Introduction to Automation for Librarians, William Saffady

William Saffady


Introduction to
Automation for Librarians
Fourth edition

AMERICAN LIBRARY ASSOCIATION
Chicago and London
1999

Integrated Library Systems

Broadly defined, an integrated library system-variously termed an "inte¬grated online library system" (IOLS) or simply an "integrated system" -is a computer-based information system that uses a single bibliographic database and a set of interrelated application programs to automate multiple library appli¬cations. That straightforward definition belies the product group's diversity. As described below, integrated library systems are marketed by dozens of vendors. They may operate on mainframes, minicomputers, or microcomputers. They may be implemented by smal1libraries or large multi-library consortia. They may be purchased as preconfigured combinations of hardware and software or as prewritten programs for use with separately purchased computing equipment.

The earliest integrated library systems date from the late 1970s. Prior to that time, library automation was dominated by single-purpose computer systems intended for specific applications, particularly circulation control, which was one of the first library operations to be successfully computerized. During the late 1970s and early 1980s, vendors of circulation control systems began offer¬ing online public access catalogs (OPAC) as widely publicized enhancements. The resulting products were often described as library management systems to reflect their expanded functionality, although their OPAC implementations were primitive by today's standards and libraries were slow to adopt them. As late as 1985, for example, only about one-third of libraries that utilized CLSI cir¬culation control systems, then the most widely installed products of their type, had implemented that vendor's public access catalog module: Similarly, Data Research Associates implemented its first integrated system at Cleveland Pub¬lic Library in 1978, but it made only three additional installations in the next five years. Product availability and customer acceptance increased in the mid¬1980s, however. By the end of that decade, there was little interest in single purpose library automation systems.

Most integrated systems are modular in design. Vendors offer suites of inter¬related programs that automate specific library operations. Core functionality is provided by certain application modules that are required in all installations; optional modules, which automate specific operations, can usually be added at any time. A minimal, but fully functional, implementation typically incorporates: three application modules: cataloging, online catalog access, and circulation control. Of these, the cataloging module-which supports data entry, database management capabilities, and, in some cases, authority control-is required. Online; public catalog access is often the principal motive for implementing an integrated system; Consequently, OPAC modules are widely installed. Circulation control, while not essential, may be inextricably integrated with cataloging and online catalog access as a standard product configuration.

Acquisitions and serials management are popular but usually optional ad¬ditions to a basic system configuration. They may be implemented during the initial installation of an integrated product or added at a later time. Other appli¬cation modules, which are selectively available and almost invariably optional, include media booking, a community bulletin board, community information and referral, homebound access, reference databases, Internet access, electronic imaging, full-text retrieval, and electronic mail capabilities. Reserve room capa¬bility may be implemented as a separate module or incorporated into circulation control. Some vendors also offer a report generator that supplements the inte¬gral output capabilities of the various application modules by producing lists, notices, and other printed documents in a variety of customized formats.

The following discussion surveys important capabilities and features associ¬ated with specific integrated system components. Integrated library systems are offered by several dozen vendors. Examples include Ameritech Library Services, Best-Seller Incorporated, Brodart Automation, CARL Corporation, CASPR Library Systems, Comstow Information Services, Contec Data Systems, Data Re¬search Associates, Endeavor Information Systems, EOS International, Ex Libris Limited, Extended Library Access Solutions, Follett Software Company, Fretwell¬Downing Informatics, Gaylord Information Systems, Geac Computers, Gores Technology Group, Innovative Interfaces, International Library Systems, MAX¬CESS Library Systems, Nichols Advanced Technologies, SIRS Incorporated SIRSI Corporation, The Library Corporation, Winnebago Software Company, and VTLS. Some of these vendors offer multiple product lines.

In library publications, integrated systems are often divided into two broad groups: (1) minicomputer- and mainframe base product, which are princi¬pally installed in medium-size and larger academic and public libraries, and (2) microcomputer-based products, which are intended for smaller public, community college, school, corporate, and government libraries. That distinction, while once meaningful, is increasingly artificial. As technology narrows the performance differences between traditional categories of computing devices, several vendors of minicomputer-based integrated systems have introduced microcomputer implementations with comparable functionality.

While integrated library systems are available for many types of comput¬ers, the newest product configurations are intended for Unix, Windows, or Windows NT installations. Mainframe-based integrated systems,' which were popular among large libraries in the 1980s, are now out of favor. Throughout the 1990s, older integrated systems that operated on proprietary minicomputers, such as VAX processors from Digital Equipment Corporation, were steadily replaced by Unix-based products, which offer excellent price/performance characteristics and expandability. At the time of this writing, the newest inte¬grated systems were designed for the Windows NT operating system and client/ server implementations in which computing tasks are distributed among a shared computer (a server) and microcomputer-based workstations (the clients). Specific computer configurations aside, integrated systems may be sold as prewritten software packages for implementation on library-supplied computers or as turnkey combinations of reselected computer hardware and prewritten software.

ONLINE CATALOGS

All integrated systems support cataloging as a core application module that permits the creation, updating, and management of a library's bibliographic database. Certain cataloging capabilities are so widely implemented that they provide little basis for product differentiation, although system-specific varia¬tions can affect the convenience with which particular tasks are performed. All integrated systems, for example, support MARC records for those types of li¬brary materials for which USMARC formats have been developed. Some sys¬tems also support other national implementations of the MARC format, such as CANMARC, UKMARC, UNIMARC, or AUSMARC. At least one integrated li¬brary system supports the COSATI/CENDI format developed by the U.S. De¬partment of Defense for technical reports and similar documentation. Most integrated systems permit partial MARC records that employ the USMARC format with some fields left blank. Some systems also support non-MARC for¬mats with library-defined fields. Typically MARC and non-MARC records can coexist within a library s database.

All integrated library systems permit key-entry of original cataloging data as well as the transfer of cataloging records from machine-readable sources. Original cataloging is typically facilitated by displayed work forms that corre¬spond to USMARC formats for specific types of library materials. Depending on the system or customer preferences, newly entered cataloging records may update a library's database immediately or at scheduled intervals. Most inte¬grated library systems can import machine-readable cataloging records from MARC-format tapes produced by national libraries, bibliographic utilities, retrospective conversion services, book jobbers, or other sources. Vendors typically provide a list of source tapes that have been successfully loaded into their sys¬tems. Microcomputer-based products may transfer machine-readable records from diskettes rather than magnetic tapes. Many integrated systems also do electronic transfer of machine-readable cataloging records from bibliographic utilities, CD-ROM information products, or other sources.

Depending on the system, authority control may be incorporated into the cataloging module or offered as a separately purchased component. In either case, authority control establishes and maintains authorized forms and cross references for designated field values. As bibliographic records are entered, field values subject to authority control are automatically checked against authorized forms, with new or questionable names and headings being flagged" for review. System-specific variations determine the fields to which authority control can be applied. Authors' names, uniform titles, series titles, and subject headings are the most commonly controlled fields. Depending on the system, authority records may be key-entered, generated from existing headings in a li¬brary's catalog, or imported in machine-readable form from external sources. Global editing capabilities facilitate the modification of headings. Some inte¬grated library systems permit multiple authority files. Some integrated systems support thesaurus-like cross-reference structures that include broader terms, narrower terms, related terms, and scope notes, in addition to the familiar "see" and "see also" entries.

THE PROBLEM OF CARD CATALOGS

As discussed above, the introduction of effective online public access catalog (OPAC) modules in the late 1970s and early 1980s signaled the transition from single-purpose circulation control systems to true integrated library systems that support online retrieval of bibliographic records in addition to mere trans¬action processing. For many libraries, card catalog replacement is the principal motive for implementing an integrated system.

Librarians have long recognized the many problems associated with the production, maintenance, and use of card catalogs. Writing in the 1940s, Fre¬mont Rider pointed out the substantial amounts of space consumed by the card catalogs of research libraries. Since that time, high construction costs and greatly diminished capital funds for new library buildings have made the efficient use of available floor space an important concern for libraries of all sizes and types. Assuming al'1aVerage of 20 square feet per card catalog cabinet and its required adjacent aisle and work space, the card catalog for a collection of 100,000 titles will occupy approximately 280 square feet of contiguous floor space, a portion of which might be put to other uses were the catalog converted to a more compact format. At a rate of 10,000 new titles per year, an additional 30 square feet will be required to accommodate annual catalog growth. If the collection is not weeded, the size of the catalog will double in ten years.

In addition to occupying valuable floor space, card catalog cabinets are ex¬pensive. A frequently cited 1969 study by R&D Consultants found that a typical sixty-drawer cabinet contained about 43,000 catalog cards when filled to 60 ., percent of capacity. At the time of this writing, such a cabinet cost several thou¬sand dollars. Assuming an average of six cards per title, the catalog of a library with a monograph collection of 100,000 titles would occupy more than $30,000 worth of cabinets. Assuming a new acquisition rate of 10,000 titles per year, over $6,000 would have to be spent every two years for new cabinets. Addi¬tional expenditures and floor space will be required for tables or other furnish¬ings that are customarily provided in the card catalog area.

For most libraries, the labor required to create and maintain card catalogs will prove more costly than the floor space and cabinets that such catalogs oc¬cupy. Catalog cards must be printed internally or purchased in preprinted sets.

New cards must be sorted and interfiled, while existing cards must be removed, modified, and refilled to reflect changes in the library's collection. Additional labor is required to create and maintain guide cards, cross-reference cards, and authority files. Of these tasks, only the production of catalog card sets has been directly affected by the automated systems and services described in chapter 5. Computer-produced cards can be ordered from bibliographic utilities, printed lo¬cally by CD-ROM cataloging systems, or purchased from book jobbers or other suppliers. These automated systems and services do nothing, however, for other aspects of catalog maintenance. While cards may be printed automatically- they must still be filed manually, a potentially time-consuming task. At six cards per set, for example, the annual acquisition of 10,000 new titles will generate 60,000 catalog cards. At a sustained filing rate of one card per minute, 1,000 hours of labor, or one-half of a person-year, will be required. In many libraries, this expen¬diture is increased by the additional labor required to check newly filed cards for correct placement before adding them to the catalog. In some cases, the lack of sufficient labor for catalog maintenance results in considerable delays in filing of newly produced bibliographic records. Consequently, some cards never enter the catalog. As an example, libraries that acquire micropublished collections consist¬ing of multiple titles are often unable to allocate the labor required to file the printed catalog cards that may accompany such collections.

Similar substantial costs are incurred to remove previously filed cards as collections are weeded, location information is added to or deleted from union catalogs, or existing file headings are modified to conform to revised cataloging practices. The cost of such heading modifications gained considerable promi¬nence when the Anglo-American Cataloguing Rules were revised. A study by the Library of Congress indicated that full implementation of the provisions of AACR2 would require changes in at least 37 percent of the form of headings then in use by LC, and changes would require modification of at least 49 per¬cent of LC catalog cards. High as they are, such percentages do not reflect the impact of revised rules for corporate authorship that necessitate additional card modifications to reflect changes in the choice of headings. Presumably, compa¬rable modifications are required by other libraries, each modification involving the removal and refilling of the affected cards. As a further source of potential expenditures to those libraries desiring to make their cataloging practices con¬form to those of the Library of Congress, LC closed its card catalog in favor of a computer-based system in 1981. As several librarians have pointed out, the Li¬brary of Congress was previously constrained in changing headings by the inconvenience inherent in having to manually modify its own card catalog. Such constraints no longer apply.

