The library catalog holds a wealth of information about library resources. The Voyager software used at Cornell, like most library systems, stores different data in a variety of ways. For example,

• Titles, authors, and subjects are stored in MARC bibliographic records.
• Locations and call numbers are stored in MARC holdings records and item records.
• Circulation information is stored within a proprietary database.

While the integrated library system is fairly successful at joining these different data sources into a coherent public interface, the system is certainly not designed to make selected data accessible to external applications. These applications may need only a small subset of data from a given record, or may require a unified view of data from multiple records.

We will examine multiple web-based projects that have approached this problem in different ways:

JAbbr, a tool for deciphering journal abbreviations, uses a large export of title and alternate title fields to help users match an abbreviation to the precise catalog record for the corresponding title.

New Books uses a monthly export of recently-cataloged monographs to provide a classified list of new additions to the library collection, complete with subject-based RSS feeds. It is also possible to filter the new titles (~6,000 each month) by language, classification, or unit library.

Mann Library Reserves uses a live database query to return the titles, call numbers, and availability of reserve materials for a given course. This information is then displayed in an easy-to-read layout within the Mann's CommonSpot-based website.

Mann Library Equipment & Rooms also uses a live database query to display the availability of study rooms throughout the library, as well as the status of various cameras, laptops, and GPS units.

We hope to spark a discussion of other methods of extracting data from the catalog, and will try to generalize the cases in which one approach is more appropriate than another.