LIRN (Library Information Referral and enquiry Network)


Nils Meulemans, Telematics and Communication Service, University of Brussels, CP230 Bd du Triomphe, Brussels, B-1050, Belgium. Phone: +32 2 629 35 53 Fax: +32 2 629 38 16 Email: meulemans@helios.iihe.rtt.be
Keywords: LIRN, WorldWideWeb, X.500, Libraries, Thesaurus

Introduction

The LIRN Pilot Information Referral Service, funded by the Libraries Programme of the European Commission, is being set up to enable network users all over the world to locate and access information providers, such as libraries and information centres, which hold information on subjects of interest and value to them.

Although the pilot is restricted to information providers located within the countries involved in the project (Belgium, Portugal and United Kingdom), mechanisms are foreseen to extend it to other countries, the rest of Europe and even the whole world. The LIRN service is however not restricted and is available on a worldwide basis.

This paper describes the system and service requirements expressed by its users and how they were covered in the definition phase of the project, which started in January 1994 and ended in October 1994. The paper also describes the implementation phase, which is about to terminate when the AusWeb conference takes place. Furthermore, the paper summarizes the rationale behind the decision to base the service on the WWW concept while the information base is stored in an X.500 directory. The project is innovating in providing a multi-lingual search and retrieval facility based on this directory service. Within each country a national service node has been established, which can be accessed through the national IP networks. The individual nodes are connected using the national research networks and EuropaNET.

The next step in the project is the start of the last phase, the evaluation phase, which will start with the establishment of a public pilot service (spring 1995) and which will finish near the end of 1995.

Project Partners

The members of the LIRN Project come from the two disciplines of Librarianship and Information Technology. The libraries of the University of Brussels, the University of Aveiro and University College London are taking part in the project.

Two other organisations from the library world are also taking part. They are LASER, which began as a co-operative service organisation for libraries in London and South-East England and now offers an international service, and the Business Information Network (BIN), which is an association of libraries and information centres in the United Kingdom providing business information services.

The project partners from the information technology world are the Telematics and Communication Service [HREF 1] of the University of Brussels, who bring to the project both development and operational expertise in the area of X.500 directories and WWW, and Level-7, who are managing the project.

Objectives and Principles

The main objective of the LIRN project was to establish a search and retrieval service for the wide range of information services used by the library community. In the first place the LIRN service should assist the user in identifying information sources. Furthermore, when available, it should provide access to as many local on-line services as possible (e.g. WWW, Gopher, Telnet, Z39.50, Inter-Library Lending, etc.).

In order to have an effective search mechanism, a classification of the information providers was needed. A classification is however purely hierarchical and as such does not use the concepts of related and non-preferred terms, as they are quite often used in a library environment, the target field of the LIRN service. For that purpose we decided to use an on-line thesaurus instead of a simple classification.

The librarians in the project considered a number of approaches to solving the problem of providing a thesaurus for the pilot service, including the development of a new thesaurus from scratch. The initial choice was made for the SCIMP/SCANP [Geoffrey E. Cleave 1992] thesaurus. A particular point in favor of this thesaurus is its European orientation; it was developed as a collaborative effort by a number of European business schools. Furthermore, an electronic version was available, which was loaded into the system using specialised tools. However, when translating the thesaurus in French and Portuguese (the other two languages used within the project), it became clear that it would be difficult to maintain the same structure within the different languages. Although the LIRN system has been developed in such a way that it can work with different structures, the adaptation of the SCIMP/SCANP thesaurus towards only conceptual terms, which have a precise meaning in all languages (and countries), is currently being considered. The results of the thesaurus analysis will be reported to the original developers.

To make sure that the information about an information provider, which a user can retrieve from the system, is represented in a uniform way, relevant to the service and worthwhile making available to the public, the information providers were asked to register themselves in the LIRN system. For that purpose a questionnaire has been produced and sent around. Among the information available for registered information providers are contact information (postal address, telephone number, fax number, e-mail), access information (hours and type of service), on-line services (electronic databases, information services, remote access, etc.) and collections held by the information provider. A copy of the thesaurus was appended to the questionnaire to assist the information providers to identify their collections as precise as possible. For those thesaurus terms selected by the information providers, the LIRN system holds pointers to the entry of the information provider in the LIRN information base. A figure illustrating this relation can be found under ß5.3.

Filled out questionnaires are still handled manually one by one to allow the project team to tune the mechanism to its optimal format. Any feedback coming from the information providers will be taken into account when updating the questionnaire. In the meantime an automatic registration service is being established which will be offered to the public when the registration mechanism is stable. This service will be provided on each LIRN system through a WWW server.

The LIRN Service

The thesaurus-based search facility allows the user to identify information providers using the logical combination (AND and OR) of one, two or three terms (three was considered by the librarians to be sufficient), a search language (see thesaurus) and a set of countries. Before launching the search, the user can ask the system to verify the terms, as a search based on non-existing terms would most likely not result in any useful information. A powerful search/browse/referral facility assists the user in locating the terms that suit his request the best. As this search is quite complex, it is possible to make mistakes in specifying the search-filter. To help the user, a detailed analysis of the problem is returned when the search fails to find any matching information providers.

