The two key components of Sage are a browseable hieararchy of resources organized by subject and type, and a full-text search engine. These are provided by a combination of locally-written and commercial software that allow users to browse and search our collection of web-based resources.

The foundation of Sage is a relational database that contains information on individual resources (such as title, URL, description), subjects (such as Chemistry or Religion), and types (such as Dictionaries, Pictures/Images or Web Megasites). These resources have been selected by UCSD library subject specialists as being valuable for research by students, faculty and staff. The database also contains a hierarchy of subjects and other information needed to build the dynamic pages. More information about our metadata can be found in our Database Structure document.

The browse functionality is provided by locally-written Java servlets and Server-Side Javascript applications that access our relational database. These programs build dynamic pages according to subject, type, the hierarchy, or any number of other database fields. Our content authors can also define their own set of resources, either by existing database fields or by selecting any arbitrary set of resources, and apply thier own formatting information - allowing for an infinite number of ways different resources can be combined and displayed to meet the needs of different users.

The search functionality is provided by the Netscape Compass search engine. Our installation of Compass has been modified extensively to meet our needs. First, instead of spidering static pages, we generate pages from our database to control which pages are spidered and to what depth. When the resources in our database are spidered by Compass, the subject and type information is collected as well, so searches can be limited by subject and type. We run several instances of Compasss: one for the resources in our database, another for E-Journals, a third for our static HTML pages, and a fourth for other sites that are linked to by the resources in our database. This allows us to control which category is searched. The output of Compass has also been customized to fit in with the design of our dynamic pages, and to have links back to our database-driven pages for each subject and type associated with each resource.

Sage also gets some of its data from another commercial software package: Roger. Because our E-Journals are cataloged in Roger, we import them from Roger instead of duplicating data entry. Each E-Journal record is retrieved from Roger in XML format, converted to SQL, and inserted into our database where it can be displayed along with our other resources and spidered by Compass. More information about how we import E-Journal records from Roger can be found in our Importing E-Journals from Roger document.