The cost of file maintenance aside, card catalogs offer only limited perfor¬mance and convenience, both for library users and for libraries themselves. Ex¬cept where telephone inquiries are accepted, a user must visit a library to consult its card catalog, even though time might potentially be saved if users could determine in advance whether a given item is in the library's holdings. Ideally, an academic library's catalog should be readily accessible to faculty members in individual departments; a technical library's catalog should be available in user work areas; and a public library's catalog might be' consulted in schools, stores, a.'1d even homes. Unfortunately, the cost of card production, floor space, cabinets, and file maintenance discussed above prevent the replication of all or portions of conventional card catalogs. In multi-branch libraries' and library systems, the necessity of maintaining a union catalog in a system headquarters or other centralized location is especially constrained. In such.cases, library users are denied convenient access to information about system¬wide holdings. In addition, the time and cost of resource sharing transactions may be significantly increased by the need to route all interlibrary loan requests to the central union catalog where library locations can alone be determined. If the union catalog were conveniently accessible to individual libraries, the hold¬ing libraries for desired items could be determined and contacted directly by the borrowing institutions.

Additional limitations make card catalogs inconvenient for library users. As an example, individual cards cannot be removed from the catalog for photo¬copying. Instead, the user must take notes, usually while standing at a table. Although many libraries consider the card catalog an important local biblio¬graphic resource rather than a mere location device, they rarely provide a com¬fortable work setting for its prolonged use. Critics of card catalogs further point out the limited number of access points they provide; they note, for example, that users cannot retrieve books by title keywords or conveniently perform searches that involve multiple headings. While there is no intellectual impedi¬ment to the production of additional catalog cards for title keywords or other retrieval parameters, such a practice would significantly increase the cost of card production and file maintenance. It would also increase the size of the cat¬alog, compounding the space consumption problems and cabinet costs noted above. Expense aside, the logical coordination of multiple headings cannot be conveniently implemented with card-form catalogs.

Among their other problems, card catalogs are vulnerable to intentional or inadvertent damage or destruction. Individual cards are subject to wear and tear in daily use, and the entire card catalog or substantial portions of it can be destroyed by fire or vandalism. Because the card catalog is a unique and vital operating record, which if damaged or destroyed would seriously impair a li¬brary's ability to function; a protective microfilming program is typically advis¬able to create a backup copy for offsite storage. In addition, older catalog cards that are in poor condition must be periodically replaced, thus further increasing maintenance costs.

Retrospective Conversion

Presumably, some or all of the problems discussed in the preceding section can be minimized or eliminated by replacing card catalogs with online catalogs. The implementation of such catalogs requires, however, the conversion of a li¬brary's cataloging records to machine-readable, computer-processible form. Such conversion can be accomplished relatively easily for cataloging records pertaining to a library's recent or ongoing acquisitions. As discussed in chapter 5, libraries that obtained cataloging services from bibliographic utilities can order archival magnetic tapes that contain machine-readable records for the items they have cataloged since becoming utility customers. Depending on the particular bibliographic utility and library requirements, such archival tapes may be available on a monthly,• weekly, or "more” frequent basis. Alternatively, some bibliographic utilities support electronic transfer (downloading) of cata¬loging records, either individually or in batches, into a library's local automation system. Libraries that utilize CD-ROM cataloging products also create machine-¬readable versions of cataloging records. As another source of machine-readable cataloging data, libraries that purchase books and other materials from jobbers or other suppliers can often obtain magnetic tapes containing computer-processible cataloging records that reflect their acquisitions for specified time periods. Some book jobbers maintain machine-readable databases of LC MARC records supple¬mented by local cataloging that can be used for this purpose.

Thus, many libraries can readily obtain machine-readable versions of records acquired and cataloged since the late I960s, when the MARC program began. Retrospective conversion of a library's catalog-that is, the conversion of older catalog records to machine-readable form-is a more complicated un¬dertaking, however. Such records can, of course, be converted by direct data entry-that is, by typing them at a computer terminal or other input devices in the manner discussed in chapter I-but that is a conversion technique of last resort. The preferred retrospective conversion method involves the use of ab¬breviated keystroking to search for and obtain machine-readable copies of cata¬loging records from existing resource databases. While a few libraries have purchased or otherwise acquired the machine-readable catalogs of comparable libraries against which their own holdings can be matched, the databases of¬fered by bibliographic utilities and CD-ROM cataloging systems are the most important and widely used sources of cataloging copy for retrospective conver¬sion. Although the LC MARC resource records maintained by all bibliographic utilities and CD-ROM cataloging products primarily pertain to post-1968 im¬prints, the LC MARC database does include cataloging records for some earlier titles. As previously discussed, MARC-format records for selected pre-I968 im¬prints were created by the RECON project and COMARC program. In addition, occasional pre-1968 imprints have been acquired and cataloged by the Library of Congress since the inception of the MARC program. More importantly, cata¬loging contributed by subscribers of bibliographic utilities includes many records for older imprints entered by libraries that either have recently ac¬quired and cataloged those items or had previously contributed them as part of their own retrospective conversion projects. Certain CD-ROM cataloging systems also contain cataloging records for older imprints, but they cannot match the huge resource databases offered by the largest bibliographic utilities. For retrospective conversion, as with current cataloging, a large resource database increases the likelihood of retrieving suitable cataloging copy for a given item.

The use of bibliographic utilities and CD-ROM cataloging systems for ret¬rospective conversion involves straightforward work steps. A bibliographic database is searched to obtain cataloging copy for each record in a library's shelflist or card catalog. To simplify input and speed response, the resource database is typically searched by Library of Congress Card Number (LCCN) or another unique identifier. Many library shelflists contain LC printed cards that include the LCCN. If an appropriate bibliographic record is retrieved, it can be edited to conform to local cataloging practice, as reflected in the shelflist card. In the case of certain bibliographic utilities, the library's holding symbol is added to the cataloging record. Records processed in this manner can be or¬dered from a bibliographic utility in machine-readable form on magnetic tape or transferred electronically into a library's local automation system. With CD-ROM cataloging support products, the records are transferred from the CD-ROM database to a hard disk, from which they may later be copied onto tapes or diskettes.

The retrieval and modification of existing resource records, while still time-consuming and labor-intensive for large retrospective conversion projects, will prove faster and less expensive than full keystroking of the information content of a library's shelflist. Use of a resource database to retrieve cataloging copy pro¬duces a MARC-format record with the least effort. Full keystroking of original cataloging records remains necessary, however, for items that are not found in databases provided by bibliographic utilities or CD-ROM cataloging systems. In such cases, a library uses cataloging copy from its shelflist or card catalog as the basis for input. Original cataloging records can be entered directly into an inte¬grated library system's database, using the cataloging module described above. Alternatively, a bibliographic utility or CD-ROM cataloging system can be used for original cataloging and the resulting records transferred to the library's local system through electronic transfer, magnetic tape, or diskettes.

Several bibliographic utilities offer products that simplify retrospective conversion through offline preparation of information about the records to be converted. The prepared information is then processed against the utility's database in the batch mode. With OCLC's Cataloging Micro Enhancer pro¬gram, for example, a library prepares diskettes that contain search keys for cat¬aloging records to be retrieved from the World Cat database. The diskettes also include call numbers, holdings information, and other local cataloging infor¬mation from the library’s shelflist. The diskettes are shipped to OCLC, which retrieves the indicated MARC-format cataloging records from World Cat, enters the library-supplied local information, and transfers the results to magnetic tapes or diskettes for delivery to the library. RLIN's Marcadia service offers similar capabilities. A library creates brief search records that contain minimal bibliographic information plus call numbers and other holdings data for the ti¬tles to be converted. The search records are submitted to the Marcadia service on computer media or via the Internet for matching against the RUN database. Search records that match entries in the RUN database generate new cata¬loging records based on the information provided.

As an alternative to in-house conversion-projects: several bibliographic utilities and other companies offer fee-based retrospective conversion services that draw on their own resource databases and cataloging expertise. As an ex¬ample, OCLC RETROCON is a customized conversion service in which OCLC staff members, working under the supervision of experienced catalogers, con¬vert a library's shelflist to full MARC records by editing existing records re¬trieved from the World Cat database and performing original cataloging, as required; to library specifications. OCLC staff members also enter the library's' holding symbol for converted items. Similarly, the OCLC provides additional database preparation and authority control capabilities for retrospective con¬version and other applications. In addition to converting a library's shelflist to machine-readable form, OCLC will compare, validate, and update authority headings in bibliographic records by matching them against specified authority files. It will upgrade unauthorized headings to authorized forms, expand ab¬breviations, and correct obsolete subdivisions. Other capabilities include prepa¬ration of smart barcodes for circulation control, manual review of bibliographic records, and enrichment of cataloging records with table of contents informa¬tion. As described above, libraries typically supply lists of LC Card Numbers or other identifiers for the records to be converted. Alternatively, a library may send its entire shelflist, or a copy thereof. to the retrospective conversion ser¬vice. In either case, the conversion company typically furnishes MARC-format cataloging records on magnetic tape for input to mainframe or minicomputer-¬based integrated systems or on diskettes for input to microcomputer-based sys¬tems. In the latter case, the cataloging records are typically provided in the MicroLIF format, which is accepted by most microcomputer-based integrated library systems.

While bibliographic utilities, CD-ROM cataloging systems, or other sources of machine-readable data can facilitate retrospective conversion, a. conversion project may prove too expensive or time-consuming for a given library. Conse¬quently, some libraries that have implemented online catalogs have decided to "close" or "freeze" their card catalogs rather than convert existing records to machine-readable form. In this context, closing means that no additional records will be entered into the card catalog after some predetermined date. In effect, a static, retrospective card catalog will coexist with a growing online cat¬alog. As time passes, the online catalog will represent an increasingly large per¬centage of the library's collection. Further, it will contain cataloging records for newly acquired titles, which are presumably of great interest to library users. In some cases, closing the card catalog is a temporary expedient that permits the implementation of an online catalog within an available budget. The library may intend to convert all or part of the closed catalog at some future time when funding is available.

In actual practice, the complexity of library activities and bibliographic records rarely permits the complete closing of a card catalog. The most com¬mon closing strategies rely on cataloging date or imprint date. Each approach is limited in several important respects. In the former instance, all records for items cataloged after a predetermined date are entered into a library's online catalog, regardless of the imprint date of the item. While this is probably the simplest closing strategy from the cataloging department's standpoint, open entries for serials, series, and multivolume sets must be removed and con¬verted before the card catalog can be closed. As a source of potential confusion, users seeking older imprints must consult both the retrospective and current catalogs, as must the library staff when using the catalog for bibliographic ver¬ification and reference.

