Enhancing Usability of Network-based Library Information System - Experimental Studies of a User Interface for OPAC and of a Collaboration Tool for Library Services

Shigeo Sugimoto, Seiki Gotou, Yanchun Zhao, Tetsuo Sakaguchi, Koichi Tabata
University of Library and Information Science (ULIS)
1-2, Kasuga, Tsukuba, Ibaraki 305, Japan
phone: +81-298-52-0511
fax: +81-298-52-4326
email: sugimoto, gotou, zhao, saka, tabata@ulis.ac.jp

KEYWORDS: User-friendliness of Library Information System, Graphical User Interface, Collaboration Support System, Reference Work, Digital Library


User-friendliness of library information systems is an important consideration in the development of the digital library. This paper describes two experimental studies of library information system usability: Graphical User Interface (GUI) oriented OPAC (Online Public Access Catalog) with a bookshelf-like browser named SOPAC, and a Collaboration Support System (CSS) and its application to a reference task. SOPAC uses a bookshelf metaphor where books are displayed and the image of each book is based on its size. SOPAC accesses the OPAC database of the university library of ULIS without any modification of that database. The CSS includes TV-phone, Shared Virtual Display, Image Tool, and White Board. An experiment applying it to a task teaching how to use a GUI-based OPAC via network is described. Both systems have been developed on a conventional workstation LAN (Ethernet, 10 Mbps).


User-friendliness of library information systems is important, because they are used not only by professional librarians but also by novice users. Advanced hardware and software technologies have brought us new concepts to build user-friendly library information systems such as GUIs, multimedia user interfaces, collaboration support tools, etc. For example, electronic document delivery systems such as the project mercury at CMU[mercury], RightPages[rightpages], TULIP[tulip], and NACSIS-ELS[adachi1][adachi2] provide users with GUIs to retrieve and read documents. Large digital library projects such as those granted by NSF will build a vast and useful information retrieval environment[nsf].

The digital library will provide us with an intellectual environment where users can interactively retrieve and read documents, books, audio-visual materials, etc. However, from a practical point of view, we cannot build such environments in one day, because not all libraries can afford high-performance computing and network facilities. In addition, since library information systems are data-centered, it is not easy to modify data-dependent functions or to add new information to an existing database in order to create new functions and user interfaces. These points imply that we need to build toward the digital library gradually, starting from existing library information systems.

In this paper, we describe two experimental systems developed on a conventional workstation LAN. The first system is an OPAC called SOPAC which has a bookshelf-like browser. The second is a CSS designed for collaborative library services between a library user and a librarian.

SOPAC has been developed on the database of an OPAC system called XOPAC that is in use at the university library of ULIS[take]. A user can issue retrieval commands and receive textual results through various windows of SOPAC. He/she can show a bookshelf window where books are sorted in the order of books on the physical bookshelves in the library. In the bookshelf window, each book is represented as a book-shaped image which is a push-button to request detailed information about the book. The book-shaped image is created only from a record of the existing catalog database, i.e., title, author, and size. Users can see the height and thickness of a book in addition to its title and author(s).

Collaboration is an important issue for the digital library, because this library will be the principal environment for intellectual activities: a user seeking information will consult with a librarian, and groups of users will meet in a virtual conference room in the digital library using library materials. The authors have developed a collaboration support system (CSS) in order to evaluate the applicability of CSCW technologies to reference services. The CSS includes a TV-phone, a Shared Virtual Display, an Image Tool, and a White Board, all of which are considered essential for communication between the librarian and the user. A CSS has been developed on conventional UNIX workstations with audio-visual interfaces connected by Ethernet (10 Mbps) LAN. We performed an experiment on the applicability of the CSS to the task of teaching how to use XOPAC. Since XOPAC is GUI-based, the instructor had to show the student dynamic features such as transition of windows, movement of the mouse cursor and key input in a window. The results with a sample of 30 students was positive.


