The Hypercatalog Graz - Budapest (HyperKGB)

Hannes Baptist, Haimo Primas, Heinrich Schadler, Gabrijela Trivic, Boris Wedl, Helmut Weitzer, Gernot Wöber, Christian Schlögl
University of Graz, Graz, Austria

Zoltan Czövek, Krisztina Katona, Gabriella Rozsa, Peter Sütheö
Eötvös-Lorand-University, Budapest, Hungary


After a short introduction to the shortcomings of present online catalogs some concepts and implementations for hypercatalogs are presented. In the main part, a prototype for a hypertext-based library catalog (HyperKGB) is described. Its design principles are laid down and the functionality is shown.


Most of today's online catalogs still have some shortcomings. One of the most severe failings is that they first and foremost support the search requirements of search specialists (Hildreth, 1989; Marchionini & Shneiderman, 1988). For instance, Boolean retrieval enhanced with a friendly user interface might be well suited for this kind of user.

The search situation is fundamentally different for people who use online catalogs for their own purposes. Most of them do not search for a specific 'known item' because they cannot, or do not wish to, precisely describe their information needs before the retrieval process. For such users a search strategy based on the 'browsing paradigm' would be better suited providing more navigation possibilities between the linked bibliographic records (e. g. 'show me more books in this series'). This could be better attained by a catalog extended with hypertext features.


Vannevar Bush (1945) is often cited as the 'founder' of the concept of hypertext. The term itself was coined by Ted Nelson (1967). Although hypertext has meanwhile won general acceptance (e.g. the boom of the WWW), it is surprising that this concept has not been considered more widely for library catalogs.

One of the most well known hypercatalog projects was performed at the University of Linköping (Björklund, Olander & Smith, 1889; Hjerppe, 1989). According to Hjerppe, HYPERCATalog was designed to be an extension and enhancement of the traditional catalog. Thus, its design incorporates the following features:

In their paper, Björklund, Olander and Smith (1989) realized these features and established a model for HYPERCAT. But as far as we know, this ambitious project was never finished.

In the Book House project (Pejtersen, 1989), a library system has been implemented for average citizens to search fiction in public libraries. The user interface of Book House is based on a building metaphor. Upon entering the library, the user can choose between a room with books for children, a room with books for adults or a room with books for everybody. Afterwards the user reaches the 'strategy selection room'. Here they can select among four different search strategies:

While in the Book House, the user can 'only' search for books, the Adaptive Network Library Interface (ANLI) project (Zhao & Kantor, 1993) was planned to enable members of a common interest group to create links between documents through interactive use. As a result, the ANLI will gradually build up a network through interaction with users. Thus it should be possible to share information about related books in a way that would naturally occur in face-to-face discussion.

Bertha (1992) also concentrates on the relation between documents. Contrary to Zhao and Kantor, she developed a detailed classification of links which require a certain degree of bibliographic description. Bertha distinguishes between

The boom of the WWW has also had some effects on the development of hypercatalogs. Some hypertext interfaces can already be found in online catalogs (see e. g. University of Arizona:; Grinnell College: or the University of Innsbruck: They share a character interface (no graphics) and simple navigation possibilities. This also applies in the OPAC module from OLIB (Finbarr, 1995) (see for instance, a commercial library information system.


Based on the study of existing concepts and realizations, the following design principles have been laid down:

  1. HyperKGB should provide hypertext features together with a graphical user interface (unlike existing WWW hypercatalogs). Graphical presentation of parts of the search process and the search results should give the user an enhanced 'search feeling' and make the information seeking much more productive.
  2. The hypercatalog should support different search strategies. At least, query search requirements and browse search requirements have to be met.
  3. Like at ANLI it should be possible for users to establish links between related books. In addition fixed links should be provided by the system. For pragmatic reasons, not all relations suggested by Bertha (see above) need to be implemented.
  4. For maintenance reasons, the links between the documents should not be static but dynamic. This design decision resulted in a two level architecture of HyperKGB. Whilst the user interacts with the hypertext interface, the data are stored with a relational database management system (RDMBS). Each time a user activates a link at the hypertext interface, it is transformed into an SQL statement which locates the corresponding documents in the database. Afterwards they are presented at the user interface. As implementation tools, we used Toolbook from Asymetrix for the hypertext interface, and MS Access for generating the database. Nevertheless, any other RDBMS supporting the ODBC standard can be employed.


The hypercatalog was implemented for the library of the Department of Information Science. The reason for this was the size of the library (around 1300 volumes) which was well suited for the prototype. In addition all data were available electronically. Below, the main functionality of HyperKGB will be described.

In the hypercatalog a user has three general possibilities for searching:

At any time it is possible to change between these search modes.

While matching, the user can choose between search by author and/or title and in more detail, where they can search by any combination of fields in the database using Boolean operators AND, OR as well as ANDNOT. After a search was executed, the retrieved books are presented graphically on the screen (Figure 1). On the spine of each book, its author and title can be seen.