A closing strategy based on an item's imprint date may be more convenient for the library user, but it does not permit a complete closing of the card catalog, since older imprints may be acquired and cataloged in the future. Records for such items must be added to the card catalog. If different practices for choice and form of headings are applied to the old and new catalogs, catalogers must be trained in both sets of rules. A closing strategy based on imprint date does, however, permit continued maintenance of open entries in the card catalog. As an additional problem for research libraries and some special collections, nei¬ther of these dosing strategies addresses the problem of cataloging records in non-roman languages. While some integrated library systems can accommo¬date vernacular records in Cyrillic, Hebrew, Arabic, and other non-roman al¬phabets, most cannot. Unless transliteration is acceptable, the card catalog must remain open for such records, as well as for those containing the ideographic symbols used in the Chinese, Japanese, and Korean languages.

Even when a card catalog is completely closed, a library must continue to maintain it. Cards will remain subject to wear and tear, which necessitates some replacement activity. Similarly, some heading changes will be required to maintain compatibility with a library's online catalog. If different rules for the choice and form of headings apply to each catalog, some cross-references will be required.

OPAC CHARACTERISTIC

Broadly defined, an online public access catalog is an organized, machine-¬readable collection of bibliographic records that represents a library's hold¬ings. The records are stored on hard drives or other direct-access computer storage media for convenient retrieval by library users working at interactive terminals or appropriately configured microcomputers. As its name suggests, an OPAC has characteristics and capabilities that are designed for library users as opposed to library staff members, although librarians will use it as well. When compared to card catalogs, online catalogs offer a number of advantages. Software permitting, online catalogs permit information retrieval operations that are not possible with card catalogs; as previously discussed, practical con¬siderations prevent the expansion of a card catalog’s access points to include such additional parameters as keywords in titles, series, or subject headings. Online catalogs can retrieve records that contain specified combinations of re¬trieval parameters. Assuming appropriate communication arrangements, an OPAC can be accessed from any location by authorized persons equipped with compatible terminals. Thus, a scientist or engineer working in a field office can consult the catalog of a technical library implemented at a headquarters loca¬tion. Similarly, faculty members can access the online catalog of an academic li¬brary to determine the availability of needed research 'materials. High school students working on term papers can use home computers to search public li¬brary catalogs to identify relevant books. Further, the labor-intensive file main¬tenance routines associated with card catalogs are eliminated. If desired, online catalogs can be updated in real time, with records being added or removed as items are cataloged or weeded. Typically, however, changes to the catalog are made in the batch mode on a daily or other periodic basis. Because catalog records are stored on hard drives, they can be conveniently backed up on mag¬netic tape, which can be stored offsite for vital records protection.

Online catalogs can be implemented' in several ways. Libraries that sub¬scribe to certain bibliographic utilities, for example, already have a portion of their catalog records accessible online. As discussed in chapter 5, several biblio¬graphic utilities store duplicate and variant cataloging records contributed by participants. Such records accurately represent a given library's holdings, in¬cluding local call numbers. Increasingly, bibliographic utilities offer powerful information retrieval capabilities that support keyword searching and logical coordination of search parameters, as well as conventional author, title, and subject searches. It is consequently conceivable that a given library could use a bibliographic utility's database as its online catalog. Similarly, some online in¬formation services discussed in chapter 7 offer "private file" capabilities that allow customers to establish their own bibliographic or other data files for ac¬cess through the same retrieval software that is used to access publicly avail¬able data files. Such private files can contain a library's online catalog.

While interesting, these implementation options are limited in several im¬portant respects. An online catalog maintained by a bibliographic utility would be limited to those records that a given library had cataloged with the utility. Unless a retrospective conversion project is undertaken, a card catalog must still be maintained for older records not in machine-readable form. The wide¬area networking arrangements required to connect a library's OPAC terminals to a remote bibliographic utility or online information service can prove costly to implement, and a library must pay online charges or other fees to search its own catalog. Library users must be trained in the search procedures employed by the bibliographic utility or online information service. While some utilities and online services feature graphical user interfaces, others employ conven¬tional retrieval procedures that require memorization of commands. Finally, li¬braries that rely on bibliographic utilities or online information services for OPAC implementations will relinquish direct control over system priorities and the management of downtime and response time.

For most libraries, an online catalog implemented on a local computer is a more practical approach. During the 1970s and 1980s, some libraries imple¬mented online catalogs based on custom-developed software. Examples in¬clude the Library Computer System (LCS) at Ohio State University, the MELVYL system at the University of California, the LIAS system at Pennsylva¬nia State University, the Catalog Access System at Mankato State University, the LSCAN system at the Dallas Public Library, the SULIRS system at Syracuse University, the Phoenix system at the University of New Brunswick, and the Dartmouth Online Catalog at Dartmouth College. while some of these systems remain in use, customized software-development requires a substantial invest¬ment of resources. As discussed in chapter 2, it is rarely the preferred approach to computer implementations. It is possible to use a prewritten database man¬agement program to implement an online catalog or other bibliographic data¬base-and several programs, such as INMAGIC from Inmagic Incorporated and the STAR product line from Cuadra Associates, have been developed for that purpose-but most libraries will be better served by an integrated system that includes an OPAC module among its functional components. The OPAC modules of integrated library systems have improved steadily and significantly over the past decade. The best integrated systems support a versatile and flexible repertoire of catalog access capabilities that can address users' information retrieval requirements in many types of libraries.

The sizable body of library literature dealing with online public access catalogs differentiates command-driven from menu-oriented systems. In command- driven implementations, a user enters a search string that specifies the field to be searched followed by a specific value to be matched in that field. Often, mnemonic abbreviations-such as "A=," "T=," and "S="-are used for author, title, and subject fields. While some command-driven systems provide excellent instruction screens with concise directions that enable new users to begin search¬ing quickly others require formal initial training for effective utilization. Menu--¬driven systems are consequently preferred for inexperienced or occasional users. '

Most OPAC modules are menu-driven, although a command-driven mode may also be provided for experienced users or library staff members. Menu-¬driven interfaces are essential for institutions such as public libraries that serve diverse clienteles and have limited opportunities to train their user popula¬tions. Command-mode searching will appeal most strongly to academic or spe¬cial libraries, where faculty members, graduate students, scientists, or other frequent users with complex retrieval requirements can be trained in advanced search techniques. Command-mode searching is also potentially useful for ref¬erence librarians, enabling them to initiate retrieval operations that are cumber¬some or impossible to perform with menu-based searching.

Designed for novice users, menu-driven OPAC modules are characteristi¬cally straightforward in concept and operation. Search options, identified by abbreviations or numbers, are typically listed in introductory screens for opera¬tor selection. Brief instructions explain search procedures and indicate the proper syntax for search statements. More detailed information is presented in help screens that may be invoked by function keys. While vendors supply the initial text of help screens, libraries can often edit their contents to incorporate local information.

Most minicomputer- and mainframe-based integrated systems support conventional character-based video terminals as OPAC workstations. In micro¬computer systems and client/server implementations, however, OPAC work¬stations are Windows-based microcomputers or, less commonly, Macintosh microcomputers. Such systems rely on pulldown menus, dialog boxes, mouse operations, and other graphical user interface components to simplify the entry of search commands and formatting of retrieved information. Several vendors have implemented special OPAC interfaces for children. One popular example, The Kid's Catalog, was developed by CARL Corporation in collaboration with the Denver Public Library. Icons that represent broad topical areas subject headings, or names of popular authors simplify searches for children with limited typing skills. It also compensates for failed searches due to typing errors, misspellings, inaccurate punctuation, or improper syntax.

As a recent development stimulated by the Internet's immense popularity, some integrated library systems support OPAC searches by microcomputers equipped with popular Web browsers, such as Netscape Navigator or Micro¬soft Internet Explorer. Some vendors also provide Web server components that make a library's catalog accessible to remote users through the World Wide Web. Since the mid-1990s, hundreds of library catalogs have been publicly available, either through the World Wide Web or by using the Internet's telnet capability. Alternatively, the Web access capability might be used to provide ac¬cess to a library catalog in a corporate or government intranet installation.

Interface characteristics aside, online public access catalogs employ indexes to retrieve bibliographic records. With all integrated systems, the OPAC mod¬ules permit searches by author, title, and subject, thereby replicating the tradi¬tional retrieval functionality of card catalogs. Typically, a user specifies a field to be searched and a name, subject heading, or other value to be matched. Other re¬trieval capabilities vary from system to system. Classification number indexes are essential if libraries are to replace their shelflists with computerized records. Many systems support retrieval by one or more unique record identifiers, such as a system control number, Library of Congress Calrd Number (LCCN), or In¬ternational Standard Book Number (ISBN), but such capabilities are more ap¬propriate for library staff members than for library users. Other possible search parameters include publisher, publication date, country of publication, lan¬guage of publication, and media type. With some integrated systems, a previ¬ously entered search statement can be qualified by these parameters.

For maximum flexibility, some integrated systems allow any library¬-designated field within bibliographic records to be indexed for retrieval pur¬poses. In other cases, indexed fields are predetermined by the system's vendor, but the list is often sufficiently broad to satisfy diverse library requirements. Root-word searching, relational expressions in search statements, and Boolean operations for logical coordination of multiple search terms are commonplace. Some integrated systems permit keyword searching of designated fields, par¬ticularly corporate and conference names, titles, series, and subject headings. An "anyword" capability will search all indexed fields for a specified value. Several integrated library systems support proximity commands, wildcard characters, and other capabilities commonly associated with full-text retrieval systems. Such capabilities are most relevant for bibliographic records that con¬tain abstracts or other long text segments. They are difficult for uninstructed searchers to use effectively. In consortia or other multi-library installations, OPAC searches can often be limited to the collection of a specific library, usu¬ally the one in which the searcher is located. Some integrated systems allow a library to define a portion of its holdings, such as a special collections catalog, that can be searched separately.

If only one bibliographic record satisfies a retrieval specification, it is usually displayed immediately and completely. Where other search results are ob¬tained, responses vary. If multiple records are retrieved by a given search, most OPAC modules provide a count of the number of retrieved items, followed by a scrollable display of brief bibliographic records that include some combination of author, title, publication date, and call number. The searcher can then request a more detailed display for all or selected items.

Full bibliographic records are usually displayed with field labels that clearly identify specific data elements. With some integrated systems, display formats are predetermined for specific types of library materials; in other cases, a library can specify the fields to be included in brief and full record displays. The inclusion of holdings information, such as the locations of copies and their circulation status, in retrieved records is an essential attribute of integrated library systems. Such information is often contained in continuation screens that can be displayed at the searcher's option. Some integrated library systems can display full MARC records with tags and subfield codes, but that format is more appropriate for technical processing operations than for public catalog access

When no bibliographic records are retrieved by a given search, most OPAC modules display a scrollable, alphabetized list of field values that most closely ¬match the search term. If authority control is implemented, cross-references are displayed for invalid or related headings. In some cases, the referenced head¬ing is automatically substituted for an invalid term. In the absence of a cross-¬reference, some systems will automatically substitute the closest matching name or word for a search term that fails to retrieve bibliographic records, the assumption being that a typographical error or misspelling led to the retrieval failure.

To expand their functionality, integrated library systems increasingly incor¬porate external information resources into OPAC searches. Reference data¬bases, obtained on magnetic tapes from information publishers, can be locally mounted and listed as OPAC menu items for operator selection. Such databases will be discussed in chapter 7. Some vendors of integrated library systems also provide interfaces to CD-ROM information products, online information ser¬vices, and Internet sites. Some integrated systems also support local mounting of library-produced databases. As noted above, an information and referral module is offered as an optional application component by some integrated system vendors. It gives OPAC users access to information about community groups and local events. Depending on the system, community information can be retrieved by the name of the sponsoring organization, the title or date of the event, and keywords.

