NOTE [From Jim Aagaard]: This document is a faithful copy of one dated October 15, 1987, with corrections to a few typographical errors and a small amount of additional material in the Organization section. It describes the status of computing in the NU Library as of Fall 1987.
As libraries have become rather painfully aware, automation does not arrive instantaneously. Reaching the implementation phase requires substantial advance planning. At Northwestern, University administrative interest was given added impetus by a librarian-faculty-student library planning committee which served as an advisory group during the design and construction phases of a new library building. In the early 1960s this group projected the development of an automated library support system and indicated the strength of its confidence by providing a special room for a dedicated library computer as the blueprints developed. This special room was not removed from the plans when the decision was made to share a computer already available elsewhere on campus.
As planning progressed, recognition that a dedicated library computer was financially impractical at that time resulted in a decision to utilize the hardware at the Administrative Data Processing department (ADP), which already maintained payroll, student records, alumni records, and accounting systems. This machine (an IBM/360 Model 30) had text processing capabilities which were superior to those of the larger computer (a CDC 6400) at the academic computing center, and the Library also was assured priority in usage.
An important early decision was to develop the system in an online mode. Unlike batch processing, the online interactive mode provides for immediate access to and modification of information in the computer file and prevents the automated system from becoming as unwieldy as the previous manual system. One unified computer record replaces a multiplicity of order and processing control slips formerly filed in separated sequences.
During 1967 a thorough analysis of all library operations was performed to determine what applications would benefit most from automation. Actual programming started in the spring of 1968, and in about a year the teleprocessing system was essentially complete and work was started on various programs for technical services functions, to be run on a daily or weekly basis. These included programs for producing catalog cards, purchase orders, and similar materials.
However, at this time the realization came that the opening of the new building was less than a year away, and the new library administration felt very strongly that it would be desirable to have an operational circulation system. Work was then suspended on the technical services part of the system, but it is important to note that the basic data management system developed up to that point eventually provided the online inquiry capability to the circulation operation.
The basic design for the circulation system was completed quickly, and while the programming continued, the massive job of keypunching nearly a million book cards was started in the summer of 1968. The system was given a brief pilot test in the Technological Institute Library in the late fall of 1969, and was in place in the new Library building in time for its opening in January of 1970.
With the circulation system in successful operation, attention returned to the development of the acquisition and cataloging capability, which was implemented in October of 1971. This was considered to be Version 2 of what eventually came to be called NOTIS (Northwestern Online Total Integrated System). Because of the unavailability of video display terminals with lower case capability at that time, the system used IBM 2740 typewriter terminals for remote input and updating.
Following the initial implementation of NOTIS there were a series of upgrades and improvements in hardware and software, both in the Library and at ADP. At ADP new Calcomp 2314-compatible (but double density) disk drives were installed to replace the IBM data cell which had served as the file storage device. The IBM/360 Model 30 was eventually replaced by an IBM/370 Model 135 and then by a Model 138, and IBM 3340 and 3344 disk drives were installed.
At the Library, IBM 3277 cathode-ray-tube terminals with lower case capability replaced the slow typewriter terminals in the summer of 1974. Initially this was done by a line-by-line emulation process, but by the middle of 1977 the programs were rewritten to take advantage of the full screen display and editing capability of the display terminals. At the same time as the display terminals were installed, the IBM CICS teleprocessing monitor was installed on the computer. A number of other academic institutions were using CICS, and it was felt that its use would facilitate transfer of the Library software elsewhere as well as adding online administrative applications for other areas of the University.
By the end of 1977 ADP had been renamed University Management Systems and had begun to take a much more aggressive role in the development of non-Library applications. This, plus an expansion in usage by the Library, resulted in a gradual degradation of response time during 1978. The obvious solution was to move to a larger computer. However, the announcement by IBM of the low-cost 4300 series of computers early in 1979 provided an alternative.
Various measurements had indicated that the load on the Model 138 was almost equally divided between the Library and other applications, and IBM indicated that the 4331 had approximately the power of the Model 138. Thus it appeared that a 4331 devoted exclusively to the Library applications would relieve the response time problem and provide both areas some growth capability. Orders were placed while further evaluations were made.
The Library was fortunate in receiving an early shipping date for a 4331 (October 1979) which meant a final decision was needed by early summer. After much discussion and consideration of many alternatives, the University administration authorized the installation of the 4331 in the Library for at least a two-year period.
The transition to the 4331 system was made over the 1979 Thanksgiving weekend, and immediately provided a substantial improvement in response time. Shortly thereafter, in February 1980, the Medical School library began to use NOTIS, and in late 1980 the Law School library embarked on a large reclassification project. (The Garrett Evangelical and Seabury Western theological seminary libraries had joined the system in May 1976; the Dental School Library started using the system in the fall of 1986.)
Hardware enhancements continued in the ensuing years. The 4331 central processing unit was upgraded to a 4331-2 in November of 1981 and to a 4361-5 in September of 1984. Nearly all of the hardware was converted from lease to purchase in November of 1982.
