FROM THE TRADITIONAL TO A DIGITAL ACADEMIC LIBRARY

Damir Kalpić, Jasenka Anzil, Hrvoje Zoković

Faculty of Electrical Engineering and Computing, Unska 3, 10000 Zagreb, Croatia

, ,

Abstract: At the Faculty of Electrical Engineering and Computing in Zagreb the development of computerised services concerning the libraries has lasted for six years. The institution consists of one central and eleven departmental libraries. It employs the staff of about 300 and enrols 3000 students. The libraries should primarily serve the insiders but also have to be open to the outside public. The software was developed closely to the student administration system. The classical alphanumeric interface has been gradually expanded with OPAC (On Line Public Access) features. The first purely digital materials were the abstracts. In the near future doctoral and master's theses will have to be submitted also in a digital form. Further enhancements derive from receiving of digital journals. A special topic are the lecture notes and examination materials and their (un)availability which calls for some legal issues to be solved.

Keywords: University, Distributed library, Internet, OPAC, Student administration, Information oversupply, Software

Introduction

Eight years ago, at the Faculty of Electrical Engineering and Computing a significant investment in the computer equipment was realised, enabling the computerisation of the basic Faculty functions. We had started our own development of the student administration (Kalpić and Mornar, 1994) as an intrinsic Faculty function. The libraries were the next important issue. We have a central library and eleven departmental dislocated libraries. At that time, attempts were made in Croatia at developing a software which could suit the major national library institution and which would be distributed and applied in other institutions comprising libraries. The idea had appeared attractive, especially as the institutions should receive the software free of charge. It had turned out that only single-user application was free and that the software was inappropriate for searching and borrowing of books. The developers proclaimed that their software was supporting distributed libraries, which was far from reality. Even to obtain a simple listing of the library contents was a tedious task.

Developments at the Faculty of Electrical Engineering and Computing

When we made the decision for our own development, the programming system for the student administration (Kalpić, Anzil, Dumanić, Zoković, 1993) was partly completed and running on a UNIX/INFORMIX platform. The natural choice was the same platform. The application functionality has been designed mostly from the user's point of view. All the functions handle the library material deposed in the central library as well as in the dislocated departmental libraries. The access is possible via LAN or WAN, according to the user authorisations.

The principal evidenced data about books are: author(s), title, publisher, scientific/professional field, keywords, publication data (edition number, year of publishing, ISBN and some other details), language(s), script(s) and the summary. All the repeating data are coded and entered only once. Data about the physical position of each copy are entered and a unique bar coded identification is printed.

The principal data about journals are: journal title, publisher, ISSN, publishing frequency, language(s), scientific/professional field. Data about a journal are entered only once, at the first appearance in the system. For each issue, the issue number and the publishing date, or month and year are entered. For each article the author(s), the title, language, script, field, keywords and an abstract can be given.

There is a standard search after the title, authors and publisher and by most other attributes upon the Query-By-Example principle, with the wild cards feature included. The keyword search fetches all the books and/or journals bearing the required keywords or their synonyms. The user can browse the keywords first to find the available keyword forms. For a group of keywords, when they first appear, the librarians have to choose a representative synonym. In this way, all the keywords in the group are defined as synonyms. For example, the keywords: OPTIMISATION, OPTIMIZATION, OPTIMIRANJE, OPTIMIERUNG, OTTIMISAZIONE etc. should have the same representative, e.g. OPTIMISATION, which makes them all synonyms.

All the books and/or journals for given UDK number(s) or hierarchically higher parts of them, are fetched. To a book or to an article in journal, multiple UDK classifications can be attributed. Look-up tables of UDK codes are available.

Currently, the abstract text search is not available. A solution, appropriate for English texts, could be attempted by application of Informix Text DataBlade software (Informix, 1998).The Croatian language is very inflected and maybe an application of language specific solutions should be required for search (Kalpić, 1994) and storage (Ristov, Boras and Lauc, 1997) of the language forms.

