) and then by ascending Title for documents with the same rank.

So, how does this translate into what to do to make your documents more "findable"? You should make sure that words or phrases that would probably be used by people searching for your page are included in all four sections of a document (this will keep the average higher). In other words, make sure you use meta keywords, make sure that your document title is well related to your page, and make sure that you use alt tags in your images. It's recommended that you don't try to craft your content to reflect searchable terms. Your content should say what it's supposed to say - the content of the document only counts a quarter of the rank, so if you work with the three other sections, you'll be able to boost your rank without having to significantly alter your content.

Finally, you should only uses these methods on documents you want to show up higher in the search results. If you apply these methods to every single document on your site, you're simply going to dilute the results of the search. Try to stay focused on important pages (such as your site's home page) so it moves closer to the top and any additional pages that have related content will appear below it.

Please keep in mind that using these rules, you won't always be able to make your document float to the top of the search results (due to the fact that it is typically difficult to construct the content of a document to contain common search terms), but you'll be able to make sure that the document at least is "findable" in the search results.