Vol. 24, No. 2/Winter 2002
PLANNING AN ARCHIVE:
DC4 may sound like the model number of a commercial airplane, but it is really the new in-house database for the USA TODAY library. It went live in mid-2001 after three years of planning, developing and testing. It replaced a custom-designed text archive using PLS (Personal Library Software), which was in use the previous five years.
The result is a powerful database that affects several aspects of library operation, as well as research capabilities for the newsrooms. All reporters and editors now have desktop access to USA TODAY, Gannett News Service, USA Weekend, Baseball Weekly and the online card catalog for books and periodicals. It also handles check-ins for these items.
Designing the database
Database development was a partnership among three entities: the USA TODAY library, USA TODAY Information Technology and Cincinnati-based GMTI (Gannett Media Technologies International), a division of USA TODAY’s parent company. The library team was led by Library Director Barbara Maxwell, supported by reference librarian Jean Simpson, indexer Jennifer Kelemen and Acquisitions & Cataloging supervisor Susan O'Brian.
According to Maxwell, "We had 100% input into the software design of the system." Maxwell and the Library were named the designers of the system. "This is very unusual when dealing with technical people," she says. "The system was built as much as possible to our specifications, and was built to mirror much of the existing functionality and workflow of the previous archive."
"The beginning of the project started with a meeting in August 1998 to go over the specifications for the new archive," Maxwell says. A new system was needed because the software powering the previous archive had been sold to AOL and was no longer being supported. The PLS system stored data in HTML and used early Web technology. DC4 uses software that works within a browser setting, except for the archiving/indexing module.
"New technology will allow us to add more features to the archive, and build a system which integrates images, PDFs, text, etc.," says Maxwell. Currently the Library's photo archive is not linked to the text archive. However, the Library is about to start reviewing new photo archiving systems and will look at integrating both products.
Simpson says, "I was involved with various aspects of the planning, development and final implementation of DC4. I worked on the initial system specs with Barbara Maxwell, gave input and opinions on screen design, created specs for search engine functionality and did a large amount of testing and timing of the search module."
David Connerth, Director of Software Development, was the main person the library worked with at GMTI. "I was the project manager, in charge of system design and construction," Connerth says. "I had three staff members who worked on the project, one in a DBA (data base administrator) capacity and two as programmers. The system underwent extensive testing. The feedback from the USA TODAY staff was invaluable."
This was a case where both the journey and destination were worthwhile. "It was very interesting to see the process evolve from beginning to end," Simpson said. "We spent time benchmarking other database systems and internet search engines and evaluated what we liked and didn’t like. This helped us when making decisions on system details such as screen design, search functionality, system speed requirements and navigation."
O’Brian designed the card catalog prep page. "I used the original form [from PLS] but added one or two fields and drop-down menus. Barbara and I also added an extra search screen for admission records based on books and magazines entered. It's able to search on any field in the database and create Excel tables based on the results."
Kelemen says that since indexing is time-sensitive and faced with early afternoon deadlines, "Planning a streamlined indexing module was of utmost importance. Eliminating any unnecessary typing for indexers was key. Each view of the module must be intuitive, clearly labeled and easily maneuvered in order to keep the workflow at a maximum."
According to Chuck Walley, Manager of Enterprise Solutions for the USA TODAY Information Technology department, "Because of the relationship between Gannett, USA TODAY and GMTI, we had a much closer partnership with the vendor than a project would normally have. This fostered a close working relationship between the USA TODAY Library Staff and the GMTI developers. Normally, a few layers of management are placed between the end-users and development staff -- here there were none."
Connerth points out that newspaper databases are different: "Several things make them unique. One is the diversity of systems, which often include public web sites, editorial systems, pagination systems, billing systems, etc. Newspaper people often have immediate time demands placed on them (especially since 9/11)."
Here are recommendations from Maxwell for librarians planning on developing new databases: "You need to make sure that the developers are educated about the system they are to develop, and their user base. We spent a lot of time in the early development educating the developers on what a library is, and what the archive should do. Once you get past this point, it makes the development easier. Flexibility is also a must. Designing a system like this, which is complicated, takes a lot of time, and change. There are always limitations because of software and hardware, which may affect how you reach your desired goals of functionality. For example, in Oracle, which is the underlying database powering DC4, there are certain search restrictions which had to be overcome -- like using quotes to define an exact phrase. Oracle will not accept a single quote."
Connerth offers this tip: "I would suggest that they look at the technology underlying a given system, and how that compares to the latest trends. Things to consider are the software vendors involved, and the possibility of using lower cost alternatives, such as open source software (e.g. Linux)."
From Walley’s perspective, he advises "Whenever you are choosing and implementing any new system, you need to understand, define and prioritize your requirements. Match your prioritized requirements against the system’s features and you will quickly separate critical needs from window dressing. When you narrow the choice, get references and plan to spend some time with them independent of the vendor." He also recommends you find out if the vendor delivered on its promises and met the schedules of the initial contract.
Finally, Maxwell advises: "In order to design a system like this, the lead librarian must be willing to think like a techie. I had to read the administrator's guide to Context in order to understand how the search engine works. I also traveled to Dayton, Ohio to discuss the search engine NEXIS uses. If you are unwilling to take these tasks on, designing a system like this will be very difficult."
Bruce Rosenstein is a reference librarian for USA TODAY. He can be reached at email@example.com.