The Chinese Academy of Sciences comprises 140 institutes of basic research and high-tech development which are located in the main cities of China. As part of the Academy, this information system includes 5 regional libraries (now also called documentation centers) in Beijing, Lanzhou, Shanghai, Chengdu and Wuhan, as well as 140 libraries and information divisions belonging to research institutes. To better serve the scientists and technical personnel, many of the libraries and information centers in the Academy began to use computers in the 1980's to create their own publicly accessible databases, which include abstracts and indexes in the pure and applied sciences. But we believe that the creation of these databases is not our main task. What we need more urgently is the creation of OPACS and online union information systems. Since there are no commercial systems available in China, the five regional libraries and their institute libraries began in 1987 to use availa-ble funds and technologies to create their own systems. At first, this labor-intensive effort was undertaken without proper collaboration. As a result, each regional library's integrated library system has software and data incompatible with the others. So now we face a number of problems that need to be solved. This paper will review the integrated library systems that are being used, those that will be replaced, and those that will be further enhanced. It will conclude with some considerations for future system development in China and for library automation in other develop-ing countries.

2. INTEGRATED LIBRARY SYSTEM REVIEW

In the last 5 years, 4 integrated library systems--what every Chinese library seeks--have been completed, and three are still under construction. The seven systems can be categorized into three types: first, HP3000, VAX 11/780, or MicroVAX II based systems; second, 286's, 386's, and PC's with LAN-based systems; third, single microcomputer systems. In the first category, DILAS (Distributed Integrated Library Automated System) started to be designed and created in 1987 by the Library of the Chinese Academy of Sciences in Beijing. This system, based on a MicroVAX II and four AST 386's with UNIX, XENIX, INFORBASE and C as programming tools, was completed in 1991 after four years of work, and library routines were automated. INFORBASE is a specific implementation of the INFORMIX database management system, which can handle variable fields and subfields. The serial control subsystem, Western-language-monograph cataloging by means of Bibliofile, the circulation subsystem, etc. are run on AST 386's used as workstations (The Chinese Academy of SciencesLibrary, 1991). All the records are shared in these workstations and in Micro-VAX II as data is transferred either by diskettes or by communication software. Since circulation on the AST 386 is not linked with the MicroVAX II, there is no way of knowing whether the needed item has been lent or is on the shelf, if one searches only the main database on MicroVAX II. With only 9MB memory and 1000MB disk space, the MicroVAX II is limited in speed and capacity. Furthermore, this system's hardware and software do not allow entry into the Educational Network with the Beijing University Library and Qinghua University Library systems, and so will be replaced by Taiwan TRANSTECH System with a network function, which runs on a Motorola computer with UNIX. Meanwhile the library will face a transition period until it migrates from DILAS and trans-fers all the records accumulated in the last few years to the coming TRANSTECH system.

Other systems on VAX series computers are SIOCL and LDICILS. SIOCL, developed by the Shanghai Institute of Organic Chemistry Library, was first completed and used in that library. With FORTRAN 77 and TDMS, a programming tool and a screen-form tool, this system ran on a VAX 11/780 minicomputer, but now runs on a MicroVAX 3300. The acquisition and cataloging sub-systems went into operation with this system first. OPAC was possible after circulation was added. The serials control subsystem was the last to be included, as MARC records were transferred into it by tape from the union serials database created by the Beijing Library of the Academy. All of the institute library's collection can be accessed on OPAC (Wang, 1992). Another system, LDICILS, is now being developed by Lanzhou Library on a MicroVAX II with a DBMS and C as a programming language and database creation tool. Cataloging is now possible and other functions are underway.

Among the first category, SDICIS (Shanghai Documentation and Information Center Information System) is a good integrated library system designed by the Shanghai Library (1992). It passed its beta test in May 1992 and won high praise for its highly integrated functions. Running on a HP925LX and a HP3000 computer with MINISIS as software, it has 200,000 MARC monograph and serial records both in Chinese and in Western languages, which cover most of the library's collection from the past 20 years. This system will be further enhanced by adding authority control, and it will link SIOCL systems in the region to create a bibliographic record center.

In the second category, 286's, 386's and PC's are connected on 3+ COM with a microcomputer as a special server. This LAN-based integrated library system was prevalent in China's libraries in the late 1980's, the Wuhan Library and the Chengdu Library of the Academy being prime examples. The Wuhan Library's integrated system was completed with Quick BASIC. The Chengdu Library intentionally took its time building its system with 3+ COM LAN because its computer staff thought this LAN was not stable and hence its database was at risk. Now both libraries face the shift from 3+ COM to NOVELL.

