Diane R. Tebbetts

University of New Hampshire Library
Durham, NH 03824-3592, USA


Keywords: Automation, Library Automation, Purchasing, Library Systems, Integrated Library System, Turkey Systems, Change, Planning, RFP, Request for Proposal, Decision-making, Evaluation, Statistics, Vendors, Matrix Analysis.

Abstract: This paper examines the decision-making that is essential before a library purchases an automated system. It is vital to understand the library's circumstances. Since a thorough analysis of the library's current operations is the basis for determining its needs, the types of statistics that should be gathered and how these will help to deter-mine the type of system required will be examined. A discussion of turnkey systems, software solutions, and proprietary versus open operating systems will explore the advantages and disadvantages of each. Also, there will be an examination of the mo-dules available for purchase and what should be considered in determining which to buy. Whether to issue an RFP and to hire a consultant are some of the other topics that will be discussed.


Purchasing an automated integrated system for your library is a major decision and one that has to be considered carefully. Whether one is purchasing the library's first, second, or third sys-tem, it is essential to think of automation as a process and not as a one-time event. Experience has taught us that library systems do not last a lifetime. In fact, it is clear that automated systems have a relatively short life span. One should think in terms of five years for the system's life expec-tancy. Then, it may be necessary to upgrade, migrate, or add new modules. Planning in this environment is quite different from planning a one-time purchase. Always, one must be thinking about adaptability and flexibility. The ability to change will be crucial to the success of the library's automation efforts.

Change means not only buying newer, faster hardware or acquiring software that is easier and more flexible, but also attempting new functions, accessing different systems, and telecom-puting to remote sites. Keeping in mind that it will be essential to do new things in new ways will make the future transition much easier.

In shopping for an integrated system, it involves several interweaved areas as shown in Figure 1.


The first admonition in any planning process, but especially in the automation decision-making effort, is to know and understand one's own situation. Before beginning, it is essential to gather as much information as possible about the library's current needs and circumstances -- internal and external.


Current statistics are essential. If this is the library's first automation effort, bring together all available statistics describing the current status of the library. This includes, number of titles, number of periodical subscriptions, number of patrons, number of circulation transactions, and number of interlibrary loan transactions. If this is a second or later system, these figures translate into number of records--bibliographic, patron, or circulation. Statistics are crucial because they determine the size of your system which is one of the first questions to be answered. A decision on size will narrow the choices of hardware and vendors.


An understanding of the history of the library's operating procedures is necessary to deter-mine what needs to be done to prepare for automation. If this is the library's first system, a dis-cussion of the cataloging procedures is essential. If the library has been producing cataloging cards through the services of a bibliographic utility, such as OCLC or RLIN, and producing MARC tapes, it will be necessary to contact the vendor and make arrangements for the tapes to be produced, processed, and massaged. This will be negotiated with the bibliographic utility and the system vendor. If the cataloging is manual, arrangements for retrospective conversion of the records to machine-readable form will be necessary. If this is a second system, an examination of the first system and the library's database will be a vital part of the migration process.

A precise recording of the procedures of the library is necessary to determine how the manual or machine readable records will transfer to the new system. If the library has a clear understanding of its past cataloging practices, the automation process will be easier. If not, it will be necessary to reconstruct this past history. From this point on, be sure to keep good notes on all decisions and operating procedures.


The operating environment of the library will be a major consideration in determining the appropriate automated system. Combined with the size will be the question of which functions need to be automated. Is acquisitions vital, is an online public access catalog the main focus, or is circulation the primary consideration? The answers to these questions will help determine what subsystems will be required. Perhaps, the library will decide it wants to automate all functions and provide a fully automated integrated system for its users. It will be essential to make these deci-sions before the purchasing process begins.

Another important aspect of the library's environment is its external relationships. The library's parent institution and its requirements will be crucial in determining the appropriate hardware, software, and networking configuration. If the library is part of an academic institution, is the campus networked, will remote access be required, with what equipment will the library's computer need to interface are essential questions. Does the library participate in any consortial relationships, will it need to access the high speed networks, such as Internet, will it be necessary to connect to other databases. These questions will need to be answered to help determine the system required.


After the current status has been determined, it is essential to determine the library's future needs. Clearly, a projection of annual acquisition rates will help to predict growth rates. It will also be necessary to consider the growth possibilities for subscriptions. Also of major importance will be the determination of numbers of future users and their requirements. Consultation with the institution's administration will help to determine its future direction. For instance, does the college or university plan to increase enrollments or will it be downsizing its operation, does a municipality expect to see an increased number of residents or is a decline predicted.

