It is now just over a decade since CLSI launched the first turnkey automated library system. In the intervening period, library experiences with automation have led to at least two conclusions that we believe can be generalized to all libraries considering automation, irrespective of their size. One is that cost effective library automation is most likely to be achieved when a library mounts an integrated system capable of supporting the automation of multiple functions. The other is the importance of ensuring ongoing support for the chosen system. For most libraries, this has meant the selection of either a turnkey system or a supported software package.
Despite the attention given in recent issues of LSN to the emergence of micro-based turnkey systems for integrated library automation, the price of these systems-upwards of $50,000 for those offered by Avatar and CTI and even more for the CLSI 11/23-based system-still place them out of reach for many small libraries. Nevertheless, these smaller libraries, aware of the increasing trend towards automation, are anxious to be part of it. In the following discussion, the editors will outline some of the current options available to them.
For the smaller library, the alternatives to an integrated multifunction automated library system are still relatively limited. A number of companies offer micro-based software to automate a single function such as circulation or serial control. For some libraries these packages can provide a solution to a pressing problem. The risk is that later, when additional functions are to be automated, the library either will have to abandon this original single-function system in favor of a new integrated system or be forced to continue down the non-integrated path for the automation of other functions.
Another approach is provided by the distributed systems such as Gaylord's automated circulation system, Baker and Taylor's and Brodart's automated acquisitions systems, and the online serial systems currently offered by various subscription agencies. Again, careful comparisons of service, ongoing support, and costs are essential.
Each of these approaches offers possibilities for the smaller library faced with the need to solve an immediate and pressing problem. But, they do not necessarily point the way in which a smaller library, with the luxury of time in which to consider a long-term automation plan, should be moving.
Another lesson gleaned from recent history is the importance of developing a high quality data base of full length bibliographic records that comply with national standards. In an automated system, the hardware is periodically replaced and the software is subject to continual maintenance or eventual replacement. The one clement which has an extensive life span is the data base, and its longevity is a function of its original design. Data bases of records developed according to national standards-AACR, LC, or the Dewey Decimal Classification, and following Library of Congress subject headings, encoded in the MARC format, and retained in full-length rather than abbreviated form--have-a better survival rate than those which do not meet these criteria.
A data base of bibliographic records is basic to all forms of library automation and the capture of machine-readable data at the time of cataloging is more economical than retrospective conversion at some later date. Therefore, libraries with the possibility of automation in the future should examine procedures and determine whether there is some way in which machine-readable records can be generated as a by-product of current operations. For example, if the services of a bibliographic utility are used for cataloging support, steps should be taken to ensure that in addition to ordering cards output, archival tapes of the records are also specified. Provided that steps are taken to check the tapes for deterioration every two years or so, and to refresh them as retired, these tapes can be retained indefinitely until such time as they can be utilized as input to an automated system. Libraries using other forms of cataloging support-a commercial cataloging service, or obtaining cards from a book jobber-should also investigate the possibility of obtaining machine- readable records as an additional output from the vendor's cataloging process.
If the library which has obtained such records subsequently decides to utilize one of the more limited approaches to library automation-an approach using a system which does not have the capacity to handle full length or MARC records-the machine-readable archival tapes can be used to generate brief records for that system with the full tapes being retained intact for future applications.
Another alternative that in some situations might provide an attractive interim solution is the use of one or more of the automated subsystems from a bibliographic utility. A variety of approaches is possible. Because of its widespread penetration of the library market OCLC is used as the example in this article. However, similar approaches are possible in various degrees, through the other utilities-- RLIN, WLN, and UTLAS.
It is relatively inexpensive for a library to use OCLC as a cataloging support system. Libraries with very small cataloging loads can cluster with other libraries and share the use of an OCLC account number and terminal. While prices for OCLC services vary according to the charges imposed by the regional networks, cataloging support is generally available for around $2.00 a title with cards priced at around $.04 each and machine-readable records at $.05 per record. The drawback to this approach is that holdings for the libraries recorded on the OCLC data base will only be definitive to the level of the account number, an exact representation of the resources of individual libraries in the cluster is not possible. While a smaller library may not have a sufficiently large cataloging workload to justify individual membership in OCLC, the use of one or more of OCLC's other subsystems might increase the work level sufficiently to then justify membership and the investment in a terminal to be used for dial access to the utility.
Currently available subsystems of interest to smaller libraries include: acquisitions, at a cost of approximately $1.50 per title; serials check-in, an activity for which a complex pricing structure is applied but which one library has estimated costs $3.50 per year per title; and use of the interlibrary loan subsystem in which lookup and referral to five holding locations is estimated to cost in the region of $1.20 per item.
Experience suggests that libraries with at least 2,500 transactions a year, of any kind, can realize benefits from use of these OCLC services, and that dial-up access is appropriate for up to 6,000 transactions per annum. At transaction levels above 6,000, use of the utility's dedicated lines is usually advisable. Ignoring location/workflow considerations, it has been found that a single OCLC terminal can handle up to 15,000 transactions a year.
The initial equipment investment costs incurred by a library beginning to use OCLC will vary according to the equipment that it already has. As detailed in items in LSN Vol. II, No. 4 and Vol. II, No. 12, a library requiring dial-up access may utilize an existing terminal, word processing equipment or microcomputer to link with the utility. Depending on the configuration of the existing equipment, additional investment in a modem, smart terminal package or communications board may be involved.
For libraries with no existing hardware, it would be difficult not to recommend the purchase of an OCLC configured IBM Personal Computer as a terminal-a unit which will become available in 1984. There are a number of reasons for this recommendation. Not only can such a unit be used to support more than library applications-including automation of office and management functions-and downloading from remote data bases, but it contains special features to facilitate library applications. These include support of the full ALA character set and capabilities which will allow the terminal to be used for dial-up access to the OCLC telecommunications network-an approach currently available only to those OCLC members who have dedicated OCLC terminals. The announcement that WLN has also chosen the IBM PC as a terminal and rumors that RLIN is likely to make the same decision strengthen the wisdom of this approach. A potential problem in relation to the OCLC configured IBM PC is availability. We are already receiving reports from libraries having difficulty in obtaining regular IBM PCs; it does not appear unreasonable to predict that the demand for the library version of the IBM PC may also outstrip the supply.