Library Technology Guides

Document Repository


Volume 5 Number 10 (October 1985)

Turnkey training provisions

As hardware capabilities increase and costs decrease, more turnkey automated library system vendors are offering multifunction systems suitable for small and medium sized libraries priced at or below $100,000. All vendors face a major difficulty in such relatively small sales: the need to reduce the expensive human resources devoted to client support.

A common approach to cost reduction and containment is to limit the amount of support provided as part of the turnkey package. The turnkey concept of single- vendor provision and responsibility for all system hardware, software, installation, maintenance, documentation and training is still accepted. However, there appears to be a trend for vendors of systems of all sizes to limit the amount of library staff training in system operation and use that is included in the basic system purchase price. The typical training commitment for a small system appears to be five days. Additional training is generally made available as an optional extra, at a fixed price of up to $1,000 a day or a daily rate of $500 plus reimbursable expenses for travel and accommodations.

Most observers would agree that this is a reasonable response to the cost containment problem, and that it is in the interests of a library--particularly one with a small automation budget--to be able to choose where it places its available dollars. However, the practice does raise the issue of how much training is essential and what steps can be taken to ensure that a library obtains the maximum benefit for those automation dollars that are assigned to training. The first issue is that of how much training is required for efficient system operation and use. The usual practice for moderately-sized systems is to provide two days of training in general system operation--start-up, routine back-up and maintenance, problem solving, and the like--and one day each for the major functions to be implemented.

Most libraries now contract for multifunction systems with the current or future capability to support data base creation and maintenance, acquisitions, circulation, patron access catalog, and serials control. Some also require media booking capabilities and access to files of abstracting and indexing information. Many libraries choose to implement the various functions in phases, often beginning with data base creation and maintenance and circulation followed by patron access catalog, acquisitions, and finally, serials control, at three month intervals. For such an implementation, two days of training in general systems operation would be followed by an additional two days of training in the chosen functions--one for data base creation and maintenance, the other for circulation control. Training in other functions would be undertaken when the appropriate software had been mounted and was ready to be implemented.

The number of staff to be trained in a session will vary according to the size of the system and the library in which it is being installed. Current industry practices suggest that groups of up to fifteen people can be adequately instructed at one time. However, some vendors seek to limit the number of attendees at a training session to as few as eight. A library pressed to limit the size of its training group to the lower figure should inquire about the reason for the limitation and examine the financial implications of accepting this limit.

Understandably, many libraries have a natural concern with the initial implementation of a system or function. The main object is to make the system operational. And, when staff training is being considered, this is the main focus. A library's plans for staff training should, however, be more farsighted. Certainly the system has to be made operational; on the other hand, continuity of support must also be provided. In large libraries more staff will need to be trained than can reasonably be accommodated within a vendor's training program; and staff turnover will occur in all libraries. Thus a library must be prepared to provide its own ongoing instruction and training. Training groups should be created with the intention of accommodating not only immediate operational needs, but also supplying the library with a cadre of in-house trainers. A library cannot expect a vendor to provide ongoing training of new staff within its basic cost framework. While customers might reasonably expect vendor training of staff in system enhancements, this is usually accommodated within the monthly software maintenance fee.

There are no industry-specific measures for evaluating the effectiveness of vendor training. As the amount of training provided as part of a system procurement decreases, the importance of performance measures increases. This is particularly so, given that most automated system training is conducted in a face-to-face classroom environment; the Editors do not know of any vendors that have developed substantial packages for computer- assisted or interactive online training.

The most readily accessible yard-stick for assessment of vendor training capabilities is the vendor's training documentation. A library proceeding to a competitive procurement without the assistance of consultants familiar with this aspect of the vendors' systems is well advised to require the submission of one copy of a full set of training materials as part of the vendor response to its RFP.