Since the mid-1980s, integrated systems have provided access to the cata¬logs of other libraries. The NISO 239.50 standard is widely supported for that purpose. It defines an interface that permits communication between the com:" pliant computer systems of different vendors. It allows an OPAC terminal of one integrated system to search an external computer system using the re¬trieval commands and operating procedures associated with the terminal's host system. The 239.50 standard is based on client/ server technology. The in¬tegrated system to be searched must be equipped with Z39.50 server software, while the integrated system at which searching originates must be equipped with 239.50 client software. 239.50-compatibility is not limited to integrated li¬brary systems. It has also been implemented by online information services, CD-ROM information products, and Internet sites.

Circulation Control

With the exception of archives, manuscript libraries, rare book collections, and other special repositories, the circulation of books and other materials is an im¬portant part of a library's mission. People who visit a public, academic, school, corporate, or government library expect to be able to borrow materials for pre¬determined periods of time for use in their homes or offices-assuming, of course, that they qualify as authorized borrowers. The principal purpose of cir¬culation control is to maintain records about the withdrawal of specific books or other library materials. Circulation records protect a library's investment in its collections by fixing the responsibility and date for the return of borrowed items. In addition, circulation records contain information about the use of library ma¬terials, which, when extracted and properly analyzed, constitutes a valuable aid to collection development, resource allocation, and decision making.

In manual implementations, a library's master circulation file usually con¬sists of paper cards or slips, each of which contains information about a partic¬ular circulation transaction. The specific information content of these records may include a brief description of the item (author, title, and call number, for example), its due date, and the borrower's identification number, name, and perhaps address. This master file reflects the library materials in circulation at any given moment. It is typically arranged in• a manner that enables a library to obtain information about the status of a given item. In many public libraries, for example, circulation cards are filed by due date for simplified identification and reprocessing of overdue items. In academic and special libraries, where it is often necessary to recall a given circulating item before the expiration of its loan period, the card file may be arranged by call number or author / title. In such sit¬uations, it can be difficult to identify overdue items.

In addition to a master circulation file, most libraries also maintain records about authorized borrowers. Such records, which may be maintained on cards or sheet-form rosters, usually include some combination of the following infor¬mation for each borrower: an identification number, name, address, telephone number, registration and card expiration dates, privilege category (such as adult or juvenile in a public library, faculty or student in an academic library), and the name of the issuing branch or library (in multi-branch or multi-library systems). Additional files may be created to address the special circulation requirements. Some libraries, for example, maintain lists of delinquent borrowers or of circu¬lating items to be held for specified borrowers on their return. Likewise, libraries may maintain circulation cards for returned items in a historical transaction file that is used for statistical analysis. Many scientific, technical, and business li¬braries maintain profiles of borrowers' interests that can be matched against newly acquired items. This approach to library service, sometimes described as the selective dissemination of information (SDI), is often considered a facet of reference service, but it can also be viewed as a form of anticipatory circulation.

Since the 1930s, libraries have sought, and more or less successfully applied, automated alternatives to the manual circulation control methodologies described above. The following discussion examines-the transaction processing and recordkeeping problems inherent in circulation control and surveys the characteristics and capabilities of circulation control modules included with in¬tegrated library systems.

Problems of Circulation Control

Library interest in computerized circulation is, in large part, based on a long¬standing recognition of certain problems inherent in manual circulation sys¬tems. Specifically, the recordkeeping operations associated with manual circulation control procedures are characteristically labor-intensive. They in¬volve such time-consuming work routines as filing single- or multi-copy book cards, checking card files to determine the circulation status of specific items, identifying overdue items with the subsequent preparation of borrower no¬tices, and calculating fines and other charges for overdue or lost materials. As¬suming, for example, an average of just one minute of work time associated with filing, removal, and other handling of each card in a master circulation file, a library with an annual circulation of 200,000 items must spend about 3,500 hours on file maintenance each year. At a labor rate of just $8 per hour, the annual cost of file maintenance will approach $28,000. In many public and aca¬demic libraries, this required labor commitment is increased by a high volume of circulation activity, long hours of service, and, in the case of multi-branch li¬brary systems, multiple service points.

As a further complication, most circulation control operations are clerical rather than professional in nature. Their routine character contributes to errors in filing and related recordkeeping tasks, as well as to low employee motivation and high staff turnover, with its associated retraining costs. Academic li¬braries, for example, often employ students as part-time circulation clerks, paying them lower hourly wages than would be required for full-time clerical personnel. Such students rarely find the job sufficiently challenging to maintain their interest. As a consequence, they may make errors or require considerable supervision, which, in turn, increases total system costs.

Errors and staff turnover aside, the scope of manual circulation systems is necessarily limited to activities directly related to the charging and 'discharging of library materials. In most cases, accurate circulation statistics-which are es¬sential to the informed management of collection development activities and appropriate expenditure of limited acquisitions budgets-are very difficult or impossible to derive. Properly prepared and presented, circulation statistics can assist bibliographers and other librarians in determining when additional copies of a given item are required and in identifying portions of a library's col¬lection that are infrequently utilized and may consequently warrant a reevalu¬ation of prevailing selection policies. Given the high prices libraries must pay for books and other materials, the ability to avoid inappropriate purchases is of obvious importance.

As a further constraint, there is typically little or no relationship between a manual circulation system's card or paper files and a library's catalog or other records that contain much of the same bibliographic information. Thus, a user who consults a library's catalog for the call number of a specified book cannot immediately determine the item's circulation status. Finally, from the service standpoint, manual circulation systems place much of the workload on the bor¬rower, who must fill out charge slips or cards for each item to be circulated. This inconvenience, combined with long waiting times attributable to the slow¬ness of ma11ual check-out procedures, can lead to user dissatisfaction. These problems are not unique to library circulation control. They are characteristic of many business applications that involve the maintenance and control of large document files. Such applications are widely encountered in insurance compa¬nies, hospitals and medical clinics, personnel departments within corporations and government agencies, registrars' offices in colleges and universities, law of¬fices, police departments, and other paper-intensive work environments. Record tracking systems developed for business applications are often based on library circulation control models.

Although computerized circulation systems are the subject of this discus¬sion, there are other alternatives to manual circulation control. Through the mid-1970s, for example, some small and medium-size public libraries utilized photographic charging systems in which a book identification card, borrower identification card, and date information pertaining to successive circulation transactions were recorded on microfilm using a specially designed camera. While such pre-computer approaches to circulation control remain in use in a few libraries, they are outdated in concept and offer limited functionality. More radical alternatives to manual circulation control have been suggested but, for the most part, dismissed. Few library administrators would, for example, aban¬don circulation control in favor of an honor system in which borrowers assume personal responsibility for returning books after a specified period of time. Even if most borrowed materials were returned on time or eventually, such a circulation method leaves a library with little or no information about the status and use of its collection at any point in time. Still, the idea of abolishing circula¬tion control cannot be dismissed completely. Some libraries apply it selectively to library materials, such as paperback fiction titles, that are not permanent col¬lection resources or that cost more to circulate than to replace.

A somewhat less radical, but still striking, departure from conventional cir¬culation control concepts involves the substitution of duplication for circula¬tion. In a paper delivered at the 1962 convention of the National Microfilm Association, Laurence Heilprin first outlined the concept of a duplicating li¬brary in which an inviolate collection of books and other materials, in full-size or microform, would be copied on demand and distributed to library users as an alternative to circulation. Recipients could keep the copies or dispose of them. The cost of duplication would presumably be defrayed by the elimina¬tion of circulation-related expenses associated with recordkeeping, the recall of overdue materials, and the replacement of non-returned items. In addition, since they are not removed from the library, materials are continuously avail¬able and multiple-copy purchases can consequently be minimized.

Unfortunately, the relative economics of a wholesale conversion from circu¬lation to duplication have never been fully explored. In addition, copyright re¬strictions may prevent copying of certain materials or significantly increase the cost and complexity of duplication. Although it is probably not a viable alter¬native for all library applications, the duplicating concept has been selectively adopted for certain types of library materials. It is, for example, the prevailing method of providing journal articles through interlibrary loan. Similarly, some libraries employ duplication as an alternative to the circulation of microfiche. Rather than circulating the library's master fiche, the requester is given a copy for personal use and disposition. Desktop microfiche duplicators are avail¬able for this purpose. In the case of microfiche report collections produced by the Educational Research Information Center (ERIC), National Technical Infor¬mation Service (NTIS), and other government agencies, there is no copyright impediment to such duplication. Copyright restrictions may apply, however, to the duplication of microforms purchased from commercial micropublishers.

In the broadest sense, the widespread availability of coin-operated copiers in libraries probably constitutes the most common and effective implementa¬tion of duplication as an alternative to circulation. A would-be borrower who is not interested in an entire book or serial publication can copy relevant sections. The user, of course, pays for the copies made, but in many libraries the total cost of copier operation is not recovered through such charges. Presumably, those libraries that subsidize coin-vend or other copying operations are doing so to encourage their use and to achieve a corresponding reduction in circulation- related costs.

Evolution of Computerized Circulation Systems

Although the concepts and methods described above have been successfully applied in certain library situations, only computerized systems offer broadly applicable solutions to the problems of manual circulation control. Perhaps more than any other library activity, the historical development of automated circulation control has reflected changes in the state of the art in data processing technology. As early as the 1930s and extending into the 1960s, a number of li¬braries used punched cards in combination with sorters, collators, and other unit record equipment as an alternative to manual recordkeeping. Tabulating cards punched with information about books, borrowers, and due dates could be sorted, for example, to select overdue items or to identify all books on loan to a given person.

Such pre-computer data processing systems were typically based on inven¬tory control models used in business. With the introduction of computers for business applications in the mid-1960s, a number of libraries developed com¬puterized circulation control systems based on batch processing techniques. Typically, such systems were implemented on a computer located in a data pro¬cessing center operated by a university, municipality, or corporation with which the library was affiliated. Information about circulation transactions was recorded on punched cards or magnetic tape for processing against a master circulation file at predetermined intervals. Following processing, a list of circu¬lating items arranged by call number or title, with due dates, was printed for reference purposes. Periodically, the master circulation file and borrower infor¬mation files were further processed to identify overdue items, produce recall notices, derive circulation statistics, or print special lists. While they eliminated the maintenance of paper files and the time-consuming preparation of overdue and recall notices, such batch processing systems were limited in several im¬portant respects. They could not provide accurate information about items in circulation at a given moment; some information in printed lists was necessar¬ily invalidated by circulation transactions occurring between processing inter¬vals. Batch processing systems could not effectively identify items to be held for specific library users when returned from circulation, and they provided no method of blocking circulation transactions for delinquent borrowers or under other exceptional conditions.