The original IBM 3310 and 3340 disk drives were replaced by IBM 3370 drives, with a capacity of 571 megabytes each as soon as they became available in June of 1980. Additional drives have been added since that time to accommodate growth in the database; in September of 1987 there were 10 such drives.
The terminal network exhibited a similar growth, from 60 in 1980 to 100 in 1982, 140 in 1984, and 200 in 1987.
The first step in providing user access to the Library database was the installation of a single public terminal during the 1975-76 academic year to display circulation status information. This was received enthusiastically, and more terminals were added. After a year-long analysis of alternatives to the card catalog by a Library committee, author/title access (LUIS) was added in March of 1980, and subject access was added in June of 1981. Full integration of circulation information with LUIS bibliographic displays did not occur until the implementation of a new circulation system in March of 1985, however.
The announcement by IBM of a program (developed at Yale) for the Series/l minicomputer which would enable ASCII terminals to emulate the IBM 3270 terminals used by the Library opened the way for remote access by Library users. A Series/1 was installed in April of 1981, followed by a cable link to the switching device at the Vogelback Computing Center. This allows access to LUIS from faculty offices, student dormitories, or wherever else a telephone is available. It has also facilitated the demonstration of the NOTIS system at conferences and other libraries across the country.
The limitations of the software used for public searching of the database (as well as staff searching) have been apparent since the capability was originally provided, and a formal project to enhance the indexes was started in 1981. The complexity of this change, plus factors related to the marketing of the NOTIS software, have caused many delays in this project. However, a prototype of the new index was first provided to staff in July of 1986 and is targeted for availability to the public early in 1988.
The original development activity was under the direction of John P. McGowan, then Associate University Librarian. He was joined by Velma Veneziano as systems analyst in 1967. After the feasibility study, when the project reached the implementation stage, Mr. McGowan enlisted the services of James Aagaard, a faculty member in the Electrical Engineering Department, who had also been serving as Systems Manager for the University's Vogelback Computing Center.
In January of 1973 the arrangement was formalized with the establishment of an Information Systems Development Office (ISDO) within the Library, with Professor Aagaard as Director. In the summer of 1974 a programmer was added to the staff. The department size has remained small, reaching six professional staff members briefly during development of a new circulation system.
The Library's technical Division recognized the need for coordination of the changes resulting from the application of automation, and in April 1971 Elizabeth Furlong, Head of the Search Department, was given the additional title of Coordinator of Automation Procedures tor Technical Services. This was during the period in which the ADP computer was being used and there were few terminals available, and so the Library established a Data Center in the space originally intended for a computer. The Data Center operated under the direction of Ms. Furlong, and provided some data entry services as well as scheduling of terminal use by other staff members. When the 4331 was installed in the Library, the Data Center staff assumed responsibility for the routine operation of the system. The coordinator position became Ms. Furlong's full time responsibility in March 1973.
The need for a similar position in Public Services became evident at least as early as 1981 when work on a new circulation system started, but requests to create such a position were not successful. Finally in December 1986, Ms. Furlong's position was renamed Coordinator of Library Automation Procedures, reporting directly to the University Librarian, and she was given responsibility in all areas of the Library.
In January 1981 the Data Center supervisor at that time, Diane Hanisch, transferred to ISDO as a junior programmer, and is now a senior programmer/analyst. With the two originators of NOTIS still active, the three senior members of the ISDO staff have over 50 person-years of NOTIS experience.
At the start of the 1982-83 academic year, responsibility for the operation of the Data Center was transferred to ISDO.
The demonstrated computer capability of the Northwestern University Library has been instrumental in securing funding for the University for several special projects. In late 1975 the Library was approached by representatives of the National Library of Venezuela regarding the possibility of establishing a project at Northwestern to identify and record bibliographic information about Venezuelan works held in United States libraries. This led to the creation of a 1.3 million dollar project, extending from the Fall of 1976 to the Spring of 1979. At the conclusion of the project the NOTIS computer software was installed in Venezuela, where it is being used to maintain and expand the data base which the project created.
Another important project was made possible by grants from the National Endowment for the Humanities and the Carnegie Corporation. With particular emphasis on African material, the project focused on ways in which cataloging could be done in a distributed environment. but still conform to national standards. As a byproduct of this project, the production of the Joint Acquisitions List of Africana, which has been published by the Northwestern University Library for a number of years, was converted to a computer operation. This project began in October 1977 and has continued since that time.
As other libraries became aware of the NOTIS system. inquiries began to arrive about possible replication. Aware of its very limited ability to provide adequate support, the Library initially declined to make the NOTIS software available. However, buoyed by the success of the Venezuelan installation and a positive recommendation from a consulting group from Educom (which reviewed the University's computing environment in the summer of 1980) the Library agreed to license the software to several other institutions. The first of these was the University of Florida, which was so anxious to obtain NOTIS that they offered to do the necessary conversion from the VSE operating system to MVS and return the converted programs to Northwestern. Other early users were the University of South Alabama and Washington University in St. Louis.