RFPs for the procurement of turnkey automated library systems--and those for the provision of supported software for internal library automation--routinely include specific training requirements similar to the following:

  • Personnel (not to exceed N) selected by the library shall be given a system overview and a demonstration of each basic function in the initial installation.
  • Similar sessions shall be scheduled as each subsystem or module is brought up.
  • For the initial installation, up to N staff members shall be trained so that they are capable of performing the following without assistance:

    • all operator functions
    • all supervisory override functions
    • recognize the common causes of system failure
    • implement the remedy for each type of system failure
    • follow oral instructions given by telephone for the correction of system problems
    • know and perform all elements of library system backup and preventive maintenance of the system.

    Similar training sessions shall be provided as each of the other sub-systems or modules is brought up.

  • Future operators who are not trained by the vendor will be trained by the initial trainees.

    Such requirements do not provide training performance measures whereby the library can assess the effectiveness of vendor training. The Editors know of no widely accepted mechanisms for such assessment but offer the wording in the following specification elements for consideration. As in all performance and penalty clauses, the objective is to increase the probability that the library and its chosen vendor will work together to achieve their joint objective of efficient and effective implementation of library automation.

  • As part of its bid, the vendor shall include a detailed outline of the functional objectives of each module of its training program.
  • Prior to the commencement of training, the vendor shall submit to the library a post-instruction evaluation test designed to measure the achievement of the function objectives detailed in its RFP. The test will be used as the basis for assessing the efficiency of the vendor's training program.
  • The library and the vendor shall negotiate: the content of the test; the "pass" score for individual trainees (which may entail a required level of accurate performance in routine procedures and the resolution of common problems rather than a "50 percent" score on a multiple choice question format); and the number of trainees required to reach that standard.
  • In the event that the agreed number of trainees do not achieve pass scores, the vendor shall provide the additional training required to meet the agreed levels of performance. The vendor shall provide such additional training at its expense and each additional session shall be of at least one half day in duration.
  • The testing procedure shall be followed for general system operation and maintenance, and for each software module.

A PC Unicorn

Sirsi Corporation, vendor of the Unicorn Collection Management System for local library automation, has announced the availability of its software on the IBM PC AT microcomputer. Sirsi indicates that the AT can be configured with sufficient storage to support a data base of up to 70,000 full MARC records or 140,000 brief bibliographic records. All Unicorn modules are available for the AT: bibliographic and inventory control, circulation, patron access catalog, academic reserves, acquisitions, serials control, electronic mail, and the bulletin board feature.

[Contact: Sirsi Corporation, 8106-B South Memorial Parkway, Huntsville, AL 35802, (205) 881-2140.]

ALIS II users load release 5.0

As of the ALA Annual meeting the vast majority of Data Phase ALIS II libraries were using a variety of software releases ranging from 4.10 through 4.13B despite the fact that these releases were not the latest available. Data Phase has, since that time, persuaded all except five sites to install Release 5.0, bringing these libraries not only to the latest software version, but to the same version. This lays the groundwork for completion of acceptance testing, says Bruce Gelb, Executive Vice President of Data Phase.

Release 5.0.1 is scheduled for September 1985. It will be a maintenance release that will remedy problems reported in the software. Although 5.0.1 is designated as a maintenance release, it will include the software for auxiliary circulation using the Techtran and PortaWand devices.

Despite these signs of progress, a number of libraries continue to hold back on acceptance because they believe that their systems are under-configured either in CPU or disk storage. Individual negotiations will probably be necessary to resolve these issues.

PAC survey revisited

Given the increasing interest in online patron access catalogs, the Editors urge that librarians read or reread the report of the Council on Library Resources sponsored study of users and non-users of online catalogs: Using Online Catalogs: a Nationwide Survey. Using Online Catalogs (ISBN 0-918212-76-6) was commercially published by Neal-Schuman Publishers, Inc. in 1983. The publication integrates and analyzes the findings of four separate projects sponsored by the Council. In the projects, survey questionnaires were completed by more than 8,000 online catalog users and nearly 4,000 non-users. The study population was located in 29 libraries. The libraries utilized 16 different online catalog systems.