By the mid-1970s, computer manufacturers and information specialists had begun to concentrate on the development of online, real-time systems. Follow¬ing their lead, several libraries implemented real-time circulation control sys¬tems that processed check-out and check-in transactions as they occurred. Widely publicized examples included the Library Computer System (LCS), which was developed at Ohio State University and subsequently replicated at the University of Illinois and State University of New York at Albany; the Bell Laboratories Library Real-Time Loan (BELLREL) system; and a self-service charge-out system developed by Northwestern University. Addressing the lim¬itations of their batch processing predecessors, these real-time systems estab¬lished the pattern for all subsequent developments in computerized circulation control. Because master circulation files are updated immediately rather than at predetermined intervals, real-time systems accurately reflect the current status of a library's circulating collection. Computer-processible files of items on hold and delinquent borrowers can be automatically checked as items are charged out or returned. Online terminals are used to inquire about the circulation sta¬tus of given items. Most printed listings are consequently eliminated, although overdue notices, recall notices, and other documents must still be produced in the batch mode.

Through the mid-1970s, computerized circulation control systems, whether operating in the batch or real-time mode, were developed and implemented on a customized basis for a particular library. Following a detailed systems analy¬sis, equipment was selected and programs written to meet the specific needs of a given application. As might be expected, the most interesting and innovative systems were developed by large academic, public, and corporate libraries that had access to institutional computing resources and technical expertise. For the typical medium- to large-size library, however, the computing facilities re¬quired to develop a customized circulation control system were too often un¬available. In most organizations, access to computing resources is allocated according to pre-established priorities, and the library's requirements are some¬times given inadequate consideration.

Even when a library is able to obtain appropriate computer hardware, prob¬lems of software development can prove significant. In computer applications generally, it is the absence of suitable software that constitutes the most formida¬ble obstacle to automation. Comparatively few libraries have resident systems analysts or programmers. Institutional computing centers allocate software sup¬port personnel in much the same manner as hardware resources. In many cases, there is a long waiting list to consult a systems analyst or programmer / analyst, and, once begun, customized software development is a slow, labor-intensive, and error-prone activity.

Prewritten circulation control software addresses this problem. It recognizes that certain aspects of circulation control are performed in more or less the same way, regardless of their specific library setting. As its principal advantages, prewritten circulation control programs can be implemented far more quickly and at lower cost than customized installations. While custom-developed soft¬ware is funded by a single library, the development cost of prewritten programs is shared by multiple customers. Speed of implementation is especially important in applications where automation of the circulation activity will yield a cost reduction or improved service when compared to the manual system it is de¬signed to replace. Prewritten circulation control software minimizes or eliminates requirements for local computer expertise. Libraries acquiring such products do not need to hire programming staff or take programming courses. In fact, cus¬tomer programming is usually prohibited by most vendors. Because the system design is predetermined, user training is limited to operational considerations.

The earliest examples of prewritten circulation control programs, intro¬duced in the 1970s, were single-purpose products designed specifically and exclusively for circulation control. Initially, they were marketed as turnkey sys¬tems-that is, preconfigured combinations of computer hardware and prewrit¬ten programs sold as a package. In most cases, such turnkey circulation control systems included minicomputers that were installed in and operated by the li¬brary itself; in the 1970s, mainframe computers were far too expensive for most libraries, and useful microcomputers did not become available until the early 1980s. As an alternative to turnkey configurations, some vendors offered prewritten circulation control software for implementation on separately pur¬chased minicomputers, including equipment already owned by a library or its parent institution.

During the 1980s, as previously noted, single-purpose circulation control programs were supplanted by integrated library systems, in which circulation control is one component or application module. Along with cataloging and on¬line public catalog access, it is typically a core module rather than an optional component. The circulation modules of integrated library systems effectively automate a variety of tasks, including check-out, check-in, and renewal of li¬brary materials; placement of holds; fines calculation, collection, and record¬keeping; creation, maintenance, and retrieval of borrower records; and printing of reports and notices. These capabilities are described in the following section.

Circulation Control Capabilities

All computerized circulation control systems maintain machine-readable infor¬mation pertaining to a library's circulating collection and its borrowers. With integrated systems, holdings information about individual copies of specific ti¬tles is linked to bibliographic records in a library's catalog. Such holdings infor¬mation is essential to the circulation activity, since a library circulates copies rather than bibliographic entities. Subject to product- and application-specific variations, typical holdings information includes some combination of a bar¬code number or other item identifier, a copy number, a loan period category, a local call number, the item's normal location or branch, a temporary location or branch, the media type, the original price, the replacement cost, the item's cir¬culation status (on shelf, checked out, or other), the borrower's identifier and due date for items in circulation, and the date of last circulation activity.

A borrower file contains one machine-readable record for each registered li¬brary user. In public libraries that serve medium-size and larger cities, this file may contain 100,000 or more records. Academic, school, and special libraries, by contrast, usually have smaller clienteles. File size aside, common data fields include the borrower's name, address, and telephone number; the date regis¬tered, expiration date, and date of last circulation activity; the borrower's priv¬ilege category (such as faculty member, graduate student, or undergraduate student in an academic library, adult or juvenile in a public library); the bor¬rower's card number, social security number, or other identifier; the name of the issuing library or branch in a multi-library system; an indication of delin¬quency status with a reason for delinquency; a count of the number of items in circulation; and a message field for several lines of text. Several circulation con¬trol modules maintain a count of the number of lost books, claimed returns, and similar problems associated with individual borrowers. Some programs also provide optional data fields for personal and demographic information¬ such as age group, occupation, and primary language-that supports statistical analysis of circulation activity for particular types of library users. Additional data fields may be provided for special situations. Academic libraries, for exam¬ple, typically include students' campus and home addresses in borrower records. Similarly, corporate and government libraries may record both office and home addresses and telephone numbers. School libraries may require data fields for a student's homeroom and parent's name.

Regardless of content, the creation of borrower files is a potentially time-¬consuming and costly activity. Key-entry of borrower information, a method of last resort, is often necessary. As an alternative, some integrated systems can import machine-readable borrower records from a previously implemented cir¬culation control system, registrars' files, personnel databases, or other sources. Although special programs must be written to convert the transferred records to a format required by the circulation control module, the cost of required pro¬gramming usually compares favorably with the cost of key-entry labor. In most cases, vendors of integrated library systems will develop such conversion pro¬grams for a negotiated fee. Rather than loading borrower records from external sources, some public libraries reregister all their borrowers at the time an inte¬grated system is installed, upgraded, or replaced, thereby purging their files of obsolete records.

All integrated systems support three broad types of circulation control oper¬ations: charging and discharging of library materials, online file inquiries, and offline printing of reports and notices. With the earliest circulation control pro¬grams, libraries were forced to accept predetermined loan periods, borrower categories, and other operating conditions. Newer products, however, are much more flexible. Drawing on three decades of library experience with computer- based circulation management, the circulation control modules supported by integrated systems are very well developed and highly parameterized-that is, they allow libraries to specify the conditions under which items will be circu¬lated, file inquiries made, and printed output generated. Rather than being writ¬ten into programs, particular operating parameters are selected by libraries from a range of possibilities.

With some integrated systems, vendor representatives establish circulation control parameters during a pre-installation "profiling" session. Alternatively, the system may include a programmable module that allows libraries to define or change their own operating parameters. In either case, a library can define the period of time-in days, weeks, or other measures-that specific types of items will be loaned to, and renewed for, specific types of borrowers. Some systems permit renewal periods that differ from the original loan period. Addressing a commonly encountered academic library requirement, most circulation control modules can charge out items until the end of a semester or another fixed date. If desired, this option can be limited to specific borrower categories-faculty members and graduate students, for example-or to particular types of library materials. Addressing the common practice among corporate and government libraries of maintaining selected items in the work areas where they are most fre¬quently utilized, some circulation control modules permit permanent loans to desk or laboratory locations rather than individuals. For chargeback purposes, some integrated systems can link the use of specific items to individual depart¬ments or cost centers within a corporation or government agency.

A library can differentiate fine rates by item or borrower type, specify fine immunity for particular types of borrowers, grant a specified number of grace days before fines are incurred, and specify the number of overdue notices that different types of borrowers will receive. Libraries can also impose limits on the amount of fines that specific types of borrowers can incur before circulation privileges are revoked. To ensure that holidays are not used in fine calculation and that due dates do not fall on holidays, most circulation control modules maintain a calendar of dates that the library is closed. For maximum flexibility, circulation parameters can be defined separately for individual participants in consortia or other multi-library arrangements. If desired, borrower privileges ¬can be limited to specific branches or libraries within a multi-library system.

All circulation control modules perform check-out, check-in, and renewal operations in real time-that is, all item and borrower records are immediately updated when library materials are charged out or checked in and the circula¬tion status of specific copies is immediately reflected in OPAC displays. With most integrated systems, circulation control is accorded the highest priority for execution, so that circulation transactions are performed before catalog searches or other operations.

Check-out procedures are straightforward. Working at a designated termi¬nal, a circulation clerk enters book and borrower identifiers. Applying predeter¬mined circulation rules, the system retrieves the appropriate records, calculates the due date, and updates holdings information for the circulating item. All cir¬culation control modules support barcode recognition to simplify the entry of item and borrower identifiers. As discussed in chapter I, barcodes use predetermined patterns of alternating vertical lines and spaces to encode numeric identi¬fiers. In library applications, barcodes are typically printed on adhesive labels that are affixed to circulating materials and borrower identification cards. Bar¬code labels may be ordered preprinted or produced by the library itself.

In either case, specific numbered labels must be associated with the partic¬ular items to which they are attached. This can be done in several ways. In one method, a barcode label is attached to a book or other item, the database record for the item is retrieved by title or some other parameter, the barcode is scanned, and its number is entered in the barcode field within the record. The database record is then saved and the item is shelved. Alternatively, duplicate labels can be affixed to library materials and shelflist cards. When the shelflist is converted, the barcode label for each title is scanned into the database record. This method is particularly useful where a bibliographic utility or other com¬pany will perform retrospective conversion from a library's shelflist. Another popular method involves "smart" barcodes that include preprinted call num¬bers, titles, or other human-readable identifiers for library materials. Smart bar¬codes are produced from a library's database and must be attached to the materials for which they are intended. Smart barcodes are typically printed in shelflist sequence to simplify the labeling of items.

Regardless of the method employed, barcoding a library's collection is a time-consuming procedure that can take weeks or months to complete. Various types of barcodes are available. Most formats encode fourteen to sixteen digits. Often, the first digit is used to distinguish item and borrower barcodes. The next several digits identify the library in a multi-library installation, and the remainders identify the item or the borrower. A check digit, used for error detec¬tion, is often appended. Circulation control programs are designed to read barcode labels in one or more formats, and different formats cannot be inter¬mixed within a given library collection. Incompatibility of barcodes can pose problems and necessitate relabeling of items and borrower cards in libraries that change circulation control systems.

Though not as common as barcodes, some circulation control modules can read OCR labels. As discussed in chapter 1, optical character recognition uses reflected light to determine the character content of input documents. OCR la¬bels are usually printed in a specially designed type font, such as OCR-A or OCR-B, that is optimized for machine recognition. If barcodes or OCR labels are absent or damaged, item and borrower identifiers can be key-entered. Key¬entry is also used for telephone renewals. As an increasingly popular capability, some integrated library systems support self-service circulation terminals.

