Keyword Search Keyword Driven • Relevance Ranking • Too much or nothing • Have to know what to ask • Machine defined

Traditional Navigation Anticipated • Fixed relationships • Drill down • XML • Consultant defined

Relational Navigation Unanticipated • Dynamic relationships • Drill, pivot & tumble • RDF/OWL • User defined

Keyword Search is the most common methodology and a good starting point if you already know what it is you are looking for. With this methodology you drill down using a series of keywords that will deliver results through algorithmbased relevancy rankings. The most obvious benefit is ease of use; you only type in a keyword and up comes the results. So if you search on the keyword “cancer” you would get 176,000,000 results with those listed on page one being presumably the most relevant.

Unfortunately search alone can not illuminate relationships across the data sets, so you have to know specifically what you are looking for in order to find it. You will either get too many results or none with no way of knowing what alternatives exist.

Traditional Navigation goes beyond traditional keyword search to deliver a list of categories available for browse in addition to any results. With this methodology you build a query dynamically based on available facets of information presented by the user interface. These facets have been hard coded as part of a hierarchal tree structure and the application editor determines which facets are visible at specific locations. In this way, the user is “guided” to specific information the editor wants the user to “discover.” So a search on “cancer” would provide categories like “types”, “pharmaceutical cures”, “treatment centers”, and “oncologists” if those were the specific categories the editor wanted the user to see.

All relationships within and across the data are fixed within the hierarchal tree structure and require a fair amount of development resource to setup and maintain. In traditional navigation the addition of new categories like “natural cures” to the example above would require a significant amount of new development.

Relational Navigation takes Traditional Navigation a step further by delivering transparency to uncovering the relationships within and across the data regardless of format or original source. More importantly, Relational Navigation relies on “graphbased” technology which allows equal navigation in any direction across all views. In other words you can pivot across subjects in addition to drilling down and expanding up. So, referring to the same example above, with Relational Navigation you could have drilled down on a new drug for cancer treatment and then pivoted to view the clinical studies on that drug for adults versus children or its effect on various types of cancer.

This allows for the blending of internal and external content for greater usability because new data and subsequent relationships are added seamlessly based on administrator controlled information feeds. You get navigation, browse suggestions, as well as revelations of previously unseen relationships across data sets. With Relational Navigation you can give the user more control over the exploration and unanticipated discovery of information while controlling who has access to what through security and data provenance.