While interesting, many of the survey findings are of only limited utility to the majority of libraries as most seek to meet their online patron access catalog needs through the purchase of an existing turnkey system or supported software package, rather than by designing a system de novo. However, there are some findings that have direct implications for library managers.

Patrons in general had no problem accepting an online catalog, even if they themselves did not use it. Online subject searching capabilities are heavily used when provided. (A finding which emphasizes, yet again, the importance to libraries contemplating eventual automation of making sure that their machine-readable bibliographic records are full-length, including subject information such as subject headings and classification data.) Other research emphasizes the importance of providing access to full title information, not just that coded as subfield "a" in a MARC record. In subject searching, keyword and term searching are heavily used; a system should therefore support this capability--on at least subject and title fields. Users also indicated a desire to have access to listings of the subject terms used in the system. (In a paperless automated environment, this presupposes a system with the capability to display lists of subject headings.)

Users find menu-driven systems easy to use, but soon become impatient with the step-by-step menu approach. Ideally, a system should continue the benefits of a menu approach with command capability. Good online instruction and help capabilities are useful in assisting users to learn the system and cope with difficulties. However, such aids are sometimes confusing. The limited online assistance needs to be augmented with individual help from public services staff. In system selection it is wise to consider both the extent of the on-line assistance and the extent to which help and instruction screens can be amended and changed by the library to reflect user experience and local terminology. Ideally, a system would provide printed commands for the more sophisticated user and summary charts at the terminals to assist others in remembering how to use the system.

Variance analysis shows that systems in which public access terminals are equipped with printers are perceived by users as being easier to use than systems which lack this capability. Therefore, it would be wise to plan on making provision for side printers on public access terminals, noting, however, that they should be simple to use and properly maintained since patrons sometimes encounter problems with their use.

Many users learn about the patron access catalog merely by encountering the terminals in the library. If a library wants its patrons to use the public access catalog, the terminals should be conspicuously located. Also, the location of terminals need not be limited to the traditional card catalog area. Presumably, a wider distribution of terminals will be more effective in encouraging their greater utilization. Finally, according to the survey, terminal hardware was not a problem area. However, adequate space should be provided to accommodate the books and papers a user brings to the terminal, and for note taking.

H-P Spectrum's implications

Hewlett-Packard is remaining close-mouthed about its new Spectrum line of computers. The Spectrum is expected to replace the HP 3000 series, the CPU on which VTLS and INLEX (formerly Electric Memory's EMILS software see LSN Vol. 5, No. 8) software operates. The company hints that the performance levels of the Spectrum will be up to five times greater than that of the largest HP 3000.

The company appears to be particularly reluctant to talk about a migration path from the HP 3000 to Spectrum. While the Spectrum could be operated in emulation mode, performance would be substantially degraded. Some programming commands and techniques would probably have to be changed because they are dependent on the architecture of the HP 3000. In the native mode, Spectrum will not support SPL, the programming language used by VTLS. In addition, the Image 3000 data base management system used by VmS also will not work on the Spectrum.

Users of older models of the HP 3000, such as series 2, 3, 30, and 33 will have greater difficulty in making the transition than users of series 48, 58, and 68.

In order for libraries to be able to use VTLS software on the new HP machines, it appears, at this point, that Virginia Tech and the users of VTLS will have to commit to substantial reprogramming and be prepared for some conversion headaches. Prospective purchasers of VTLS software should therefore ascertain whether they have a Spectrum alternative.

INLEX/EMILS users are not expected to face the same problems since the software has been written in Pascal to operate on both the HP 3000 and the Spectrum.

IBM PC discounting

PC sales have tapered off and the market is becoming increasingly competitive. IBM PCs, purchased through dealers, are now being discounted as much as 33 percent below list. These big discounts are not available directly from IBM for small orders. (Those buying direct from IBM must purchase 150 machines to qualify for a 20 percent discount.) However, most dealers routinely offer a 20 percent discount to an educational institution or government agency. And buyers with quantity purchasing agreements can expect discounts averaging 30 percent.