Besides the thesaurus-based search, the LIRN system also contains a thesaurus browser which can be used to locate an information provider based on one thesaurus term only. This feature is also interesting for helping first-time users to get familiar with the system.

A third service consists of an information provider search facility based on the collections held by the service. E.g., the user could specify a collection (e.g. Conference Proceedings) and a country (e.g. Portugal). As a result of this particular search the LIRN system will return a list of all registered information providers in Portugal holding Conference Proceedings. This is only a very simple example of this type of search. In practice the user can base his search on a set of collections and filter out collections for a specific continent, collections which are held for a certain period of time (1 year or retrospective) and collections which can be lent to external enquirers.

At the end, these three services return the same result: a pointer to an information provider which contains information matching the request of the user. If the information provider offers some on-line services, the LIRN system will provide direct links to them. Among the supported services are E-mail, Gopher, WWW, FTP, Telnet, WAIS, Z39.50, etc.. A detailed description of the LIRN service can be found in [Nils Meulemans 1995].

A fully operational LIRN service access point [HREF 2] is already available at the University of Brussels in Belgium. Two other access points are currently being established in Portugal (University of Aveiro) and the United Kingdom (LASER).

The LIRN Information Base

The LIRN information base contains two types of information: the registered information providers and the thesaurus. For both types of information, the LIRN system uses the X.500 [X500] directory. Similar projects have been carried out in the past, and before finalising the structure of the LIRN information base we have carefully studied two relevant projects: ABDUX [Paul Barker 1993][HREF 3] and CNUCE [Giovanni Armanino et al 1992]. In the following paragraphs we explain why we decided to go for this solution. A detailed description of the X.500 feasibility study carried out for the LIRN project can be found in [Nils Meulemans 1994].

Information Providers

The most efficient way to allow identification of information providers is to have them registered at pre-defined locations according to some classification. As the information held by these information providers can be very dynamic, and as such requires simple maintenance mechanism, it is best to store the information as close to the information provider as possible. To keep the information up-to-date, any modifications should be applied by the information provider itself.

Currently the most wide spread distributed information system in Europe, which meets the requirements specified above, is the X.500 directory. The open character of the X.500 information base and its multi- functional search facility provides a very good basis for LIRN. The fact that an X.500 infrastructure already existed, was an extra point in favor of it.

As long as they follow the pre-defined format of LIRN, it does not matter where an information provider is registered within the X.500 directory. This means that an information provider could install its own X.500 system (DSA - Directory System Agent) and link it to the existing X.500 infrastructure, or use an available DSA service. Within the pilot project, such a service is provided by the three LIRN access points: University of Brussels, University of Aveiro and LASER. These systems are hooked up to the European X.500 service operated by DANTE (NameFLOW project) [HREF 4]. This approach also guarantees a high scaleability of the information provider information base.

Thesaurus

Most of the currently used search and navigation systems within the WWW space are based on centralized indexes (e.g. Aliweb [HREF 5], Lycos [HREF 6]). This creates an enormous bottleneck at these locations. To avoid this problem, within the LIRN project we decided to distribute the thesaurus over several systems. Having the X.500 infrastructure at hand, this seemed to be the most appropriate place to store the thesaurus. The original thesaurus was the English SCIMP/SCANP thesaurus. This thesaurus was translated by the project team into French and Portuguese. By storing these translated terms as X.500 alias terms, pointing to the original English term, searching in any of these languages always has the same result. The following figure illustrates the alias mechanism used for multi-lingual terms.

ěAlias

Each thesaurus term also contains an indication of the language in which the term is specified (e.g. cn=projects, language=UK). This mechanism is used by the LIRN system to search and browse the thesaurus using the language selected by the user.

Below you can find a list of the most important advantages of this mechanism:

Distributed thesaurus
As the thesaurus is distributed over several systems, the load on any of these systems is reduced.
Specialised thesaurus
Specialised parts of the thesaurus, which might not be of general interest, can be kept on a local system. However, due to the underlying X.500 mechanism, this part is still available to the other systems and will appear as being local to all the other systems.
Multi-lingual thesaurus
The LIRN thesaurus is available in three languages. As these three thesauri point to the same concepts, it is not necessary to have direct access to all these languages in each country. Normally the copies are held on the systems located in the country from which the thesaurus originates (i.e. French in Belgium, Portuguese in Portugal and English in the United Kingdom).
Information replication
Distributing the information over several systems might cause a lot of traffic between these systems to exchange information. This can be solved by using the X.500 replication mechanism with which some more used parts, which are only kept by one specific system, are copied automatically at regular intervals (only when updated) to other systems. As a result of this, for each term in the thesaurus there is only one master copy. All other copies are updated from the master copy when it changes.
Thesaurus maintenance
Maintenance of the thesaurus (which can only be applied to the master copy) is also done using X.500 mechanisms and is therefore not restricted to locally stored terms.

