FRBR: SOME COMMENTS BY ELAG
(European Library Automation Group)

Versione italiana

This brief presentation gives an overview of the analysis by ELAG of FRBR. The Group has made a summary examination of the IFLA model but only some technical aspects have been tackled. ELAG members are librarians and information specialists, computer experts and system analysts who for more than 25 years have been working on issues of information technology in bibliographic services. (Participation is free and new members are welcome. The next annual meeting will be held in Paris, at Musée de la Villette, from 12th to 14th April and the subject will be: Integrating library, museum and archive collections). From early 1996 the Group started to examine the FRBR study in one of its workshops and interest has been constant since then. On the basis of the Group's reflections some documents have been produced: among them are a written declaration of intent addressed to IFLA for further technical investigation of the model and a study project proposal to the European Commission for the application of FRBR in the real world. The proposal was quite complete and detailed, but due to organisational and time schedule problems, it was not sent for approval, despite the consensus of the Library Project officers in Luxembourg.

This short report gives an outline of what has been elaborated so far.

ELAG work on FRBR is based on the belief that the reasons to launch the FRBR study identified 10 years ago at the Stockholm Seminar are even more important now, the reasons being mainly the rising costs of cataloguing, increasing drive to reduce duplicate efforts, growth of publishing outputs, proliferation of digital materials and the need to integrate them with actual bibliographic collections. It is then essential to work for the development of FRBR.

Three main issues have been addressed by ELAG: the structure of the model, with its inadequacies as well as its strengths, user benefits, practical implications.

1. The IFLA model and its structure

Its design is not only of a high theoretical value, but also a practical one and it is necessary to start working on its experimental implementation in the real world of library systems, evaluating benefits and shortcomings. The model has the potential to make the bibliographic records independent from the language, culture, and country of cataloguing. By separating the cataloguing elements and tagging them by language/culture/region identifiers, it allows automatic delivery of the culturally correct version of the object to the end user. By defining different interrelated entities it is possible to assign the correct information to the different objects: each of the entities of the 1. group is by consequence an object which derives from the preceeding one and inherits its properties according to requirements of Object Orientation methodology. ELAG discussed how such approach could be used to develop an advanced bibliographic model. As a consequence of the inheritance principle coherence of representation and economy can be achieved when documents are analysed, certainly in a long term perspective.

The examination of some entities has shown that some relations may be realised in different ways: For example when different works are related to each other, it is possible to link them by horizontal relationships and leave it to the retrieval system to follow the links; another solution is to define a 'common top' object (a sort of 'archetypal' original work) for closely related works when very complex cases occur, thus allowing for a quicker and more precise retrieval. It is obvious that this is not an easy procedure.

As regards expression (in agreement with the AIB Study Group on Cataloguing) the accumulation in this entity of different types of expressions, such as various editions, translations, parallel editions, forms, causes problems. Also determining relationships is a complex activity in that relationships between expressions are not always horizontal, owing to the fact that an expression can be based on another (a translation made on a specific edition or on a translation). Furthermore it is useful to group the expressions according to their form (text, sound, braille...) or according their medium (paper, electronic...). An attribute in the expression description would take care of this functionality thus allowing a correct identification of it and a consistent relationship with the others.

Concept-subject heading. At the moment the model does not intentionally include (the same happens for persons-corporate bodies) any structure for this entity. When applying the model in the real world the various levels and relationships between concepts and terms should be defined. Besides that, the link is made to the work, as well as to the expression? And can the expression be designated through different concepts from the work?

Mandatory entities and library systems. The IFLA model has more entities than most systems today. The actual library systems usually handle manifestation and item since cataloguing is done mainly at manifestation level. Systems will have to include the levels for work and expression and implement the necessary functionality and control procedures.

Copyright. Not only can the catalogue become the tool for dealing with the copyright of documents, but also the property rights of records themselves should be handled. The question is raised then who owns those rights, how are they managed and if copyright for records at each level can be realised.

2. Benefits of the model

FRBR allows a greater clarity and precision in the retrieval and understanding of the relationships, thanks to the placement of bibliographic elements at the right level of abstraction, while at the same time ensuring an increased flexibility of representation. The model structure makes its maintenance easier and an accurate processing can be achieved, according to precise rules. The 4 level bibliographic record is very much in line with the results of the analysis of digital resources, while making it possible to integrate them with 'traditional' material.

Benefits for end users. Searching and retrieval functionality will be improved. Individual objects (i. e. paper documents or video) can be grouped under an individual work. Through the OPAC it will be possible to navigate between different objects, for example displaying only works or expressions and consequently, following the downwards links, retrieve the other entities.

Thanks to this framework, integration can be realised between different media and worlds which are still isolated from each other. That is to say merging library catalogues with museum and archive catalogues, thus offering the user increased information capabilities.

As regards cataloguers and library staff in general, the precise placement of attributes to the various entities will further improve bibliographic exchange and shared cataloguing: this could now be possible at the individual levels rather than as a complete self contained bibliographic MARC record. The records for the different entities will be smaller and mapping from a format to another is likely to be easier.

3. Practical implications

Record management. Once the model has been implemented, record management and control will require more efforts since the records will be segmented and inevitably dynamic, as well as their relationships. The history of records will be important to keep and this raises the question of who should be responsible for keeping such history.

Cataloguing. More data must be entered into the systems than today and a supplementary work is required for the attributes of works and expressions as well as for the relationships between works, between works and expressions and manifestations. The task for preparing records of works and expressions should be centralised and this data should be made available for all cataloguers.

Library systems. Almost all current systems are 'single record' based, despite the spread use of DBMS. Those few systems which have implemented a different bibliographic model (SBN, Servizio Bibliografico Nazionale system is one of these) use a scheme which is approximately that proposed by the 4 level model. Modifications of library systems will involve mainly cataloguing and searching modules in order to make a better use of data and relationships. An investigation of the minimum and maximum work needed to utilise the model is necessary and the problems of converting the systems should be tackled.

Cataloguing rules must be revisited, mainly in the overall organisation of their specifications and recommendations.

The traditional formats must be modified to incorporate the relationships and the characteristics of the individual entities. UNIMARC, with its emphasis on links, seems to be the closest format to the model. The new or readapted formats should be structured as XML documents to promote interchange not only within the library community, but more broadly.

Preparation of examples based on FRBR. Preparation of examples by a small group of ELAG volunteers who have tried to hesitantly venture to complete the IFLA model structure, has shown inevitable uncertainties and difficulties, due not only to the unavailability of information on the bibliographic history of a work, but also to the problems encountered in assigning some relationships, especially between expressions, and in placing attributes at the various levels.

Therefore much work has still to be done in different areas and ELAG is ready to contribute for the development of the model.