MICRO VAX III announced

Only five months after introducing the Micro VAX Il--and before the library automation industry has absorbed the new Digital supermicro into its product lines--Digital has announced the Micro VAX III. The III is in the power range of the VAX 780, but the price is under $20,000. The Micro VAX II is in the power range of the VAX 750 and costs just over $15,000.

The two Micro VAX products will make it possible for library automation vendors such as Data Research Associates and Universal Library System to offer very powerful systems beginning as low as $100,000.

Fax market booming

The recent drop in prices of digital facsimile equipment has spurred sales. Sales are up 55 percent over 1984, with a total of 139,500 units expected to be installed during 1985. The average price per machine is now $4,000, as compared with $4,300 a year ago, but the more im-portant development is the reduction in price on many machines to less than $3,000.

Xerox, which was the dominant com-pany in the industry five years ago with a 50+ percent market share, now trails industry leader Rapicom, which has a 17 percent market share. Other ma-jor suppliers are Panafax, Murata, and 3M. The leader at the upper end of the market is NEC, the company which provides Federal Express the hardware for Zap Mail. Top-of-the-line NEC machines cost $10,000 and more but they are Group IV devices (adhering to the latest international standards). These machines can transmit a page in two to six seconds if special telecommunications lines are used. Reso-lution is equal to a high quality office photocopier.

Fax use in libraries continues to be primarily experimental, but industry representatives believe that the library market will grow rapidly at the end of the 1980s.

MCI increases international reach

MCI now offers international long distance service to 31 countries and is negotiating to provide service to several more, including Denmark, France, and Japan. The services include voice, data, and telex communication.

While MCI is only one of several special common carriers which is going international, it is currently the most aggressive. MCI expects to have services to countries representing 80 percent of the international calling volume by the end of 1985.

AT&T may consolidate marketing

AT&T customers may soon be able to arrange both long distance service and equipment purchase or lease with a single phone call. Under an FCC proposal expected to be approved and implemented within the next two months, AT&T would no longer be required to separate the marketing of long distance services and equipment. AT&T would be required only to separate the bookkeeping for the two types of business so that the profits from long distance services could not be used to subsidize the equipment business.

Since deregulation, the separation of functions has not only required that customers make two phone calls, but has caused serious problems in coordinating the delivery and installation of equipment and communications services. The major reason for the FCC's relaxation of its former proscription against a single AT&T marketing organization is that AT&T's share of the long distance market has shrunk from over 90 percent two years ago to its present 64 percent. In any case, if it wishes to hold market share in the long distance industry, AT&T cannot afford to cross-subsidize other segments of its operations.

Make your own CD-ROM

A number of libraries and library consortia are beginning to "publish" catalogs or special data bases on CD-ROM, the computer storage medium based on the same technology as the highly successful compact audio discs. A single CD-ROM can accommodate approximately 550 MB of data. In addition to the specialized bibliographic services provided by LSSI (currently on videodisc rather than CD-ROM) and the Library Corporation, described in LSN Vol. V, No. 2, the following firms provide a range of services including assistance in data preparation, mastering, and production:

  • Denon America, Inc.
    27 Law Drive
    Fairfield, NJ 07006
    (201) 575-7810
  • LaserData
    One Kendall Square, Bldg. 200
    Cambridge, MA 02139
    (617) 494-4900
  • Reference Technology Inc.
    1832 N. 55th Street
    Boulder, Co 80301
    (303) 449-4995
  • 3M
    3M Center Building 223-5N-01
    St. Paul, MN 55144
    (612) 733-5211

Data preparation charges are usually $7,500-$l0,000 per disk side, plus $250 per reel of input tape after the first reel. Disk mastering and replication is approximately $200-$250 per disk for minimum quantities of 50, but only $24- $30 per disk for quantities over 1,000.