Information Access Support Tools in Libraries

Consider a user in a typical library. He/she would search for a book using an OPAC terminal, find a book and go to a bookshelf designated by the OPAC. Then, he/she would locate the book on the bookshelf and pick it up to browse it. In some case, he/she would find another book which is more suitable to his/her requirements than the retrieved one. In the case he/she cannot find any book suitable to his/her requirements at the bookshelf, he/she would go back to the OPAC terminal or consult with the reference librarian concerning such problems.

The preceding paragraph describes the typical behavior of a user in a conventional library. The OPAC and the librarian are essential aids for the user to access books in the library. From these considerations, the authors found the following aspects to be important to improve the information environment based on exisiting library information systems;

  1. Bridging the gap between OPAC and bookshelves: A user often goes back and forth between an OPAC terminal and bookshelves, so that a tool like a bookshelf browser coupled with an OPAC will be helpful.
  2. Improving accessibilty to a librarian: It is crucial especially for remote users to have improved communication tools for collaborating with the librarian. This is because, even if OPACs and electronically accessible materials are well-provided, the users need specialized assistance.
From these considerations, the following guidelines for OPAC usability improvement were derived. It seems obvious that a collaboration environment for a library will be implemented using a well-designed CSCW tool on a high-performance workstation on a high-performance network. However, from a practical viewpoint, we should examine the feasibility of collaborative reference work in a conventional environment consisting of conventional workstations, personal computers and network.

Enhancing Usability of OPAC

OPAC has been the most common tool for library users and librarians, and it will be also commonly used in digital libraries. It is obvious that well-designed GUI improves user-friendliness especially for novice users. OPAC with a bookshelf-like browser of bibliographical data would be useful from a practical point of view, even if the bookshelf on a display does not look like a physical bookshelf. The advantages of the bookshelf-like interface are: 1) users can find not only a book retrieved via retrieval commands but also from a glance at other books placed on the same bookshelf, 2) the structure of bookshelves is semantically organizable, and 3) the display order of books is modifiable as needed.

HOPAC[hopac] has a bookshelf-like interface organization based on the Dewey Classification System. Its hierarchical structure helps a user move his/her viewpoint in the space of books and documents. The Book Shelf window of HOPAC does not look like a physical bookshelf. In addition to the text-based presentation, the authors consider that images of books and bookshelves would help users find books. The number of book images displayable on a display at a same time is obviously smaller than the number of book title texts. This fact looks disadvantageous, but too much information on a display is not friendly to users. The image-oriented bookshelf has the advantage that users can immediately see the size of a book without reading detailed bibliographic information. In addition, we can embed icons into a book image which represent such information as the book being checked-out, having duplications, one of series books, and so on.

Collaboration by Library Users and Librarian

The librarian is an important information resource in a library, and we cannot fully take advantage of a library without his/her help. A digital library will not necessarily be a fully automated librarian-less library, but will have to provide intelligent collaborative services of librarians.

Yamamoto mentioned the roles of librarians in the era of digital libraries as follows[yamamoto],

He also mentioned the types of reference services via network as follows,

Figure 1: Snapshot of SOPAC.

Figure 2: Bookshelf Window.

In addition to the collaborative services of librarians, the digital library will have to provide environments for group work[lotus][marshall]. For example, researchers sometimes use a conference room in a library for research meetings using journals and other materials in the library. They store their new knowledge and information which are got from their discussion into the library. This means that the digital library should provide a virtual conference room where users can hold discussions through the same displays where library materials such as electronic journals and digitized images are shown.

Thus, collaboration support is one of the important features that should be explored in order to build the digital library. Reference services could be carried out using conventional media such as telephone, facsimile and email. However, these media do not offer the face-to-face communication that is quite important both for librarians and users; librarians must be able to correctly understand questions and needs of users, and users would be able to appropriately express their questions in their conversation with librarians.