In borrowing of library material, the data are recorded about the book copy (its accession number) or the journal issue number, the borrower, the dates of borrowing and due return and possible extending. Bar codes are used to identify the book and the student borrower. They are printed on stickers and attached on the inner side of books or student indexes (matriculation pass). Periodical standard reports can be printed concerning the borrowing activities. They can be grouped/sorted by type or field of the library material, by client characteristics etc. Warning notices to borrowers are sent via e-mail, if available.Evidence of borrowing is connected to the student administration system and the Faculty personnel database. Therefore, for most of the library clients their personal data are already available. A user can reserve a book and he/she gets a personal notice via e-mail, as the book becomes available. Catalogues are printed in a standard card format, or they can be specially formatted. The library material can be presented in an abridged version of UNIMARC. Information exchange with other libraries is supported for any requested data format.

All our students have free Internet access via the Faculty. Our student administration software has supported students' applications for examination through graphical interface on touch screen monitors (Kalpić, Zoković, Anzil and Pukljak, 1994) or from home via Internet. The most required information from the library database has been hyperlinked to the Faculty WWW site. The search is performed as Query-By-Example, wild cards included. The software enables anyone to use the OPAC features to search through books and journals. Any user can get the information about the books in the Library evidence and the number of currently available copies is given.We have not displayed on the Web yet the data about the articles, authors and the article abstracts because these data are scarcely entered. Similarly, the keyword search and the UDK search are not widely displayed because the quality of the entered attributions is currently still questionable.

The software has been developed on a Sun SPARC Station, under SunOS. It runs also under SCO UNIX. Informix On-Line DS Ver. 7.22 is the chosen RDBMS and the programming language for the character-based applications is Informix-4GL Ver. 6.02. Client-server parts are written in Visual Basic Ver. 4.0 on Windows95/WindowsNT platform. The protocols in use are TCP/IP and HTT. The server is attached to the Faculty LAN and to the Internet through the Croatian Academic Network (CARNet). Any Internet user can browse the library material and remote users with authorisation can access the database via telnet and character-based applications. The current relational database consists of 89 strictly normalised tables.

The experienced problems and proposed solutions

The main problem is the acquisition of correct and up-to-date information. There is a notorious lack of time and/or competence and good will to complete this task and keep up with the changes.

A very unpleasant surprise was the poor quality of the existing data in departmental libraries, which had been formerly collected under some amateurish software. To transfer the data, practically every record from such a database has to be touched and edited. The data from some departmental libraries have not been transferred yet. A lot of library material has not been entered yet at all, because it is regarded as the job to be performed only by few library specialists.

A rather serious problem is the assignment of the proper UDK classification to the library material. On one hand, the librarians are acquainted with the classification method, but on the other hand they do not understand much about electrical engineering and computing, to which most of our material refers. Therefore, it happened that a master's thesis about CASE tools was classified within the electromagnetic theory. Even more serious is the problem with the keywords, if they have not already been declared by the author of material. The problem is for the librarians to choose the appropriate keywords and on the other hand to form appropriate groups of synonyms. While testing for the purpose of this article, the authors found out that OPTIMISATION and OPTIMIZATION have been put to different synonym groups! A possible solution for a proper description with the keywords could be to authorise the person who had decided on or suggested buying a book or a journal. S/he should make a selection from the set of existing keywords which would best describe the new object. The software was implemented but human engagement is also required. The other solution can be in acquiring the interesting data from a referral library or institution.

If the purchase of some library material had been subsidised by the Ministry, the obligation of the library is to lend it to any borrower. Clearly, this obligation does not exist for the material bought with one's own money. The departments are reluctant to put on display such material. Therefore, duplicate titles are often simultaneously bought by different departments.

The Faculty library has among its principal aims helping to improve the educational process. Data mining of the borrowing habits and the examination outcomes for the same students could discover unexpected patterns and functional dependencies. This could help the library to properly cluster its material available for free access, to acquire additional copies of useful books or journals and possibly do something else that remains to be discovered.

At this moment (September 1998), our library cannot pretend to be a digital one in the sense of definition at ( but some steps in that direction have been taken. The summaries of new doctor's and master's degrees are already collected in the digital form. An access to the library is possible via Faculty WWW home page.