Only 3M does its own mastering and replication, CBS also has a facility, but it accepts only very large orders. The reason why there are fewer than 10 plants in operation or under construction worldwide is that the cost is $20-$30 million per facility.

A typical system to retrieve data on a CD-ROM disk includes a disk drive, controller, personal computer, and software. The cost is presently from $4,500 to $7,000. If an IBM PC is already available, the cost for the additional hardware and software is $3,000 to $4,000. Few interfaces are available for other personal computers.

OSI not a panacea for interfacing

In recent months the literature of library automation has included considerable discussion of the Open Systems Interconnect (OSI) Reference Model--see, for instance, LSN Vol. IV, No. 10. While the development of standards within each layer of the model can be expected to facilitate intersystem linkage among computer systems using different hardware and software, these standards will not, by themselves, automatically bestow the ability to connect different library automation systems.

The implementation of OSI standards will provide a smooth, well-lit highway between the "doors" of the different systems, but system vendors and users will still have to follow certain internal standards and practices to permit the "key" to the door of one system to unlock the doors of other, different, systems. At a minimum, those involved in library automation will need to develop a common command language whereby an instruction such as "AU ABCDEFG" is universally recognized as a command to search for the occurrence of "ABCDEFG" as an author in a bibliographic record. Each different system vendor will have to develop software that will permit this common command to be translated into the specific author search command required by the vendor's system. Such an interface will need to be bidirectional--to be able to process both incoming and outgoing commands.

When and if such capabilities are implemented, total interfacing of heterogeneous automated library systems will still not be realized. Some libraries will have set up their systems to perform author searches only on the contents of the lxx fields of the MARC record, others will have mandated that the contents of author added entry fields also be accessed by such searches. System linkage will also be affected by the internal data base standards adopted by different libraries: the contents of author added entry fields cannot be searched if a library has chosen to set up brief machine-readable records in which the only author data is that recorded in the 1xx fields.

In order to be as well positioned as possible to take advantage of future developments in automated library systems linkage, libraries developing data bases of machine-readable records and those planning for current and future automation should give careful consideration to adopting national and international standards for both system interfacing and machine-readable record creation. This requires an internal commitment to full-length bibliographic records in the MARC II format, and the specification of adherence to emerging OSI standards in automated system development.

For the immediate future, a library's approach to interfacing its local automated library system to standalone systems in other libraries is likely to be determined by the purpose of the linkage. Libraries which require such linkages for staff use have the greatest number of options. These include "vanilla flavored" terminal to computer linkage via dial access or dedicated lines. When no special interfacing software is used the terminal operator on the local system must manipulate the data base of the remote system using the searching language and commands of the remote system. Networking software developed by companies such as MSI and LSSI (see LSN Vol. V, No. 9) provides varying degrees of support to allow the use of local system commands or a common network language to search and manipulate remote systems which differ from that of the local library. However, such software still requires a level of knowledge and expertise that can only be expected from a library staff member experienced in the use of the interfaces.

Libraries that wish to provide their users with full and direct access to the automated systems of other libraries have two options, only one of which utilizes interfacing. Direct access can be provided by a group of libraries with mutual interest in reciprocal access joining together to select and implement a single shared automated system. This eliminates the need for interfacing. However, a number of factors, particularly telecommunications costs, can present major barriers to such system sharing. When two libraries at a considerable distance from each other wish to provide their patrons with easy access to each other's systems, the only viable solution is for the libraries to agree to implement common data base standards, to purchase identical systems from one vendor, to cooperate in system implementation, and to provide telecommunications linkages between the systems.


Publication Information


PublisherLibrary Systems Newsletter was published by the American Library Association.
Editor-in-Chief:Howard S. White
Contributing Editor:Richard W. Boss
ISSN:0277-0288
Publication Period1981-2000
Business modelAvailable on Library Technology Guides with permission of the American Library Association.