Another question concerns the life expectancy of the proposed system. As stated earlier, the normal life span is approximately five years. Will it be necessary for the library to maintain a state-of-the-art system? According to Mark Hinnebusch "This is intimately tied up with what you project the future of your library to be, and also your perception of the future of the library system marketplace. How demanding of technological innovation are your patrons?" (Hinnebusch, 1991) Equally important in determining the life span of your system is determining the economic environment. Hinnebusch states: "The purchasing strategies run from buying the newest and the best as soon as it becomes available to planting a system in your library and using it until parts availability, or rather the lack of availability, forces replacement" (Hinnebusch, 1991).

Projecting the answers to these questions will help to determine which systems will be appropriate for the library's current and future needs. Planning carefully at this stage will save much time and worry later.



Once the library's statistics have been gathered and its needs determined, the question of the size of the system becomes more manageable. Although there are all kinds of configurations of hardware and software, determining whether the system should be supported by microcomputer, minicomputer or mainframe applications will help to narrow down the options. Certainly the size of the library -- the number of records, the number of users, and the number of annual acquisitions -- will help to determine the appropriate hardware. Once this has been decided, the list of possible vendors can be narrowed down to those supplying systems with the correct size and configuration.


In addition to size, the type of system will be determined by the library's environment. If the library is part of a consortium or larger system, it may be necessary to connect or be an integral part of that system. If so, the question of whether to set up a distributed system or a centralized system may be extremely important. If telecommunication costs are a major factor, it may be more cost effective to set up a distributed system in which each library maintains its own system but is interfaced with the other libraries. A centralized system will probably be necessary for a main library and its branches.

The system's architecture is now becoming a major consideration. The latest developments suggest that client-server architecture will be the direction of the future. "Client/server architec-ture involves the physical separation of the user interface and its search commands from the mechanisms that extract the information from the database" (Bridge, 1993). In this type of confi-guration, processing is downloaded to workstations where much of the actual work is performed. This means that the central hardware does not need to be as large as in systems where all the processing is performed on a central CPU. However, it does mean that the peripherals must be microcomputers rather than terminals. There is, of course, a cost factor involved in this type of system. If the library is embarking on an automation project for the first time, it may be appro-priate to consider seriously the client-server systems.

Proprietary or Open

Whether to purchase a system with a proprietary operating system or an open system, pro-bably UNIX, is a burning question in the automation field at this time. Of course, purchasing a proprietary operating system limits the hardware setup, whereas, purchasing a UNIX-based system allows a greater degree of flexibility. Although a study of the UNIX operating system reveals that there are many versions currently available and that transferability may be difficult (Bridge, 1993). Nevertheless, planning for the future requires flexibility and an operating system which will run on various hardware platforms provides for a certain amount of maneuverability. It is essential to discuss these issues with the vendors. If a vendor offers a UNIX-based operating system, be sure to find out precisely what this means. Does it run on different types of hardware or is it a specific version of UNIX. Again, it is important to avoid, as much as possible, limitations incurred by non-standard adaptations of any part of the system.

Turnkey or Software Only

The next major decision is whether to buy a turnkey system in which the hardware and software are bundled or to purchase a software only system. There are advantages and disadvan-tages to both. If the library has very little experience with automation and computer expertise is limited, it may be safer to purchase a turnkey system in which the vendor supplies hardware, software, and accompanying support. When a library purchases software only there is always the problem of troubleshooting the system. In situations like this, the question of software versus hardware problems can become troublesome. Unless the library has local computer expertise the "software only" solution may be problematic.

The vendor may increase the software licensing fees if the library chooses to purchase it separately from the hardware. It is important to consider all of these factors before a decision is made in this regard. However, a library with good technical support, available hardware, and financial considerations may choose to mount and maintain the software locally.


Crucial to the success of the automation effort is the reliability, viability, and stability of the vendor. It is essential to find out as much as possible about the vendor to determine the prospects for future support and development of the system.


Clearly, it is necessary to require statistical documentation about the vendor. Some of this information is readily available in published sources such as Library Journal's annual survey of the "Automation Marketplace" by Frank Bridge. This covers such statistics as number and size of installations, estimated revenues, market share, and comparisons based on current data.

In addition to unbiased evaluations and comparative data from the literature, it is necessary to obtain as much information as possible directly from the vendor. Annual reports, financial statements, and company publications and brochures should be forthcoming upon request. Certainly, this is part of any formal RFP (Request for Proposal) process. Number, size, location, and date of installations should be readily available. These data should help to determine the vendor's current status in the library system market.


A study of the vendor's experience in the library automation field will reveal its length of time in the field, types and sizes of libraries in which there are installations, and the type of system the vendor specializes in. For instance, is the system fully integrated or are there specific modules on which the vendor places emphasis. A survey of the literature over a few years will reveal the role of the vendor has played in the development of library automation. It will indicate whether the vendor has kept up with research and development so that the system has the latest features or whether the system is becoming outdated and outmoded. News items in the latest library and computer journals will provide background on the history of the company, its stability and financial outlook. All of this information is important in trying to determine the reliability and stability of the vendor.