If a user wants more information about a specific book, they have to double-click it. The corresponding book is 'opened' and all bibliographic data will be displayed (Figure 2). In the opened book it is not only possible to get information about it, but also to make personal annotations. These remarks can be read by all other users.

After a book was selected (single mouse click) or opened, it is possible to browse through the catalog. In browsing mode, the catalog allows the user to connect to

These links can be activated either by clicking on the corresponding field contents (e. g. [editor:] Henderson Diane) which are marked in blue color in the current book, or by selecting them from the menu 'More works' ('Weitere Werke') (Figure 1).

Browsing is also possible in the thesaurus. After selecting a descriptor from a list, the user is presented a local map of the thesaurus (Figure 3). In this local browser, the selected descriptor is located at the center. Surrounding it, all terms with direct relationships are displayed. Above the broader, more general terms are displayed, while below the narrower, more specific terms can be found. Other semantically related terms are displayed to the left and the right. The relationship is also expressed by the colors of the descriptors.

Now the user can navigate through the thesaurus by clicking on one of the neighboring descriptors. As a result, a new local map appears. On request, all books which have been indexed by the currently selected descriptor will be shown.

If a user finds relevant books during their search, they can put them into the rucksack (Figure 1, bottom right-hand corner). At the end of a search, they can print the contents of the rucksack and/or store it as a so-called user-defined collection. These collections make it possible for users to link related books. They can be accessed by other users as a further search strategy and may also be used for browsing (see above).

When browsing through a hypertext (as well as a hypercatalog), it is easy to get disoriented (Conklin, 1987). This has been countered by various means. As already mentioned, a user can change from browsing to matching at any time. In addition a bookmark and a history function were implemented. Whilst via the bookmark function, one can return directly to a browsing stage determined before, this is only possible step by step from the history function.


In the meantime, a laboratory experiment has been carried out. Around 50 students in the Department of Information Science have tested the catalog. The results can be summarized as follows: HyperKGB proved to be easy to learn and easy to use. Nearly all students preferred the catalog to the existing OPACs at the University of Graz though there was criticism on some parts of HyperKGB. Most criticism focused on the detailed version of querying which was found to be very complicated. The graphical representation of the search results as a heap of books received varied responses. Some students liked it very much whilst for others the graphical representation of the books was too badly arranged.

After some improvements, HyperKGB will be tested in a real library environment. If the results are encouraging, it is planned to develop HyperKGB to a universal access interface tool for online catalogs.


Bertha, E. (1992). "Beziehungen zwischen bibliographischen Dokumenten." In H. Zimmermann, H.-D. Luckhardt, & A. Schulz (Eds.), Proceedings of the 3rd International Symposium of Information Science (ISI'92) (pp. 316-323). Konstanz: Universitätsverlag Konstanz.

Björklund, L., Olander, B., & Smith, L. C. (1989). "The Personal Hypercatalog." In Proceedings of the 52nd Annual Meeting of the Association for Information Science (ASIS'89) (pp. 115-120). Meadford, N.J.: Learned Information.

Bush, V. (1945). As We May Think. Atlantic Monthly, 176, 101-108.

Conklin, J. (1987). Hypertext: A Survey and Introduction. IEEE Computer, 20, 17-41.

Finbarr, J. (1995). Extending the Third Generation OPAC. Journal of the Library Association Information Technology Group, 32, 27-32, see also: FDI/x3gopac.htm.

Hildreth, Ch. R. (1989). "OPAC Research: Laying the Groundwork for Future OPAC Design." In Ch. R. Hildreth (Ed.), The Online Catalogue: Development and Directions (pp. 1-24). London: The Library Association.

Hjerppe, R. (1989). "HYPERCAT at LIBLAB in Sweden: A Progress Report." In Ch. R. Hildreth (Ed.), The Online Catalogue: Development and Directions (pp. 177-209). London: The Library Association.

Marchionini, G., & Shneiderman, B. (1988). Finding Facts vs. Browsing Knowledge in Hypertext Systems. IEEE Computer, 70-88.

Nelson, T. (1967). "Getting It Out of Our System." In G. Scheckter (Ed.), Information Retrieval: A Critical Review (pp. 191-210). Washington D.C.: Thompson Books.

Pejtersen, A. M. (1989). "A Library System for Information Retrieval Based on a Cognitive Task Analysis and Supported by an Icon-Based Interface." In Proceedings of the Twelfth Annual International ACM SIGIR Conference (SIGIR'89) (pp. 40-47). New York: ACM Press.

Zhao, S., & Kantor, P. B. (1993). "Development of an Adaptive Network Library Interface: Progress Report and System Design Issues." In Proceedings of the 56th Annual Meeting of the Association for Information Science (ASIS'93) (pp. 211-216). Meadford, N.J.: Learned Information.