Among the first data available in a digital, but still not properly organised form were the problems from former written examinations. A problem of exposing lecture notes on Internet remains for some lecturers questionable. First, lecture notes are not books or articles, so they normally contain errors or non-elegant expressions. For this reason, many of us would be ashamed to expose it to "the whole world". The next issue is the question of copyright. It is useful for a lecturer to take whole chunks of material from books and journals. It is legal and normal and nobody would expect him/her to ask for a written permission from the author of a book, often meant as teaching material! The situation may change if this material is displayed on the Web as part of the lecture notes. There is also the question of the lecturer's copyright. Practices differ, as probably in the rest of the world.

The scientific production of the Faculty members should be submitted in a digital form to the library. A search by keywords could detect potential experts. No need to say that the staff should only take care to update their personal information. Standard reports from this evidence of the scientific production could be made automatically to suit different purposes, as promotion, grant, specialisation etc. Whole institutions could be evaluated accordingly.

The classical library will most probably survive (Krieger, 1998) to store the non-volatile information and to enable our leisure-time reading. We are not so convinced about the secure future of classical libraries in fast developing and emerging fields and their role concerning scientific activities. From personal experience we know that the time from submitting a first paper version to any respectable journal to its printing can be measured in years. Many authors already develop strong reserves regarding their texts until the moment they appear published. Some may find the stuff published in journals too speculative and non-mature, and they prefer to read the books that usually contain a more systematic material. On the other hand, those who wish to be extremely up-to-date informed, browse the Web to find the newest material. It is well known that there is no guarantee regarding the quality of material displayed on the Web. One can expect further advancements of digital journals. To make a substantial change, the journal should exhibit the articles as soon as they are received. At that stage they should bear a flag as "unreviewed", then as "under correction" and eventually as "accepted". This would enable the most eager ones to get the first-hand information, risking its value. The reviewing process would be democratised because not only the assigned reviewers could make their remarks. That would not make the editor's job simpler, but the finally accepted articles would pass through a tougher procedure.

Once libraries helped to keep people informed. A future digital library may have a seemingly opposite task - to protect its users from an oversupply of information. The task of the librarians in such virtual digital libraries would be to train the intelligent agents (Gams, 1997) according to the specific personal profiles, requirements and preferences of their users and to supply the users with an appropriate digestible dose of information.

Acknowledgements: We owe our thanks to the former Faculty dean, Prof. Ivan Ilić, who during his mandate initiated the development, to the former vice dean and now dean, Prof. Slavko Krajcar who has offered a standing support to the project and to our colleagues, members of the Computer Science group, who directly or indirectly have been helping the project realisation. To Ms. Jadranka Lisek we are thankful for useful professional advice and for her support to efficiently apply and improve our software. We are thankful to Ms. Andrea Vrkljan for her valuable suggestions which had lead to improvements of the article terminology, grammar and style.

Bibliography

Informix (1998). DataBlade Development Kit, User's Guide, Version 3.4, Informix Software

Gams, M. (1997). Intelligent Agents - A New SW Breed, Proceedings of the 19th Intern. Conference on Information Technology Interfaces, Pula, June 17-20, pp 111-116

Kalpić, D. (1994). Automated Coding of Census Data, Journal of Official Statistics - Statistics Sweden, 10, 449-463

Kalpić, D., J. Anzil, V. Dumanić and H. Zoković (1993). Student Administration System, Proceedings of the 15th International Conference on Information Technology Interfaces, Pula, June 15-18, 1993, pp 123-128

Kalpić, D. and V. Mornar (1994). Student Administration System, European review conference proceedings "University-Enterprise Information Systems",Graz, September 15-16, pp 124-131

Kalpić, D., H. Zoković, J. Anzil and D. Pukljak (1994). Graphical Interface for the Student Administration System, Proceedings of the 16th International Conference on Information Technology Interfaces, Pula, June 14-17, pp 325-330

Krieger, D. (1998). Bibliophiles of the world, unite! A 21-st century spin on the fabled Library of Alexandria, Networker, The Technology Newsmagazine of the University of Southern California, 8, 26-36

Ristov, S., D. Boras and T. Lauc (1997). LZ compression of static linked list tries, Proceedings of the 19th International Conference on Information Technology Interfaces, Pula, June 17-20, pp 283-288