All circulation control modules block check-out transactions when bor¬rower delinquencies or other exceptional conditions are encountered. Exam¬ples include borrowers with overdue items or outstanding fines, borrowers who have exceeded the library's predefined charge-out limit, borrowers who are using an identification card reported lost or stolen, items that are being held for another library user, items that are already checked out to another borrower, or items that are not part of a library's circulating collection. Some integrated systems will automatically block circulation transactions if a borrower's circulation privileges will expire before an item's due date. Most systems distin¬guish between "soft" blocks, which can be overridden by a circulation clerk or other authorized person, and "hard" blocks, which require correction of the ex¬ceptional condition before the circulation transaction can be completed.

With most circulation control modules, check-in of library materials is a re¬peating transaction that is initiated by a command, followed by scanning of barcodes for the returned items. The check-in command is entered once for multiple items. During check-in, most integrated systems will automatically detect such exceptional conditions as items that are on hold for other borrow¬ers, items that were never properly charged out, and items returned to the wrong location. To reflect the in-house use of library materials in statistical reports, some circulation control modules permit the discharging of items removed from shelves but never checked out. Fines can be calculated for immediate collection when overdue items are checked in. If overdue items are dropped in book bins, or the borrowers are otherwise unavailable or unable to pay the fines, the amounts are typically added to their records. In most cases, a library can specify the amount of fines that borrowers can owe before circula¬tion privileges are revoked. Some systems support amnesty days, when all fines for overdue items are forgiven.

To facilitate the completion of specific circulation transactions, item records can be retrieved by barcode numbers or by any of the search parameters sup¬ported by the integrated system's online public access catalog. Once a desired bibliographic record is retrieved, its associated item records can be displayed. Borrower records are usually retrievable by card number or other numeric identifiers or the borrower's name. The content of displayed records varies from system to system. Typical data elements include the borrower's name, ad¬dress, telephone number, home library (in a multi- library installation), the date borrowing privileges were granted, and the date of last circulation activity. An operator can usually request additional information, such as a list of the spe¬cific items charged out to the borrower, any items the borrower has requested, a list of the borrower's overdue items, and the borrower's delinquency status. Library permitting, some integrated systems give registered borrowers online access to their own records, thereby reducing staff time spent on borrower in¬quiries. Borrowers can review items they have in circulation, items they have on hold, fines owing, and any blocks on their records.

Effective management of holds and recalls is one of the features that clearly distinguish real-time circulation control systems from their batch-oriented pre¬decessors. As used in this context, a hold causes an item in circulation to be trapped on its return and held for a borrower who previously requested it. A hold is a passive request. The borrower waits for the desired item to be returned, although he or she must usually specify a cancellation date after which the item will no longer be required. A recall, in contrast, is an active request that changes the due date for a circulating item in order to obtain its immediate return.

Most circulation control modules support the placement of holds at the bib¬liographic level-that is, on all copies of a specified title-or on a specific copy only. Bibliographic-level requests will trap the first available copy of a specified title involved in any circulation transaction. In multi-branch or consortium installations, copy-level requests are usually made by borrowers who do not want to travel to another location to obtain an item. Recalls are usually placed at the copy level and backed up with holds at the title level. When multiple re¬quests are received for a given item, a hold queue is established. In most cases, the request date determines a borrower's position in the queue, although au¬thorized library staff members can prioritize requests. Some circulation control modules will generate a purchase alert when the hold queue for a given title ex¬ceeds a specified number of names. Library permitting, borrowers can use OPAC terminals to place holds on circulating items retrieved through catalog searches. Depending on the system, holds may also be placed on items that are on order or awaiting cataloging

To conserve computer resources and preserve response time for online op¬erations, integrated library systems typically print circulation reports and no¬tices at night or during other off-peak hours. Most circulation control modules provide a selection of preconfigured operational and statistical reports that ad¬dress a broad range of library requirements. Some integrated systems also sup¬port a general-purpose report-writing program that the library can utilize to generate additional or customized circulation reports. Examples of preconfig¬ured circulation reports include borrower rosters with identification numbers, addresses, phone numbers, and similar information; lists of borrower records added or changed since a specified date; lists of borrowers with excessive over¬due items, fines, or other delinquencies; lists of borrowers with more than a specified number of items in circulation; lists of items being held for specific borrowers; lists of items on hold in call number or title sequence; lists of excep¬tional items on hold, including lost, missing, or long overdue items; lists of items with hold queues longer than a specified number of names; lists of re¬questers, with phone numbers, for items on hold; lists of fine receipts by work¬station location; and lists of overdue items for shelf searching prior to printing borrower notices. Examples of circulation reports that support inventory con¬trol activities include lists of lost or missing items arranged by call number, ma¬terial type, or other parameters; lists of items withdrawn from the library's 'collection since a specified date; lists of delinquent items claimed to be re¬turned; and lists of unsatisfied holds since a specified date.

Statistical reports provide useful information for collection development as well as for management of the circulation activity itself. Perhaps the most im¬portant example provides circulation totals for specified call number groups. It may be produced monthly, quarterly, or at other specified intervals. For com¬parison purposes, such reports often provide statistics for the same period dur¬ing the preceding year. Other useful statistical reports summarize circulation activity by borrower category or the type of library material; charge-out, check¬ing, and hold activity for specified time periods; and workstation activity by day, week, month, or other time periods.

All circulation control modules can generate overdue notices, recall notices, fine notices, hold availability notices, and similar borrower notices, as well as bills for lost items. Libraries can specify the content of borrower notices. They are usually printed on letter-size paper with the borrower's address formatted for compatibility with window envelopes. Some systems can also print notices on postcards. Notices may be sorted by zip code prior to printing. As an alternative to printed notices, some integrated systems support an automatic tele¬phone dialing and voice notification capability for holds and overdue items. Such capabilities, which utilize voice synthesis technology, can store multiple message scripts.

. Designed specifically for academic and school libraries, reserve room capa¬bilities maintain circulation records that are linked to courses and instructors. With some integrated library systems, reserve room capabilities are incorpo¬rated into the circulation control module; in other cases, a separate reserve room module is available as an optional component. Implementation pattern aside, reserve room materials are identified in the online public access catalog as subject to special, usually short, loan periods. Records for items on reserve can be retrieved through OPAC searches or by course number or instructor. In addition to managing items from a library's own holdings, most reserve room modules can create brief catalog records for photocopies and instructors' personal copies. Circulation periods for reserve materials are customarily measured in hours or even minutes, although most systems will permit overnight circulation if the allowable charge-out period extends beyond the library's clos¬ing time. Most integrated systems will print lists of items on reserve for partic¬ular, courses reports of charge-out activity arranged by course number and item, lists of reserve items that are overdue, and lists of items to be withdrawn from reserve.

As another optional system component some integrated library systems support a media booking module that provides circulation control for video¬tapes, films, and other audiovisual media, as well as equipment and facilities, that are subject to advance reservations. An online calendar is consulted to de¬termine availability and reserve items for specific dates. Holiday and weekend reservations can be prohibited. Check-out and check-in functions are similar to those for circulation control. Reservation records may be accessed by borrower name, item identifier, or other parameters. Typical management reports and notices include media catalogs, pick lists, packing slips, mailing lists and labels, reservation work slips and schedules, media activity reports, overdue notices, and usage summaries. The media booking module is typically integrated with other system components. With some integrated systems, OPAC users can make reservations for media or items.

As previously discussed, most integrated library systems support acquisitions and serials management modules as optional application components. While acquisitions and serials management are often relegated to separate depart¬ments within a given library, they share an important function: the procure¬ment of library materials. This discussion will follow common library practice in using the term "acquisitions" to collectively denote those tasks that support the procurement of library materials that are published on a nonrecurring basis, including books, technical reports, government publications, musical scores, sound recordings; and visual materials. The acquisition of journals, magazines, and other periodical publications is considered a facet of serials management, the scope of which extends to such activities as cataloging and binding of serial publications. The distinction between acquisitions and serials' management is rarely unequivocal, however, since annuals or irregularly published materials may be treated as monographs in one library and as serials in another.

Since acquisitions and serials management initiate the procurement of li¬brary materials that will eventually be cataloged and circulated, they would seem to be more appropriately discussed at the beginning rather than at the end of this chapter. They are treated here, however, because an understanding of automated acquisitions and serials management requires some familiarity with the application components described above. As with automated circulation control and computerized catalog production, libraries have had decades of experience with automated acquisitions. Pre-com¬puter book ordering systems-using keypunch equipment, card sorters, and tabulating machines-were developed in the late 1950s by public and academic libraries. The first computer-based acquisitions systems, implemented by a number of academic libraries in the early to mid-1960s, employed batch pro¬cessing technology to automate procurement and recordkeeping tasks. During the late 1960s and early 1970s, several academic libraries developed online ac¬quisitions systems. Widely publicized examples included the Book Order and Selection System (BOSS) at the University of Massachusetts, the Library On¬Line Information and Text Access (LOLITA) system at Oregon State University, and the acquisitions component of the previously discussed BALLOTS system at Stanford University. The first minicomputer-based turnkey acquisitions sys¬tem was installed by CLSI at the Cleveland Public Library in 1972. Other ven¬dors subsequently introduced similar special-purpose products; the most popular example, the INNOVACQ system from Innovative Interfaces, auto¬mated both acquisitions and serials management.

Integrated library systems began incorporating acquisitions modules as op¬tional application components in the 1980s. The earliest examples were notice¬ably weaker than their OPAC and circulation control counterparts, but they have improved steadily and significantly in recent years. As described below, the latest versions offer excellent functionality for ordering, fund accounting, and record¬keeping. For many libraries, particularly those that operate an integrated system for online catalog access and circulation control, the addition of an acquisitions module is the simplest, least expensive way to automate that activity.

Library motives for computerizing acquisitions operations are well estab¬lished. Acquisitions funds represent a large and important component of the typical library budget, and library administration is accountable for their re¬sponsible expenditure . As with circulation control, manual acquisitions and fund accounting systems cannot readily generate the financial and statistical in¬formation essential to scientific planning and management. In libraries where book budgets are allocated annually, for example, certain acquisitions funds may remain characteristically underspent over a period of many months, neces¬sitating intensified expenditures at the year's end, a situation that is not con¬ducive to carefully considered selection practices. Similarly. certain funds may be expended too quickly, leaving little reserve for important materials pub¬lished later in the year.

In either case, periodic reports of fund status can alert subject bibliographers and library administrators to exceptional situations that require attention and possible corrective action. While such reports can be produced manually, the available labor in most manual acquisitions systems is fully occupied with pa¬perwork processing, accounting, and related work routines pertaining to book orders. Among other advantages, computer-based acquisitions can reduce cleri¬cal labor .requirements and costs, thereby making more money available for col¬lection development and professional services. Cost reduction aside, many librarians perceive an added value in the potentially faster paperwork processing and more timely receipt of materials permitted by computerized acquisi¬tions capability. Further, automation of acquisitions promotes the integration of library operations. Bibliographic information, captured in machine-readable form at the time an item is ordered, can later be enhanced or modified to sup¬port information retrieval, circulation control, or other library operations. In in¬tegrated system implementations, a library's online public access catalog can contain information about materials that are on order or awaiting cataloging.

