This report is intended to facilitate discussions about if, when and how Computing & Communications should provide and centrally support a document image storage and retrieval system for the administrative offices of the University of Washington.

Over the past five years, several departments have considered imaging applications. Some have launched substantial projects to define requirements, calculate costs and benefits, and meet with vendors. Most of these projects have stalled at one point or another, frequently due to the expense of commercially available imaging systems. None of the organizations within the scope of this report have demonstrated a favorable financial return on investment for a departmental imaging system. Some believe imaging is justified based on broader criteria, but they have not acquired or installed solutions yet.

In Fall, 1994, the Student Loan Office was prepared to purchase an imaging system. The chosen application was a proprietary LAN-based product, which did not fit in well with the University’s technical vision. In the interest of ending up with the best overall solution, the Student Loan Office delayed their project while Computing & Communications considered alternatives that could meet departmental requirements for imaging within the long-term technical direction of the University.

Key elements of that direction include:

• Information access from X terminals and PCs
• Distribution of information over the campus network
• Consistency with campus and higher education standards
• Supportability of systems by Computing & Communications

In order to determine whether or not such a system could be justified, it was necessary to inventory the potential applications of such a system and to estimate the costs. This report is the result of that analysis.

In cooperation with Records Management, a series of questions was assembled to evaluate the applicability of imaging to any given set of files. Interviews were held with people familiar with each of the major record series maintained by administrative departments. A screening process identified which files are good candidates for centrally administered imaging and which are not. Combined with additional departmental objectives and constraints, the list put into a recommended implementation order.

This report addresses the imaging requirements of administrative offices on the upper campus which would normally be supported by Computing & Communications. Included are most of the offices in Business and Finance and Student Affairs. Excluded are the medical centers, academic departments and self-sustaining units. Scope was further restricted to consideration of systems which have broad applicability as opposed to highly specialized application software, such as Resumix. There are certainly opportunities in these other areas which could provide additional benefits.

A generic electronic document storage and retrieval system will meet the business needs of most of the candidate departments. The only high-level requirements are to scan, index, store, search for, display and print business documents. The architecture presented here is designed to meet those requirements within the technical environment of the University.

This architecture is modular, so that any piece of it may be replaced without impacting the overall system. If a new browser becomes available, it can be used without affecting the storage, indexing, or searching modules. If the indexing database is changed, the software used for scanning and displaying images will not be affected. And if the volume of data for a given site justifies an optical jukebox, only the image server would need to be changed.

By keeping the index server separate from the image server, document searches can be very fast. It also makes it possible to use common tables in the indexing of different record series. For example, a single employee table storing name, employee number, and social security number could be used to validate entries into the benefits, personnel, and payroll image files. All three keys could be used to search for a document even though only one key was entered when the document was stored.

Each department will do their own scanning using an appropriately sized scanner for their application. Off-the-shelf scanning software will be used to capture the images and store them in an industry standard format on a network host. Custom-developed routines will be used to index and store the images. The indexing can be done separately from the scanning, allowing greater flexibility.

Images will be stored on httpd servers and the index will be stored separately in a database that can be searched using Willow. All required storage software is already in use at the UW.

Users can search the database based on pre-defined keys using Willow. They can bring up descriptive fields and thumbnails to help them identify which document they want. Then they can select a document for retrieval, which is passed to either Mosaic or Netscape on their computer for display. All required retrieval software is already either in use or in development at the UW.

It is extremely difficult to accurately predict the benefits of an imaging system. Each of the departments that have evaluated imaging have developed their own estimates of benefits as part of their reports. Excerpts from Payables and Accounting Operations and Student Loan Office reports are included as appendices. Those estimates do not include some of the key benefits, such as the real value of increased office space and decreased legal exposure due to lost documents.

One crucial benefit of implementing imaging technology centrally in the near future is the decreased risk of multiple departments implementing multiple different and incompatible imaging systems on their own. Over half of the top candidates, and a good number of the other candidates, have already seriously evaluated imaging technologies. At least three offices (Student Loans, Purchasing and Accounts Payable, and the Registrar) have already met with vendors. Many more have spoken with vendors and reviewed vendor literature. One is considering purchasing a CD-WORM drive to experiment and another plans to contract with a service bureau before the end of the current biennium. Furthermore, industry trends indicate that imaging will continue to grow and spread in coming years.

If Computing & Communications does nothing with imaging, it is easy to imagine ten incompatible departmental imaging systems in place on campus within five years. By that time, some of the vendors may already be out of business, some of the proprietary applications may no longer be supported, and C&C will not have the skills to maintain these systems.

On the other hand, if C&C provides strategic direction and support for departmental imaging requirements ahead of time, then most departments will wait for the centrally supported solution. Five years from now, the University will have imaging systems aligned with the technical vision and skilled resources to maintain them.