One of the best ways to determine the reliability and support of the vendor and the perfor-mance of the system is to survey current library users. Genaway writes: "Clearly users of library systems are in the best position to evaluate their systems. They are more familiar with the pro-blems, nuances, and peculiarities of the system they are using than any third party 'objective' evaluator. They know how well the system works and how reliable it is" (Genaway, 1989). It is important to have done one's homework and to have specific questions to ask. Genaway goes on to list some of the questions that might be asked such as How reliable is the system? and How good is service? These are the types of questions that only experienced users of the system could answer. Along with this type of survey, it is useful to conduct onsite visits to see the operation of the system in real-life situations. Vendor demonstrations are useful, but onsite visits present the system in actual operation. Of course, it is important to pick sites as similar to one's own library as possible. Again, it is essential to be prepared for these visits. If possible, people who will be directly involved with the operation of the system should be included in the visit. As Genaway says "Have reference librarians talk with reference librarians and catalog librarians with catalog librarians at installed sites" (Genaway, 1989).


Obviously the purpose of collecting as much data as possible is to determine not only the current performance of the system and the vendor but also to predict the future reliability of the system and the support of the vendor. A vendor with a strong research and development program is more likely to provide updates to the system and keep up with the latest technological develop-ments. A company with strong financial credentials is more likely to keep up with the heavy competition in the field and to provide ongoing support for the system. A system with several installed sites is more likely to have been "debugged" whereas, a new system is more likely to have "glitches" in the program. However, a new system may have more capabilities and provide new approaches. It is important to weigh the pros and cons carefully in a consideration of the relative merits of the systems.


Once the library has analyzed its own situation and needs, researched the vendors and the systems, surveyed current users and visited operational sites, it is necessary to evaluate the systems to determine the best possibilities for its own situation.

Matrix Analysis

One way to systematically analyze the systems in terms of the library's requirements is to develop a matrix based on the needs and environmental concerns of the library as developed by the information gathering strategies listed above. Figure 2 illustrates a sample matrix. The columns represent the criteria established by the Library such as size considerations (number of records, number of terminals), hardware requirements, networking needs, and gateway capabi-lities; the rows represent the vendors. By filling in each cell one can determine how the vendor rates in that category. This will indicate quickly the vendors that do and do not meet the Library's basic criteria. This is a method for narrowing the possible vendors to a few that will require more in-depth scrutiny.

RFP (Request for Proposal)

The decision whether or not to issue a formal RFP to the major vendors under consideration has to be weighed in the context of the Library's particular situation. The RFP process is lengthy, detailed, and costly both for the Library and the vendor. At the same time, it is thorough and rigo-rous. Very often the RFP process is demanded by the parent institution or governmental authority.

The RFP process requires the thoughtful analysis of the Library's needs and operations. Likewise, it requires the vendor to supply specific responses to the Library's specifications. It provides the mechanism to evaluate the vendors' systems based on established criteria.

Clearly, the RFP provides a mechanism for the careful evaluation of the systems based on the specific needs of the Library. If time, money, and staff are sufficient, the RFP process will ensure the rigorous analysis the purchase of an integrated automated system deserves.


Another major consideration is whether or not to hire an automation consultant. The answer to this question has to be determined by the Library's particular situation. If the Library is migrat-ing to a second- or third-generation system, has internal or local expertise in computer-related issues, has carefully analyzed needs, and has projected automation goals and objectives, then the services of a consultant may not be required.

On the other hand, if the library is involved in its first automation project, is unsure of the process, needs assistance in defining needs and specifications, then the consultant may be very helpful. Consultants may be hired for very specific purposes, such as performing the library's needs analysis, preparing the RFP, or advising on contract negotiations. There are many excellent

consultants who will tailor their services to meet specific needs. Therefore, the decision to hire a consultant has to be based on the specific needs and environment of the library.


The decision to purchase a new automated integrated system is certainly an important deci-sion for the library. It involves a large sum of money, a heavy commitment of staff time and effort, and the organization and access of the library's resources. But more than that, it will to a large extent determine the services the Library can offer, the appearance the Library will present to its community whether that is made up of faculty and students or townspeople and government officials, and the possibilities for providing users with the latest in information and a wide variety of resources. Therefore, it is extremely important to analyze carefully the needs of the library, to determine logically the specifications required, to review systematically the available systems, to scrutinize thoroughly potential vendors' viability and stability, and to project future requirements to the extent possible. Finally, it is important to recognize the fast-paced nature of technological development and the necessity to plan for change. Flexibility and adaptability will be crucial to the long-term success of the library's automation program.


Bridge, Frank R., "Automated System Marketplace 1993," Library Journal, pp. 52-64 (April 1, 1993).

Genaway, David C., "Micros, Minis, and Mainframes," Technicalities, 9 (10): 4-7 (October 1989).

Hinnebusch, Mark, "Integrated Library Systems," Academic and Library Computing, 8 (5): 4-6 (May 1991).