RSS
Automating Document Delivery: A Case Study
George R. Thoma, Susan E. Hauser, Glenn Ford
National Library of Medicine
Rosalie Stroman
National Institutes of Health Library
Bethesda, Maryland
Interlibrary loan (ILL) services are currently heavily labor-intensive and largely mail-delivered. An approach is presented here to transition to a more automated, possibly cheaper and more timely service. The approach discussed assumes that the current mix of delivery modes (mail, fax and Internet) will migrate to a predominantly fax- and Internet-based document delivery. To this end we have developed a prototype system we call the Workstation for Interlibrary Loan, or WILL.
The approach is based on the concept of replacing the labor intensive manual ILL activity by a system that, by automating every function that can be automated, reduces the load on human operators, and thereby cost. The system is designed to automatically retrieve ILL requests received by a request routing software package (e.g., DOCLINE), and to use the information in the request (requester name, address, fax number, Internet address, preferred delivery mode, etc.) to channel the scanner output (the images) to modules that automatically convert the images to the required file formats, and send them out. In a medical library environment, WILL would receive an ILL request from DOCLINE, process the request to determine the required delivery mode and print the ILL request with a barcode. The printed request informs the operator of which document to go get from the shelves. The operator's initial step in the scanning process is to identify the request by scanning the barcode. The WILL software uses this number to link the scanned images with the stored request data and, after the scanning is done, channel the images to the delivery mode software module appropriate for the request. After the document images are delivered, the WILL station would automatically update status in DOCLINE. The manual activities in current ILL services that are eliminated by WILL are: physically distributing ILL request printouts, conducting separate activities for the different delivery modes, photocopying bound volumes for subsequent faxing, and updating DOCLINE status. Activities that are not automated: retrieving the physical documents from the stacks and scanning. WILL also allows the operator to modify the request to correct IP addresses or fax numbers if necessary.
WILL integrates the following hardware: Pentium PC, high resolution monitor, Kofax image processing boards, Ethernet board, Intel SatisFAXion board, Fujitsu 3096 scanner, laser printer, barcode scanner. The software environment is MS Windows, and the modules developed in C language are for user interface, interaction with DOCLINE, document capture, image format conversion, image compression, and communications for each delivery mode. The modules determine and set the status of each request via a database containing information about the all requests in the WILL system.
A single WILL station has been shown in the lab to deliver 200 ten-page documents a day, a limit placed by the physical retrieval of paper documents and the manual scanning of pages by an operator. To replace the entire or a large part of the ILL activity at a large library, a suitable system would consist of several such workstations networked over a LAN, rather than a monolithic mainframe-based system. The motivation for this is both economics as well as the opportunity for gradual transition where a cycle of design, development, beta testing, redesign and operational use may be put into place. The economics argument relies on the recognition that a system built with affordable components (scanner, printer, computer platform, image processing boards) will reflect the performance limitations (in terms of scanning speed, printing speed, CPU speed, etc.) of these components. Hence, a reasonable approach is to build many workstations, rather than a single system that might require a mainframe as a central controlling component.
Each workstation, however, may be functionally unique (i.e., doing one job such as interacting with DOCLINE, doing scanning, or printing or faxing, etc.) or functionally global (i.e., capable of performing all necessary functions but tailored to do any subset of these functions), which is the WILL approach. One advantage of this latter approach is that any machine could be quickly reconfigured to perform the role of another that fails, thereby increasing overall system reliability and minimizing downtime. In this networked system, one machine would serve as the Front End Processor (FEP) that logs on to DOCLINE, receives the ILL requests, and distributes them to other machines. This distribution could be based on location, for example, so that the request would be sent to the station nearest to the shelf containing the requested document. In this case, a distance measure using the call number of the requested document might be used. Distribution of the ILL requests among the machines could also be based on delivery mode in case we find that it is more effective for machines to be dedicated to either faxing, or Internet delivery, or printing.
In summary, the guiding philosophy for WILL is that we separate document delivery functions into those that can be automated and those that cannot, the only manual functions being for staff to retrieve documents from the stacks and to scan them, and that all other functions should be automated.
A self-contained WILL workstation has been in beta testing in the NIH Library since April of 1996. In this environment, WILL receives DOCLINE requests for documents to be delivered to local and remote NIH researchers as well as to other libraries. Several functions were added to the initial WILL design to assist the NIH Library in request management. These include a program to print accumulated requests in batches that are sorted by floor, delivery type and journal title, a management tool that displays the numbers of requests in each state within the WILL system, and a program to set the status of a request as filled even though it was delivered via the photocopy process rather than WILL. By setting the status as filled, the WILL operators are able to utilize WILL's automatic DOCLINE update function without having to use WILL to deliver the article. The purpose of the beta test is to (1) measure document delivery in actual operation; (2) identify the mix of delivery modes optimum for a single machine in line with the current hardware capabilities; (3) evaluate the effectiveness of the operator interface, and identify shortcomings; (4) determine what changes in current workflow could be done in light of automating ILL request retrieval from DOCLINE, automating update of ILL delivery status in DOCLINE, and eliminating the need to set up workgroups to handle documents by different delivery modes; (5) investigate the requirement to interface with request routing systems other than DOCLINE. The performance evaluation is designed to elicit information such as speed of delivery by each mode, image quality, the optimal combination of delivery modes for a single station, and the legibility and ease of use of the operator interface.
Although the WILL DOCLINE interface functions proved to be valuable, time-saving tools for the management of traditional document delivery functions at the NIH Library, WILL was used to deliver less than 1% of the requests that it processed. Of those, about one third were delivered electronically and the rest were printed for delivery by mail. The average time per article to scan and perform Quality Control was 81 seconds, which is about 20% longer than the time expected based upon laboratory studies. For those articles that were delivered electronically, those 81 seconds plus the time required to get and return the paper volume to the shelves represent the total amount of labor needed to deliver the document. Further testing is expected to address some of the other goals of the beta test, but other issues are raised, for example, the organizational factors that have to be overcome to increase utilization of a system such as WILL.
National Institutes of Health (NIH)
U.S. Dept. of Health