Blacksburg, Virginia, U.S.A.—VTLS Inc., The Princeton University Libraries, and Princeton's Computing and Information Technology Department are working on a joint project to scan the libraries' public union card catalog, containing 6.5 million 3x5 catalog cards, to create a database of digitized images. The database represents 1.75 million titles acquired and cataloged before 1980, predating Princeton's library automation and online catalog. Initially, the image database will not be integrated with the libraries' online catalog, but will work in conjunction with it. The 6.5 million images represent the largest image database created in a library.
The project is planned in several phases. In the first phase, already completed, library staff analyzed the card catalog and enhanced the guide cards to create an index for the database of card images. In three months, 65,000 index points were created. In the second phase, beginning in June 1993, VTLS Inc. will use single-sided scanners, located on site at the Firestone Library, to generate digital, bit-mapped images of each card in the catalog. Princeton staff will monitor the process for quality control, using workstations to view statistically significant samples of the scanned images.
In the third phase, which will overlap the second, VTLS will develop an interface for the display workstations-40 computers running NEXTSTEP for Intel processors—to be used by library patrons and staff for browsing the database of card images. VTLS will use NeXT Inc.'s NEXTSTEP object-oriented development environment to build the graphical user interface. The NEXTSTEP operating system was selected because of its multimedia and client-server capabilities, and because NEXTSTEP enables the rapid development and deployment of this "mission critical," custom application, as compelled by the continuing deterioration of the catalog cards. Using the NEXTSTEP¬based interface, users will browse a hierarchical index tree to retrieve, view and save card images; library staff will also edit scanned images, using graphics software. Initially, 40 display workstations, powered by Intel 486 processors, will be located in the Firestone Library and at all branch locations where there are now card catalogs.
In a later phase, optical character recognition (OCR) will be used to convert the digital image of each card to fully tagged and indexed records of text in standard MARC format such as those found today in Princeton's online catalog. With the present rate of rapid improvement, OCR technology may be reliable enough to make feasible this conversion of the card images.
Several factors led Princeton University to rely on scanning rather than a standard retrospective conversion to combine their card and online catalogs. Scanning will be far more cost- and time-effective than a labor-intensive retrospective conversion. Much of the local information in the card catalog, which is specific to Princeton's collection, is complicated and esoteric and would require extensive editing by library staff. A project in retrospective conversion would therefore create an unacceptable backlog in current cataloging. Moreover, since many of the cards are in poor condition and rapidly deteriorating, the information they contain may be quickly preserved by creating digital images through scanning. Due to the uniqueness of Princeton's catalog, preserving the database by the traditional approach to retrospective conversion, as opposed to scanning, would require inordinate time and money. According to a recent estimate, retrospective conversion of the card catalog would require 11 years to complete and cost $18 million. In contrast, Princeton's image database project will be completed in nine months and will cost less than $1 million.
VTLS was selected to become a partner in this project after Princeton University Libraries and Princeton's Computing and Information Technology Department conducted an intensive, joint evaluation of options and systems developers.
Dr. Vinod Chachra, president of VTLS Inc., said, "VTLS is delighted to have the opportunity to work with Princeton University on such a unique project. The developments from this project related to handling and storing lar e image databases will surely be made available to other libraries having similar needs.
Princeton is very excited about this imaging project. The project will preserve valuable information, permit network access from non-library locations, provide enhanced searching capabilities to the libraries collection, create more efficient tools for maintaining the catalog, and provide a relatively cost-effective and error-free alternative to standard retrospective conversion. According to Project Director Eileen Henthorne, "The project may serve as a model for other universities and repositories that want faster, more effective access to information and that are concerned about the preservation of their information. Princeton is pleased to be working with VTLS on such an ambitious project."