Relation between information providers and thesaurus

The next picture describes the relation between the thesaurus and the registered information providers.

ěRelation

The LIRN System

So far we have only discussed the infrastructure and protocols supporting the LIRN service. All of this is however transparent to the user and could be considered as a black box that might have been implemented with other mechanisms, e.g. relational databases.

The only thing which is important to the user is the LIRN service. As we wanted this to be available on a wide range of platforms and as we were aiming for a high scaleability, we decided to go for the WWW approach.

The following picture illustrates the three LIRN service access points and how they are linked together.

ěLIRN

Currently there is one access point in each country involved in the project. Each of the sites is linked to the national X.500 and internet services. The LIRN User Agent (LIRN UA) communicates with the WWW server of the LIRN system and therefore it can be any WWW browser. For compatibility reasons the project is focusing on Netscape Navigator for Macintosh, MS-Windows and X-Windows and Lynx for simple terminals. A local Lynx [HREF 7] access point is available on each LIRN system and PPP/SLIP servers are being considered.

The LIRN system itself is described by the following picture.

ěLIRN

Access to the X.500 directory is provided through a specialised WWW to LDAP (Lightweight Directory Access Protocol for X.500 [RFC 1487]) gateway which has been developed within the project and which is operated on each LIRN site. The gateway handles all aspects of the LIRN service: thesaurus browser/search, resource search and access to some external services. The gateway also has a language switch with which it presents the LIRN service in the selected language. The language tables are dynamic and can easily be extended with other languages. The full language tables are available on each site but by default the gateway will select the local language.

Last, but not least, the gateway also has a switch for the search language (language used by the thesaurus). By default the search language will also be set to the local language but it can be changed in the course of a session. As the three used versions of the thesaurus (French, Portuguese and English) have the same coverage, this language switch is only available to allow the user to search using his own language. The underlying thesaurus structure, which is based on X.500, allows however to have a different coverage; e.g. a Portuguese library having only material in Portuguese might decide not to register its servers under French or English terms.

Conclusion

Although the evaluation phase of the project still has to start, we can expect a rather good acceptance of the system, as it has been built in close co-operation with the people for whom it has been designed. The most important aspects of the system are: the guided search (thesaurus driven), the open information base (all kinds of information providers can be registered), the open subject tree (the subject list can be extended), the high scaleability and the multi-lingual search facility.

Although originally aimed at a library environment, the LIRN system provides a good platform to support a general search facility for several kinds of network services (WWW, Gopher, FTP, Telnet, etc.). Further initiatives to setup a public registration, search and referral service based on the LIRN concept and developments, can and will certainly be considered.

More information about the LIRN project can be found in the deliverables which are available on the LIRN FTP server [HREF 8].

References

Geoffrey E. Cleave (9th Edition 1992) "SCIMP/SCANP Thesaurus", United Kingdom, University of Warwick Library.

Nils Meulemans and Andrew Colleran - LIRN 4.2 (February 1995) "LIRN DUA Development Specification", Belgium, University of Brussels.

X.500 (1993) "ISO/CCITT X.500 - Directory Services" Paul Barker (December 1993) "Project ABDUX - Accessing Bibliographic Data Using X.500", "Schema to support Searching for Information Resources and of On-line Archives", United Kingdom, University College London.

Giovanni Armanino, Antonio Blasco Bonito, Maurizio Martinelli, Giuseppe Alberto Romano and Giuliana Tamorri (1992) "An X.500 extension to provide a Database Information Service", Italy, CNUCE - CNR.

Nils Meulemans - LIRN 3.1 (June 1994) "X.500 Feasibility Report", Belgium, University of Brussels.

RFC 1487, "LDAP - X.500 Lightweight Directory Access Protocol".

Hypertext References

HREF 1
http://www.iihe.ac.be/ - Home Page of Telematics and Communications Service of the University of Brussels
HREF 2
http://sun7.iihe.ac.be/index_uk.html - Belgian LIRN access point.
HREF 3
telnet://blu@blu.brunel.ac.uk:23/ - British Library ABDUX Project Demonstration System.
HREF 4
http://www.dante.net/ - Home Page of DANTE
HREF 5
http://web.nexor.co.uk/public/aliweb/aliweb.html - ALIWEB
HREF 6
http://lycos.cs.cmu.edu - Lycos - The Catalog of the Internet
HREF 7
telnet://lirn@sun7.iihe.ac.be/ - Belgian LYNX access point for LIRN
HREF 8
ftp://sun7.iihe.ac.be/pub/ - LIRN FTP server

Copyright

© Southern Cross University, 1995. Permission is hereby granted to use this document for personal use and in courses of instruction at educational institutions provided that the article is used in full and this copyright statement is reproduced. Permission is also given to mirror this document on WorldWideWeb servers. Any other usage is expressly prohibited without the express permission of Southern Cross University.
Return to the AusWeb95 Table of Contents

AusWeb95 The First Australian WorldWideWeb Conference