Mirago's search technology has evolved substantially over the last few years. Now in its third generation it has come a long way since inception and will undoubtedly continue to evolve as the dynamic world of the internet moves forward.

In common with other leading search engines, Mirago does rather more than just index the words on web pages. Almost without exception the major engines use the links between pages as well as the text on those links to determine the importance of web pages. This citation based model greatly improves the ability of engines to determine the subject of a page. Typically links use short focused phrases such as 'Contact Widget plc' from which search engines are able to infer that target page contains contact information for an imaginary company called Widget plc.

Mirago take this model further by looking at the themes of the page from which the link originates and comparing it to the theme of the page to which the link points. Why is this relevant or indeed beneficial? Put simply, links between similarly themed pages are of more use when determining the relative merits of pages. Lots of people put links on their site to download the latest version of Microsoft's Internet Explorer. Few are from other browser manufacturers. More useful are the links between sites which are about the same subject. Dog owners' clubs who have web sites may well link to other sites which contain information about specific breeds. Such links are of far greater importance than just links to the latest version of the browser required to view a web page.

Of course, Mirago's technology considers many other factors when deciding what information to present in response to a search. Factors such as how deeply within a site the information is buried play a part. How often the page changes and when it was last changed has some influence.

The most important factor for any web page is the authenticity of its information. There are unfortunately certain individuals and companies who attempt to subvert the impartial nature of search results by artificially manipulating the text on a page as it is indexed by search engines. This is known generically as spamming. Mirago has developed a wealth of leading edge technology to detect and deal with it. An obvious example of spam is hidden text. Another type is the so-called doorway page which immediately redirects a searcher to a different web page. Web pages which employ such tricks do not benefit from their endeavours.

Essentially Mirago has one guiding principle in the design of its search technology. Namely that the technology should evaluate the relative merits of diverse web pages in the same manner as would a human doing the same job.

Q3, the Mirago Query System, is used to search for information drawn from web pages; the so-called organic search. It is also used to search information drawn from database systems, otherwise known as the Trusted Feed Program. An index is created from each of these information sources. The former is created by Henry, The Mirago Robot. The latter by an XML data feed from Mirago's Trusted Feed partners...

Trusted Feeds

Not all useful data is readily accessible through static web pages. Large repositories of data are stored in database systems. Frequently these databases are interrogated by proprietary search techniques. Consequently robots such as Henry, The Mirago Robot, are unable to read the information for inclusion in Mirago's web index.

To overcome this, Mirago operates a Trusted Feed Program. The program is managed by nominated resellers and allows sites containing large quantities of information to submit their data automatically for inclusion in Mirago's searchable index. The resellers manage the data on behalf of the sites and present it to Mirago as an XML feed.

Mirago operates a special robot which reads just these feeds and produces similar indexes to those created by Henry. Being independent, the frequency of update of these indexes is not tied to the normal cycle. As such the Trusted Feed Program updates the searchable information each day.