While differences in purchasing practices and procurement regulations may lead to local variations, most library acquisitions departments share certain basic characteristics and work steps. Depending on the type of library involved, a request to purchase a specific item may be transmitted to the acquisitions de¬partment by subject bibliographers, other library staff members, library users, or other persons or groups. Typically, the library's own acquisitions files and cata¬log are first consulted to determine whether the item is on order or already in its collection. Assuming that the item is not already owned or that an additional copy will be purchased, the bibliographic information in the request must be verified, the item's availability and price determined, and a vendor selected. A purchase order is then prepared in multiple copies, several of which are filed by the acquisitions department. When an item is received, the corresponding docu¬ments are removed from files and updated as required to reflect the full or par¬tial receipt of an order. Payment is then authorized and a payment voucher or check is issued.

Automated acquisitions systems retain these basic characteristics but replace typing, filing, and related manual work steps with data entry and computer pro¬cessing. Automated acquisitions control typically employs a combination of three data files: (1) an order file, sometimes called an in-process file, contains one record for each item purchased; (2) a vendor file contains one record for each publisher, book jobber, or other procurement source; and (3) a fund file contains one record for each account that supports the purchase of library materials. While specific details will necessarily vary from one installation to another, most order. files contain a combination of bibliographic and order-specific data, including au¬thor, title, and imprint information; an edition statement; an International Stan¬dard Book Number (ISBN) or other unique numeric identifier; an order number; the order date; a fund account number; a vendor number, code, or similar identi¬fier; the type of procurement (conventional order, standing order, prepayment, approval item, etc.); the number of copies ordered; the price, plus discounts and service charges; the currency type; the estimated receipt date; the ordering branch, department, or agency; and the name of the requester, bibliographer, or other person who initiated the order. Some systems also include a. text field for free-form messages intended for acquisitions clerks, catalogers, or others.

As noted above, vendor files contain one record for each publisher, book jobber, dealer, sales agent, or other procurement source. To accommodate gifts and exchanges, the vendor file may also include donor records. Again subject to variation from installation to installation, commonly encountered data fields include the vendor's name, address, telephone number, fax or telex number, and e-mail address; an assigned numeric or other coded identifier that links specific order records to the vendor file; pertinent contract numbers and dates; applicable discount schedules; a claim period indicator that specifies the inter¬vals at which notices will be sent for overdue orders; and the total volume of vendor order activity, as an item count or dollar value, for a specified period of time. Some vendor files also incorporate performance statistics, such as the av¬erage time each vendor requires to fill an order and the number of claims or cancellations experienced in a specified time period.

Most fund files contain one record for each account or other budgetary unit established by the library for acquisitions purposes. These funds may reflect subdivisions of a library's collections, academic departments within a college or university, branch libraries within a public library system, special accounts established for the purchase of particular types of materials, or gifts from spe¬cific individuals or groups. For each fund file record, typical data fields include the account number, a brief account description, the original dollar amount al¬located at the beginning of the budgetary cycle, total expenditures since the be¬ginning of the budgetary cycle, encumbrances for items ordered but not yet received, and the current fund balance.

Some automated acquisitions systems also include a requester file, which contains names, addresses, and other pertinent information for persons or or¬ganizations, including branch libraries, that may initiate acquisition requests. It may be used to print notification slips when requested material is received by the acquisitions department or as other information about the status of re¬quested items becomes available. In some cases, the requester file and vendor file are merged in a master name and address file, although vendor records may contain more information than requester records. A few automated acqui¬sitions systems maintain invoice files to permit the retrieval of information by invoice number in answer to questions about the payment status or history of particular transactions. Similarly, some systems feature an online desiderata file that contains records for items being considered for purchase. In the case of antiquarian or out-of-print items, lists of desiderata records may be printed for circulation to book dealers.

Vendor, fund, and requester files are typically created in advance of system operation, although new records may be added to them at any time. In most cases, their information content is derived from files of source documents used by the automated system's manual predecessor. Records in the order file are key-entered when items are ordered.

The acquisitions modules supported by integrated library systems can ac¬commodate a broad range of procurement transactions, including new orders, multi-copy and multivolume orders, standing orders, subscriptions, blanket or¬ders, approval plans, depository items, rental books, items obtained--through memberships, prepaid orders, gifts, and exchanges. Like their circulation control counterparts, acquisitions modules are highly parameterized. Libraries, or indi¬vidual branches in a library system, can specify file contents and operating con¬ditions. Because acquisitions modules are fully integrated with other system components, records for ordered items are included in a library's bibliographic database, and on-order status is reflected in the online public access catalog. With some integrated systems, as previously noted, library users can place holds on materials that are on order. Further, the integrated system's bibliographic database can be used for preorder searching to identify items that are already on order or owned by the library. While this capability does not eliminate the need for external reference sources for bibliographic verification, it can prove espe¬cially useful for minimizing duplicate ordering in multi-library or multi-branch installations where the integrated system maintains a .union catalog. In most cases, bibliographic information retrieved from the online catalog can be trans¬ferred directly into order records, thereby reducing key-entry requirements.

One of the principal advantages of automated acquisitions is the substitu¬tion of computer-based file maintenance for the labor-intensive sorting, filing, and other paper-handling procedures associated with manual acquisitions sys¬tems. When an order record is entered, the acquisitions module automatically encumbers the estimated purchase price in the indicated fund account. As items are received, order records are searched and their status updated to re¬flect full or partial shipments. When an order is closed, the corresponding records can be automatically purged from the order file to a historical file. Ac¬quisitions modules can produce claiming and cancellation notices, thereby eliminating a time-consuming work step encountered in manual systems.

Superior file inquiry and order tracking capabilities further distinguish automated acquisitions systems from their manual counterparts. Acquisitions modules permit the online retrieval of order records by bibliographic and acquisitions-specific parameters, including an order control number, purchase order number, order date, vendor identifier, requester identifier, or fund num¬ber. Retrieved records typically include brief bibliographic data, accompanied by a summary of the order's essential characteristics and status, including ven¬dor and fund information. Operators can usually request more detailed displays, which may include the number of copies ordered, the price, the payment type, the currency type, the vendor's discount, and the specified cancellation date. Vendor files are typically searchable by vendor name or code. Fund records can be retrieved by account number, account name, or, occasionally, account balance.

More than any other library activity, the acquisitions activity relies on printed output for many operations. The acquisitions modules supported by in¬tegrated library systems can produce procurement documents, notices, and re¬ports in a variety of formats. The most important examples of procurement documents are purchase orders and payment documents, including checks and vouchers. As an alternative or complement to printed output, acquisitions mod¬ules increasingly support the electronic transmission of order information to publishers, book jobbers, or other vendors. Claiming of overdue orders has a direct counterpart in the circulation activity. Automated acquisitions systems will print claiming and cancellation notices o vendors at predetermined intervals and arrival notices to requesters when items are received. In some cases, librari¬ans must specify an anticipated arrival date for ordered items; in other cases, the acquisitions module uses performance statistics stored in vendor records to cal¬culate an anticipated arrival date. Most systems can generate lists of overdue or¬ders for operator examination prior to printing claiming notices.

While online file inquiry can be used to quickly determine the status of a given order or fund balance, most acquisitions modules rely on printed reports to provide aggregate financial, statistical, or other information on a regularly scheduled or demand basis. Examples of such reports include fund status sum¬maries, arranged by account number; fund history reports that list closed or¬ders by title within each account; lists of daily accounting transactions; charts of accounts; purchase order lists arranged by title, vendor, fund, workstation, or other parameters; claimed item lists; lists of canceled orders; lists of open or¬ders; lists of orders outstanding longer than a specified period of time; lists of orders received but not invoiced after a specified number of days; vendor ros¬ters, including discounts and performance statistics; lists of newly acquired items arranged by author, title, call number, or other parameters; and lists of or¬dered items on hold for requesters.

Unlike online catalog access and circulation control, where integrated sys¬tems have supplanted other automation options, libraries have several alter¬natives to the acquisitions modules offered by integrated system vendors. Customized software development is, of course, an alternative in every com¬puter application. As noted above, a number of libraries developed customized acquisitions systems during the 1960s and 1970s, and some of those systems re¬main in operation. For the most part, they are implemented on mainframes or minicomputers operated by a university, municipality, corporation, govern¬ment agency, or other organization with which a library is affiliated. Since the 1980s, some small to medium-size libraries have utilized microcomputers and database management software to develop acquisitions systems.

As its principal disadvantage, customized software development can be time-consuming and expensive. Libraries developing customized acquisitions systems run the risk of significant delays in implementation and substantial, often unanticipated costs. Post implementation costs will likewise be incurred for software modification and other maintenance to address changing application requirements. Consequently, customized development is often viewed as an im¬plementation alternative of last resort, to be considered only in those situations where other approaches to automated acquisitions are clearly unacceptable.

Among the bibliographic utilities, RLIN offers an acquisitions subsystem that libraries can access on a time-sharing basis. The RLIN acquisition system is integrated with bibliographic records and supports a variety of program activi¬ties and situations, including selection decisions, single item procurement, direct extensions, and standing orders. Acquisitions records can be created by adding in-process information to the RLIN database. RLIN's central bibliographic files can be used for preorder verification and derivation of bibliographic data for a given title. Order numbers, prices, supplier identifiers, claiming intervals, ship¬ping instructions, fund numbers, encumbrances, and other acquisitions-related information are entered in fields provided for that purpose. To simplify input procedures, codes can be defined for suppliers' names and addresses, selector and requester names, shipping and billing locations, fund codes, and other fre¬quently repeated data values. Default data values can also be specified. The RLIN acquisitions module will track the status of orders through receipt and cat¬aloging. Claims can be generated automatically or on demand. Printed output includes purchase orders, claims, cancellation notices, transaction logs, and fiscal reports.

Among its advantages, the RLIN acquisitions subsystem can draw on a large cataloging database for bibliographic verification. The ability to transfer biblio¬graphic data from cataloging records to order records can facilitate order prepa¬ration, while the online availability of holdings information for other libraries is useful in cooperative collection development. The use of a single system for ac¬quisitions and cataloging also simplifies staff training requirements and pro¬motes the integration of technical services operations. As disadvantages, the RLIN acquisitions subsystem may not meet local accounting or auditing re¬quirements. Compared to acquisitions modules supported by integrated library systems, it provides less flexibility for production of information or statistical re¬ports, although certain types of financial reports, such as fund activity and his¬tory reports, are offered. Finally, the RUN acquisitions subsystem does not offer check-writing capabilities.

As an alternative, various booksellers offer software and. time-sharing ser¬vices for online ordering of library materials. Such systems not only automate the creation and maintenance of acquisitions records but also transmit orders to the bookseller electronically. Because their developers are in the business of selling books and other library materials, their online ordering systems are designed to expedite delivery to customers, although they may also help libraries manage their acquisitions operations by maintaining records and preparing purchase or¬ders for materials to be acquired from other vendors.

Online ordering systems provide time-shared access to booksellers' com¬puter systems. In the manner of the bibliographic utilities, formatted screens simplify order entry. File maintenance and fund accounting are performed auto¬matically, and online inquiries about the status of a given order are supported. Libraries receive periodic reports that reflect their acquisitions and fund ac¬counting activities. Because the booksellers' databases contain inventory infor¬mation, the availability of a given item can be immediately determined.

