The nature of typical mechanized data bases with which university information centers and libraries will be working is examined to provide background information for data base retrieval efforts. Format and content of the data must be taken into consideration before meaningful retrieval can be achieved. Hence both of these aspects are treated in detail. The discussion covers a range of data bases, including nationally sponsored projects such as Medical Literature Analysis and Retrieval System (MEDLARS), Machine-Readable Cataloging (MARC), the United States Census Tapes (1/1000 sample), and the Educational Resources Information Center (ERIC) report resumes. Also discussed are some local projects, including the University of California, Los Angeles, Brain Information Service system and a generalized Metropolitan Center Data Bank. Various types of formatting are presented in this array of data bases, and several examples among types of requests that can be directed against such data bases are considered. The report is followed by another which assesses the applicability of generalized or task-oriented programs to these data bases (LI 000 499). (Author/CM).