SOPAC has been implemented based on the client-server model on a unix workstation which has the Motif GUI environment. The database of SOPAC is identical to the database used by XOPAC, which has 110,000 titles.

Users communicate with SOPAC via multiple windows. A snapshot of the SOPAC GUI is shown in Fig 1. This GUI includes the following windows:

  1. Retrieval Command Window -- accepts retrieval commands as well as set operations AND, OR, and DIFF for intersection, union, and difference operations on retrieved data.
  2. Result Window -- displays retrieved data. Users can get detailed bibliographic data of a member of the list on the Detailed Information Sub-window.
  3. Memopad Window -- for user annotation.
  4. Bookshelf Window -- browse a portion of bookshelf by designating a book on a result window. A bookshelf image of the designated book and its neighboring books is displayed. The order of the book images is primarily determined by the identification numbers of the books.
  5. Detailed Information Sub-window -- all of the bibliographic information of a book which is selected on a result window or a bookshelf window.

Bookshelf Window

Users can browse a bookshelf by clicking the Display Bookshelf button to open the window shown in Fig.2. SOPAC displays a bookshelf with the book highlighted on the Retrieval Window and placed in the center. The order of books is the same as that on a physical bookshelf. The order is determined by the unique identification number given to each book, and users can scroll the bookshelf horizontally.

Each book image is a push-button to request detailed bibliographic information, which is displayed on a Detailed Information Sub-window attached below the bookshelf window. The shape of a book image is determined by the bibliographic description of the book, i.e., height and number of pages. The book title is presented from top to bottom and can be vertically scrolled.


It is not easy to re-create an OPAC database or to modify the scheme of its records. SOPAC has been developed on the XOPAC database. It would be better for bookshelf-like browsers to be able to show physical images of books, i.e., covers and title pages. However, it was not practical to scan in the physical image of every holding and to add it to an existing OPAC database, so the SOPAC book image is created only from existing data.

Books are displayed vertically in order to enhance realism. This means that title texts on the images are drawn vertically. This is convenient for titles in Japanese.

Several issues remain unsolved in the current SOPAC system:

  1. Bookshelf reorganization: functions to re-organize bookshelves in accordance with semantical requests from users,
  2. Browsing Support Tools based on Semantical Structure of Bookshelves: tools to show classification hierarchy and thesauri, and
  3. Active Information Tools: icons to show active information such as "checked out", "in bindery", etc.


In this section, we firstly discuss a model of reference work, and then describe the system structure of the CSS. Some experimental results using the system are also described.

Modeling Reference Work as Librarian's Desk

The principal components of reference work in conventional libraries are as follows;
  1. a librarian,
  2. a library user,
  3. library materials, which include primary information sources in a library such as books, journals, etc., and secondary information sources such as OPACs, thesauri, dictionaries, and encyclopedias,
  4. user's materials such as books and memos, and
  5. a librarian's desk.
Based on the objects listed above, we can model an environment for CSS. The library user and librarian are modeled as the participants in a conversation carried out via the system. Library materials and user materials, which may or may not be in digital form, are modeled as objects to be displayed on the terminals in front of the participants. The librarian's desk is modeled as a display screen in front of each participant where the participants can see their partners and view the same materials.

It is natural to design a library collaboration support system based on the librarian's desk metaphor. We can find memo pads, a computer terminal, a telephone, a copy machine, and so on in addition to the library materials on/around a librarian's desk. These things should be modeled as system components visible to the participants. For example, the participants would enhance their discussion with visual materials and memos displayable on their terminals.

Figure 3: System Overview.

Figure 4: Snapshot of the Collaboration Support Tool(TV-phone and SVD)

Figure 5: Lan Environment.