In an attempt to "test the waters," the Library administration agreed to fund the rental of a suite at the annual American Library Association meeting in San Francisco in July of 1981. Two ISDO staff members, with the assistance of many Library staff members from other departments, demonstrated NOTIS to over 100 visitors using the new remote access capability installed in April.
It was apparent that there was a market for the NOTIS software, and efforts began within the Library to separate the marketing and customer support functions from ISDO. After at least one false start, a librarian with extensive marketing experience with another vendor of computer systems was hired in the fall of 1983. The growth of the marketing effort since that time has been rapid; by August of 1987 the staff had grown to 50 people, outdistancing the Library's ability to provide space and computer support. Effective September 1, 1987, the organization became NOTIS Systems, Incorporated, wholly owned by Northwestern University, and located in its own rented quarters in downtown Evanston.
The growth of the marketing effort was not without repercussions on the ability of ISDO to continue development of the NOTIS system. ISDO has responsibility for the operation of the Library's computer, and the additional load contributed by a dozen new programmers, unfamiliar with the system, had a severe impact. Also, the need to answer questions and provide advice, plus the marketing-driven pressure for certain developments, seriously distorted past procedures.
In the Spring of 1980 the University joined the Research Libraries Group (RLG). This organization, which includes such prestigious institutions as Stanford. Columbia, Yale, Princeton, Dartmouth, and the Universities of Michigan and Pennsylvania, is attempting to promote cooperation among the research libraries of the United States. One of the most important aspects of RLG is the Research Libraries Information Network (RLIN), a computer-based system located at Stanford. The Stanford computer will eventually have a data base which includes the holdings of all of the members of RLG, as well as the machine-readable cataloging produced by the Library of Congress. The potential implications of this network for resource sharing and cost sharing are enormous.
It has been Northwestern's goal to work toward a direct link between the RLIN and Northwestern computers, which can bring access to the RLIN data base to anyone who has access to the Northwestern system. The RLIN connection, however, cannot be expected to eliminate the need for a local computer at Northwestern, since most of the requests will be for items held by Northwestern, and the communication costs to Stanford, as well as the additional computer hardware needed by RLIN to handle this large volume of transactions on the central computer, would far exceed the cost of the local computer.
The general need for such linking has been long recognized by many persons in the library community. In 1975 the National Commission on Libraries and Information Science prepared a report which described the goals and objectives of a national library network. One result of that report was the establishment of a Network Development Office at the Library of Congress. There have been a number of other groups which have addressed the policy and technical issues of networking since that time, many of them organized by the Network Development Office and funded by the Council on Library Resources.
The Northwestern Library was a participant in one of these groups, the Network Technical Architecture Group (NTAG), in the 1977-78 period; this group attempted to develop the technical specifications for a library network. This effort did not lead to the funding necessary to develop the network, but in 1980 a new project was established. The Linked Systems Project initially involved the Library of Congress, RLG, and the Washington (now Western) Library Network, joined in 1984 by the Online Computer Library Network (OCLC). This group had the advantage of being able to take advantage of the rapidly emerging international standards for computer networks, and unlike the original NTAG project, its efforts have resulted in operational links between several of the participants.
ISDO participated in the early stages of the Linked Systems Project with a grant from the Council on Library Resources to develop the application protocols for the network. Staff members from ISDO have been members of the project's Technical Committee since July of 1985 and the Bibliographic Analysis Committee since February of 1987. ISDO is actively developing the software which will allow NOTIS systems to participate in intersystem communications.
Many other libraries, including NOTIS licensees, have been members of one of the two large library networks (RLIN and OCLC). This participation has involved a number of local terminals which are used to access a large central database. However these libraries realize that this approach is not satisfactory for providing patron access to their collections, and see a local computer plus a link to RLIN or OCLC as a means to provide high-volume access to a local database as well as occasional access to a much larger database when it is needed.
The present 4361 configuration supports approximately 200 terminals in the main, Science-Engineering, Deering, Garrett, Medical, Dental, and Law libraries. Nine dial-in lines are also available The NOTIS-4 software provides support for a wide range of library functions, beginning with pre-order search, and continuing through the generation of purchase orders, recording receipt of materials (monographs and serials), production of temporary catalog cards, claim notices, book marking materials, permanent catalog cards, and circulation control.
During the normal academic year the computer system operates 16 hours a day, 6 days a week, and 12 hours on Sunday.
The main bibliographic file now contains over 900,000 records, and additional smaller files exist for the Joint Acquisitions List of Africana, transportation literature, and Africana conference papers. The bibliographic file is growing at the rate of about 75,000 items per year.
Operational cost for NOTIS has been a major consideration from the beginning of design. Nearly all programs are written in assembly language; this was originally forced by the limitations of the available hardware and by the lack of adequate text processing capabilities in other languages. The result has been a system which is much more economical in usage of hardware resources for equivalent function than other administrative systems in operation or under development.
The Library budget for the 1987-1988 year for computer hardware and software is approximately ,000. However, this does not provide adequately for the eventual replacement of existing equipment; an additional ,000 per year is needed to provide this and allow for modest growth.