Serials Management Modules

As previously defined, the term serials denotes publications that are issued in successive parts on a recurring basis, usually, but not necessarily, at regularly scheduled intervals. In additional to scholarly journals, popular magazines, and other periodical publications, the term encompasses newspapers, proceed¬ings and transactions of professional societies, newsletters, and numbered monographic series. Unlike multivolume books, which may likewise be issued in successive parts, serials are characteristically open-ended. The publication of successive issues is expected to continue indefinitely, although external circum¬stances may force the eventual suspension of a given serial.

Since the 1960s, library interest in automated serials processing systems has been motivated by the same factors that have encouraged the development of automated approaches to acquisitions, circulation control, and cataloging-the desire for improved operating efficiencies and cost reduction. The cost of ac¬quiring and maintaining serial publications represents a significant percentage of many libraries' budgets. In technical, medical, and business libraries, for ex¬ample, serials typically constitute the major collection component, and ordering, claiming, binding, and related paperwork processing tasks can require many hours of labor. Even in academic and public libraries, where the ratio of serial to monographic publications is usually lower, serials departments may have large staffs. At a time of simultaneously rising subscription prices and re¬duced library budgets, operational economies are essential if a library's serials collection is to be maintained at its present level, let alone increased.

Although library interest is high and cost-reduction potential considerable, the widespread development and implementation of effective computer-based serials processing systems are complicated by certain peculiarities of serial publications. Circulation control, cataloging, and, to a lesser extent, acquisitions are characterized by well-defined, predictable work steps that are performed in a predetermined sequence on a regular basis. Serials processing tasks, on the other hand, are distinguished by a variety of exceptional circumstances that must be anticipated in the design of any automated system. While bibliographic infor¬mation pertaining to monographs remains relatively stable once entered into a library catalog or other computer database, it is a rare serials record that will not eventually require some change in title, publisher, issuing agent, frequency of publication} numbering sequence, or other attributes. Such changes can render a computer-generated union list obsolete, complicate the development of an automated claims processing system, or alter previously established binding practices for a given serial. While changes and exceptional conditions can be handled by computer programs, they increase the complexity of automated se¬rials implementations.

Over the past two decades, a number of libraries and vendors have developed systems that automate one or more aspects of serials processing, includ¬ing the production of serials holdings lists for single or multiple libraries and ordering, check-in claiming, and routing of received issues to designated per¬sons. AB with monographic acquisitions, various academic, public, and special libraries developed customized serials control systems during the 1960s and 1970s. Examples include the University of California at San Diego, where an automated serials check-in system was implemented as early as 1961; the Wash¬ington University School of Medicine, which introduced its much-publicized PHILSOM serials management system in 1962; the UCLA Biomedical Library; the San Francisco Public Library; the University of Massachusetts; the Univer¬sity of Arizona; the University of Washington; Brigham Young University; and the Universite Laval in Quebec, one of the first libraries to implement an online serials control system.

During the 1970s and 1980s, several vendors introduced special-purpose serials control systems and prewritten software. One of the most widely in¬stalled examples, the INNOVACQ system from Innovative Interfaces, was mentioned in the preceding discussion of automated acquisitions. Since the 19OOs, most integrated system vendors have offered serials control modules as optional application components. As with their acquisitions counterparts, some early versions were not as well developed as other system components, but se¬rials control modules have improved steadily and significantly in recent years. For libraries that operate an integrated system for cataloging and circulation' control, they usually offer the simplest, least expensive method of automating serials management.

The serials control modules supported by integrated systems are designed to manage magazines, journals, newspapers, monographic series, and other materials that libraries receive on a continuing basis, whether at regular inter¬vals or as irregular supplements or special publications. With some integrated library systems, ordering and cancellation of serial subscriptions, as well as claims for missing or damaged issues, are handled by the acquisitions module, while the serials control module supports check-in, routing, and binding of received issues. Alternatively, a serials control module may handle all aspects of serials management, including procurement and claiming. That approach does not require implementation of the acquisitions module as a: precondition for serials control. With a few integrated systems, acquisitions and serials man¬agement are combined in a single module.

In any case, order and renewal of serial subscriptions are typically per¬formed online, although purchase orders and renewal orders may be printed in batches. Received issues are checked in at computer workstations by retrieving the appropriate serials record and modifying designated holdings fields. The most flexible systems simplify data entry by displaying information about an expected issue, including the volume and issue number, cover date, and num¬ber of copies anticipated. The workstation operator simply modifies those data elements requiring correction. When all modifications are completed, the new record enters the serials data file. If serial publications circulate, item records can be created for individual issues.

Serials management modules keep track of publication patterns and will pre¬dict the receipt of specific issues. The library specifies the periodicity for succes¬sive issues and the number of days after the anticipated receipt when claiming notices are to be sent. The most flexible predictive algorithms can accommodate a variety of publication patterns. When overdue issues are to be claimed, the sys¬tem searches the serials data file for issues that have not arrived according to their previously defined frequencies. Claiming notices for missing issues may be pro¬duced manually or automatically. Query notices, rather than claims, can be gen¬erated for irregular publications. Based on information stored in serials holdings records, most serials control systems will also prepare bindery orders, print in¬struction slips, check in items on their return from binding, and print claiming notices for items that have not been returned on schedule.

Some integrated systems support electronic transmission of subscription orders and claims for missing issues. Serials holdings information is included in the library's online catalog, which is updated automatically when new' issues are received. Typical management and statistical reports include serials cata¬logs and union lists, on-order lists, lists of issues received, lists of issues past due, lists of claiming notices sent, and vendor performance reports. Some inte¬grated systems will also print routing slips for received issues and pull-slips for items to be sent to the bindery.

As an alternative to the serials management modules supported by inte¬grated library systems, several serials subscription services offer computerized, serials management capabilities on a time-sharing basis. Such services provide online- access to large databases of bibliographic records for serials titles and monographic serials, including the serials segment of the LC MARC database. Records can be retrieved by various parameters, including International Stan¬dard Serial Number (ISSN), title, and publisher. Serials management capabili¬ties include online ordering, subscription renewal, and claiming of missing issues; confirmation of transactions via electronic mail; online access to account balances, payment histories, and other financial information; and electronic transfer of serials records to a library's local automation system.

An integrated library system is a computer-based information system that uses a ' single bibliographic database and a set of interrelated application programs to automate multiple library operations. Since the mid-1980s, integrated library sys¬tems have steadily supplanted custom-developed software and single-purpose library automation products that were designed for specific applications, such as circulation control or acquisitions. Integrated systems are offered by several dozen vendors for various types of computers. Products are available for li¬braries of all types and sizes.

Most integrated library systems are modular in design. Vendors offer suites of interrelated programs that automate specific library operations. Core func¬tionality is provided by certain application modules that are required in all in¬stallations, while optional modules, which automate specific operations, can usually be added at any time. In most cases, the standard system modules sup¬port cataloging, an online public access catalog COPAC), and circulation control. Acquisitions and serials management are popular but usually optional addi¬tions to a basic system configuration. Other optional application modules, not necessarily offered by all vendors, include media booking, a community bul¬letin board, community information and referral, homebound access, reference databases, reserve room capability, a report generator, Internet access, elec¬tronic imaging, full-text retrieval, and electronic mail capabilities.

All integrated systems support cataloging as a core application module that permits the creation, updating, and management of a library's bibliographic database. All integrated systems support MARC records for those types of library materials for which USMARC formats have been developed. Most integrated systems permit partial MARC records that employ the USMARC format with some fields left blank. Some systems also support non-MARC formats with library-defined fields. Typically MARC and non-MARC records can coexist within a library's database. Cataloging records can be key-entered or transferred from machine-readable sources, either electronically or on magnetic tapes or diskettes. Depending on the system, authority control may be incorporated into the cataloging module or offered as a separately purchased component.

The replacement of a card catalog by an online public access catalog is often the principal motive for implementing an integrated library system. Suitable for novice users who receive little or no training in information retrieval, most OPAC modules are menu-driven, although a command-driven mode may be provided for experienced users or library staff members. Some systems provide graphical user interfaces with pulldown menus, dialog boxes, mouse opera¬tions, and other features that simplify the entry of search commands and for¬matting of retrieved information. Several vendors have implemented special OPAC interfaces for children. As a recent development that is likely to be widely adopted, some integrated library systems support OPAC searches by microcomputers equipped with popular Web browsers such as Netscape Navi¬gator or Microsoft Internet Explorer.

All OPAC modules- permit searches by author, title, and subject, thereby replicating the traditional retrieval functionality of card catalogs. Typically, a user specifies a field to be searched and a name, subject heading, or other value to be matched. Other search parameters, selectively supported, include a classi¬fication number, Library of Congress Card Number, International Standard Book Number, publisher, publication date, country of publication, language of publication, and media type. For maximum flexibility, some integrated systems allow any library-designated field within bibliographic records to be indexed for retrieval purposes. Special retrieval capabilities include keyword searching, root-word searching, and Boolean operations.

To expand their functionality, integrated library systems increasingly incor¬porate external information resources, such as locally mounted reference data¬bases and CD-ROM information products, into OPAC searches. Some systems also provide Internet access from OPAC terminals. If an integrated system sup¬ports the NISO Z39.50 standard, OPAC users can search the compliant com¬puter systems of other vendors. With some systems an optional information and referral module gives OPAC users access to information about community groups and local events.

Drawing on three decades of library experience with computer-based cir¬culation management, the circulation control modules supported by integrated systems are very well developed and highly parameterized-that is, they allow libraries to specify the conditions under which items will be circulated; file in¬quiries made, and printed output generated. All circulation control modules perform check-out, check-in, and renewal operations in real time. Circulation procedures are straightforward. Barcodes are typically used to simplify the entry of item and borrower identifiers. Circulation transactions are blocked when borrower delinquencies or other exceptional conditions are encountered. Holds can be placed at the title or copy level. Most circulation control modules provide a selection of preconfigured operational and statistical reports that ad¬dress a broad range of library requirements.

Designed specifically for academic and school libraries, reserve room capa¬bilities maintain circulation records that are linked to courses and instructors. Records for items on reserve can be retrieved through OPAC searches or by course number or instructor. In addition to managing items from a library's own holdings, most reserve room modules can create brief catalog records for photocopies and instructors' personal copies. Some integrated library systems support a media booking module that provides circulation control for video¬tapes, films, and other audiovisual media, as well as equipment and facilities that are subject to advance reservations.

The acquisitions modules supported by integrated library systems can ac¬commodate a broad range of procurement transactions, including new orders, multi-copy and multivolume orders, standing orders, subscriptions, blanket or¬ders, approval plans, depository items, rental books, items obtained through memberships, prepaid orders, gifts, and exchanges. Because acquisitions mod¬ules are fully integrated with other system components, records for ordered items are included in a library's bibliographic database, and on-order status is reflected in the online public access catalog.

The serials control modules supported by integrated systems are designed to manage magazines, journals, newspapers, monographic series, and other materials that libraries receive on a continuing basis, whether at regular inter vals or as irregular supplements or special publications. Capabilities include online ordering and renewal of subscriptions, check-in of received issues, claiming of missing or damaged issues, and control of bindery orders. Some integrated systems support electronic transmission, of subscription orders and claims. Serials holdings information is included in the library's online catalog, which is updated automatically when new issues are received.