  1. The librarian (L) explains a retrieval window.
  2. L shows a simple retrieval operation to the user (U).
  3. L shows and explains the result window where the retrieved result is displayed.
  4. L copies the contents of the result window into a Memopad window, and explains its usage.
  5. L explains and shows an example of a composite retrieval which implies AND, OR, and NOT operation.
  6. L shows the result.
  7. L explains the operation to terminate the OPAC.
  8. U does a simple retrieval and displays its result with assistance by L.
  9. U uses a scratch window with assistance by L.
  10. U does a composite retrieval and displays its result with assistance by L.
  11. L asks U whether U has any questions about the usage.
  12. L tells U to do practice until U gets satisfied.
  13. U says Good-bye. (Note: The OPAC has a retrieval window to input retrieval commands, result wind ows to display retrieval results, and scratch windows to make personal annotati ons.)

Figure 6: Retrieval Instruction Scenario

SONY NEWS-3000 workstations
CPU: MIPS R3000, 20/25 MHz, 32/64 MB RAM, 17/25 MIPS
Display: 1024(W) x 768(H), 15 bits true color (32767 colors)
Video Input: NTSC
LAN interface: Ethernet
Window System: X-window Version 11, Release 6 OSF/Motif
Backbone LAN: FDDI (100 Mbps)
Front-end LAN: Ethernet (10 Mbps).
Audio Sound
Sampling Rate: 8 KHz
Quantumization: 8 bits
Channel: Monaural
Unit Data Size: 8 KByte (0.4 sec/unit)
Video Image
Capturing Rate: 2 frames/second
Size: 160(W) x 120(H)
Compression/Decompression: None
Image Size: 38 KByte/frame

Table 1. Hardware Specifications used in Experiment

System Structure

The CSS has been developed on two workstations of the ULIS campus LAN. The graphical environment of the workstation is OSF/Motif on X-window (X11R6). The overview of the system on each workstation is shown in Fig.3. Each component is realized as a window. A snapshot of the terminal display is shown in Fig.4. The following paragraphs are brief descriptions of the components. Overview of the LAN environment and specifications of hardware used in this study are shown in Fig. 5 and Table 1, respectively.

Applying Collaboration Support Tool to Library Service

We applied the CSS to the task written below.

TASK: A user who wants to use a GUI-based OPAC requests and receives instruction from a librarian. Librarian teaches him/her how to use the OPAC.

The user is assumed to have basic knowledge about OPAC and to have some computer experience, not necessarily with GUI. The librarian (instructor) shows how to use the mouse and keyboard to input parameters and invoke operations. The user (student) learns and practices the example operations given by the instructor. The windows on SVD accept equally keyboard and mouse inputs from both participants. The mouse cursors of each participant are visible to both participants.

The scenario given to the librarian for the experiment is shown in Fig. 6. Since this scenario involves interactions which are common among interactive services by librarians, we can evaluate the applicability of the system not only to the service written above but also to interactive services in general.

Experiment and Evaluation

30 users performed the task in the scenario given above using two remotely placed workstations. They were asked to fill out a questionaire shown below before and after the experiment. For comparison, four users received instruction in the same task from an instructor sitting next to the user in front of a workstation.

Questionaire (Definitely Yes: 5 - Definitely No: 1)

All of the participants were students of ULIS. About 20 % were novice users of workstations, not used to a mouse of GUI-based applications. All students had learned information retrieval with the ULIS-OPAC on a main-frame with a character user interface. The OPAC used in this experiment, called XOPAC, is a new one, so that not all of the students had learned it.

question   5   4   3   2   1   total 
Q.1        6  10   8   2   4      30
Q.2        7  10   8   4   1      30
Q.3        1  5    8  14   2      30
Q.4        5  8    7  10   0      30
Q.5       16  12   1   1   0      30
Q.6       15  11   2   2   0      30
Q.7       12  10   6   2   0      30
Q.8        7  7   13   3   0      30
Q.9        8  16   5   1   0      30
Q.10      16  10   4   0   0      30
Table 2: Responses to the Questionnaire

minutes | 10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25 
persons |  1   0   2   0   5   2   4   4   4   2   3   2   0   0   0   1 
Table 3: Minutes for Instruction by Librarian Using the Collaboration Support System