Single-microcomputer-based library systems using dBASE are the third category, which are not actually integrated library systems, since such one-user operational systems handle only one library routine at a time while others have to wait. The Shengyang Institute of Computer Technology Library and the Shanxi Astronomical Observatory Library developed their own systems to deal with library work routines. Many small libraries belonging to the research institutes in the Academy use this kind of system. The drawback to such a system is that it cannot handle collections larger than 100,000 items.

3. SERIOUS ISSUES FACING THE INTEGRATED LIBRARY SYSTEMS

As we can see from the system development of the past years, library automation in the Academy is moving down a treacherous road. When we started to design and build our library systems, there were no Chinese MARC records, which became available only in the late 1980's. Because networks and fully integrated systems were not commercially available, individual libraries had to design their own integrated systems. As a result, we have seven systems that do not communicate with each other, and librarians have to do original cataloging for their own institutions without sharing Chinese MARC records. All this, because these libraries did not properly collaborate from the start. If librarians and computer staff in the libraries of the Academy cooperated in designing and building one or two integrated systems, significant funds and time would be saved, and hardware, software and data would be compatible, and communication would be easier.

In the process of system development and maintenance, we discovered that we had neglected the user environment in terms of planning, system analysis and user instruction. Thus, some systems need to be modified to better serve users and accommodate library work routines. Some systems have problems in integrating their subsystems. All the systems have had to be modified after their startups, which could have been avoided, if we had done careful system analysis and design first. And all systems depend on maintenance support from the technical personnel who designed and built them.

Maintenance personnel are important in the promotion and use of the integrated library systems. When such personnel are moved to other institutions, their systems may face a crisis. Consider what happened to DILAS. Three leading computer staff who designed the system left the library, which made further development of the system difficult.

Another issue that we face is communication among systems, that is, one system's catalog records being read by another system via electronic media like tapes or diskettes. So each system needs an interface so that Chinese MARC records can be communicated. This avoids the duplication of effort in cataloging. Now, except for the SIOCL and SDICIS systems, which have over 100,000 records each, the systems all have a limited number of records. To speed up database building, the systems will try to share existing MARC records.

These systems are also quite different in hardware configuration and programming, which we will have to overcome by making the software compatible. This leads to another problem. If we have different hardware, it is difficult for software to run on different computers, since the operating systems are different. So our only recourse is to select and use better systems together or to add more functions to the present systems.

We face many problems. To avoid both old and new ones, we must raise some considerations for future system development. We hope these will be useful to people planning and building integrated library systems both inside and outside the Academy as well as to those in other developing countries.

4. SOME CONSIDERATIONS FOR FUTURE DEVELOPMENT

First, we need to develop standard integrated library system software for the regional libraries. Such systems should be shared among the institute libraries or information divisions for their library work routines. We think multi-user 386's and 486's with NOVELL LAN is the best system configuration for the institute libraries with no regional libraries nearby, and with this LAN they can connect to other computers in the institute. The five regional libraries should try to perfect their systems and connect with the institute libraries to create and share a common bibliographic database. The Shanghai area of the Academy is best suited to connecting SDICIS with SIOCL in order to create online union catalogs which other institute libraries can share.

Second, we should download records from other cataloging sources like Chinese MARC and Bibliofile data to minimize the amount of original cataloging. The creation of an onlineunion catalog database is a big project. Unless we convert our card records to MARC records, we will have difficulties. The present systems have problems in dealing with records in communication formats. We need to design a reasonable data structure in each of our systems and take advantage of others' cataloging resources.

Third, we need to devise operating manuals and other necessary materials in support of our systems so they can be used more efficiently.

Fourth, we need to pay special attention to user training and instruction. Librarian training should be conducted at the sites of the systems. Thus far, patron instruction has been neglected, and the optimum use of OPACS de-emphasized. As systems develop, this type of training becomes increasingly important, and so we believe that user instruction is essential to the development of integrated library systems.

Fifth, in addition to these integrated library systems, regional networks of shared databases should be established in accordance with area needs. The TRANSTECH system, which will be replacing DILAS, should join the Educational Network in the Zhongguancun Area of Beijing, which will connect to CHINAPAC, linking bibliographic and information databases in the main cities of China. Most of the institute libraries in that area can join the system for shared cataloging, inter-library loan and searching. Other regional libraries should emulate this network in sharing their resources. To achieve the above goals, we should fully utilize new information technologies, available funds, managerial talents, and personnel resources.
 
 

REFERENCES

Chinese Academy of Sciences, Library. (1991). Testing Report of DILAS.

Wang, Yuan. (1992). "SIOCL System and Some Considerations," New Technology of Library and Information Services, vol. 2.

Shanghai Library of the Chinese Academy of Sciences. (1992). "SDICIS Recognition, Special Section," Library News, No. 5.