The contents of the CSTR electronic library are currently stored on both magnetic disk (primarily for indexes and programs) and on an optical disk jukebox containing the full text and page images of the technical reports. Because most of the technical reports were not available in a machine-readable form, the database is made up primarily of scanned page images, and the output of OCR run on those page images. Some of the more recent reports are available in PostScript or in some mark-up language (such as troff or TeX). The database was created by scanning each report using a PC and commercial OCR software, and creating a simple bibliographic record for each item. These bibliographic records use a format developed for exchange of CSTR information via email[\protect\citeauthoryearCohen1992].