Responses to the questionaire on usability of the system is summarized in Table 2. Response to the system was mostly positive, and criticism was due mainly to the low quality of the TV-phone, such as poor sound, image size and frame rate, and half-duplex communication. Since such criticism was mainly of the performance rather than structure, these problems can be solved in future high performance network and computers.

minutes | 10  11
persons |  2   2
Table 4: Minutes for In-person Instruction

Instruction time required using the network CSS is summarized in Table 3, and using in-person instruction in Table 4. The average instruction times of the CSS and the in-person groups are 16.8 and 10.5 minutes, respectively. The fastest and slowest times of the CSS group are 10 and 25 minutes, respectively. The instruction times of 26 CSS users (87%) are between 14 to 21 minutes, and the median is 17 minutes. The instruction time of the CSS group has a wide range compared with the four students of the in-person group.

We attribute the difference betweenthe two groups to the fact that the CSS students had to learn two systems, the OPAC and the CSS, since the experiment was the first experience to use the CSS for all of them, and some were novice users of workstations.

We feel that the difference in instruction times between the two groups is reasonable and acceptable, because all users learned to use the OPAC without going to the librarian's desk at a cost of only 7 additional minutes.


`Digital library' is a fascinating term, but many problems remain to be solved and a high-performance computing and communication environment is needed to achieve the ideal digital library. Nevertheless, in this paper, we described two systems developed in a conventional environment which indicate that the usability of library information systems will be greatly enhanced by the future digital library.


[adachi1]Adachi, J., Premises for Establishment of an Electronic Library System, Digital Libraries (ISSN 1340-7287), No.1, 1994.8 (in Japanese) (http://www.dl.ulis.ac.jp/DLW_E/Journals/no1.html, Abstract in English)

[hopac] Allen, R.B., Navigating and Searching in Hierarchical Digital Library Catalog, Proc of DL'94, pp. 95 - 100, College Station, TX, 1994.6

[lotus] Ehrlich, K. and Cash, D., Turning Information into Knowledge: Information Finding as a Collaborative Activity, Proc. of DL'94, pp. 119 - 125, College Station, TX, 1994.6

[take] Fujita, T. et al., Transporting an Online Public Access Catalog from Mainframe to Distributed Environment, Proc. of 47th FID, pp.444-448, Ohmiya, 1994.10

[rightpages] Hoffman, M.M., et al., The RightPages Services: An Image-Based Electronic Library, Journal of the American Society for Information Science, Vol.44, No.8, pp.446-452, 1993.9

[marshall] Marshall, C.C. et al., Putting Digital Libraries to Work: Issues from Experience with Community Memory, Proc. of DL'94, pp. 126 - 133, College Station, TX, 1994.6

[mercury] Trol, D.A., Project Mercury and Development of the Library Information System, Mercury Technical Report Series, No.7, 1993

[yamamoto] Yamamoto, T., The Digital Librarian: Work and Tools, Digital Libraries (ISSN 1340-7287), No.1, Tsukuba, 1994.8 (in Japanese) (http://www.dl.ulis.ac.jp/DLW_E/Journals/no1.html, Abstract in English)

[tulip] ELSEVIER SCIENCE B.V., TULIP - The University Licensing Program http://www.elsevier.nl/info/projects/tulip.htm, March 1995.

[nsf] National Science Foundation, NSF ANNOUNCES AWARDS FOR DIGITAL LIBRARIES RESEARCH, http://www.nsf.gov/nsf/press/pr9452.htm, NSF-PR 94-52, Sept. 1994

[adachi2] The National Center for Science Information Systems (NACSIS), NACSIS Digital Library Project, http://www.nacsis.ac.jp/dl-j.html (in Japanese)