Harriet W. Harrison, President of the FIAF Cataloguing Commission.

As Chairman of FIAF's Cataloguing Commission, I read last issue's Guest Editorial by Geoffrey Nowell-Smith on the European Filmography with great interest.

I had, of course, known about ACCE/Projecto Lumière's efforts to establish a database for films from EEC countries, and had met and talked with Nowel-Smith. All the Commission members have been following developments on this project, which so closely relates to the Commission's work. Last issue's Guest Editorial both helped to clarify some questions, and raised some important new issues and areas of concern.

The filmography seems to have gotten off to a good start. Nowell-Smith has recognized the complexities caused by major gaps and inconsistencies in existing data, as well as by the several languages in which the data is held. He has identified a number of data elements which he wants incorporated into the database and has decided to adapt and use a standardized exchange format, UNIMARC, which was developed by librarians as a means of equalizing the ways in which data are held in various extant sources which will be incorporated into the database.

These are the first steps, but more questions remain. Nowell-Smith alludes to some of them when he talks of the necessity of utilizing filmographic experts to rationalize differences in the information itself. His comment is, I think, a recognition of the limited functions which formats can play in the building of databases. Formats in general, and UNIMARC in particular, serve the function of setting record structures and providing content designation for fields and subfields, etc. What they do not do is to regulate the contents, the actual data in records, and this is the very thing which is necessary to achieve in order to reach the goals of a unified and functional database.

The work of standardizing data - the information we actually want from the database - has traditionally been the function of another set of standards, those which librarians have termed cataloguing rules. Cataloguing rules analyze information, and name and define the pieces of information contained in bibliographic records. Additionally, they prescribe ways for conforming data, functioning as a set of normalizing conventions which serve to guarantee the predictability and comprehensibility of the retrieval results. For example, cataloguing rules tell me how to choose a main title or a nationality or date from among several possible alternatives, thereby assuring to the greatest extent possible that both you and I will catalogue the film in the same way and that the researcher with his request will find both your information and mine.

Contrary to what is commonly thought, librarians have learned that the introduction of automation into the cataloguing process has brought requirements for increased exactitude and logic within cataloguing conventions. As databases become larger and contain data from a greater number input sources, the greater becomes this need for adherence to standards. As Canadian librarian, Tom Delsey has explained,
"The application of computer technology to the cataloguing process has brought with it increasingly exacting demands for precision and logic in the recording of bibliographic data. With the advent of the machine-readable catalogue record and computerized processes for record matching, filing, and indexing, the level of tolerance for variations in format and for literal and logical inconsistencies within the catalogue was radically diminished"(1).

To explain the situation in another way, it is the conventions which assure me that I have found everything in the database in which I am interested and, conversely, that I have not retrieved an unmanageable or frustrating level of unwanted items, commonly known as "false-drops".

In the arena of moving image cataloguing, the FIAF Cataloguing Commission has designated cataloguing rules for use as normalizing conventions in the building of inter-institutional databases. The first major result of our efforts was the FIAF Cataloguing Rules for Film Archives (Munich, K.G. Saur, 1991). The rules contained in this volume were based upon the very considerable work of creating cataloguing standards already contributed by the International Federation of Library Associations, the series of International Standard Bibliographic Descriptions (ISBDs), now comprising a family of standards for separate types of materials, e.g., monographs, serials, cartographic materials and non-book materials. The very purpose of all this effort at designing rules, i.e., normalizing conventions, was to create an international environment in which records created by one organization could be understood and utilized in another. The work of rules elaboration undertaken by our Cataloguing Commission was to harmonize the principles of ISBD with the particular needs of film archives. The FIAF rules vary from ISBD precisely in defining and prescribing means of relating, both physically and bibliographically, all the varying copies, issues, and versions which typify holdings of moving image works - the complexities of which are alluded to by Nowell-Smith in his examples about Voyages en Italie and Le Mépris.

Once conventions have been established, the variety of sources such as catalogues, files, databases and published texts, which will comprise the centralized file must be conformed to the conventions and then converted to the centralized database format - in this case UNIMARC. This process requires not only filmographic analysis and revision, but also computer programming, i.e., the writing of software for the conversion of each separate machine readable format represented among the sources. Moreover, those sources which are not already machine readable must be character-scanned, keyed, or otherwise converted to machine form. This whole process may be termed the execution stage of database construction and is highly resource intensive. Indeed, there is a real danger of seriously underestimating the time and energy required on the part of a large number of highly skilled and knowledgeable persons to bring the database to fruition. Numerous database plans have foundered at this juncture, and it is well to remember that the efficacy of the database is directly dependent upon the time and energy spent in normalizing the data.

Commission experience with the issue of identification is paralleled by experiences in the book library profession. Author, Ronald Hagler, in his article on "The Consequences of Integration," points out:
"The variety of cataloguing practices and errors evident in every large file, and the variety of bibliographic circumstances which can confuse the identification of two different bibliographic items, make it impossible to automate the task (of merging retrospective files) using only the natural language of authors, titles, etc."(2).

Thus, we would caution that the European filmography will require data normalizing according to common cataloguing conventions and conversions programs, in addition to the choice of a common format in order to achieve its goals. We would further recommend that the filmography participants, being mostly FIAF archives, follow the cataloguing conventions set by the Cataloguing Commission in the FIAF Cataloguing Rules for Film Archives. Both programming and data normalization are labor intensive, and the costs of achieving them should not be underestimated.

In a different vein, I would note that I am confused about the degree to which the filmography is projected to be used as a programming tool. To be truly useful for this purpose, it would seem that the filmography ought to include information about holdings. As every archivist knows, the quality and kinds of film holdings are subject to change, and therefore mechanisms for submitting and incorporating holdings updates should be included into the database design. An additional desirable function might be a lending/tracking module capability.

NO2
Ronald Hagler, "The Consequences of Integration" in The Conceptual Foundations of Descriptive Cataloguing. (San Diego: Academic Press, Inc., 1989), p. 203.

Filmographie Européenne

En réponse à l'éditorial de Geoffrey Nowell-Smith paru dans le bulletin précédent, Harriet Harrison, présidente de la Commission de catalogage de la FIAF, soulève quelques interrogations concernant des points de méthodologie auxquels pourraient se heurter les chercheurs de l'Association des Cinémathèques de la Communauté Européenne (ACCE) chargés d'établir la filmographie européenne dans le cadre du Programme Lumière.
Les commentaires de Harriet Harrison devraient permettre à G. N.-S. et à ses collaborateurs de vérifier leurs hypothèses de travail.

Filmografía europea

En su respuesta al editorial de Geoffrey Nowell-Smith publicado en el Boletín precedente, Harriet Harrison, presidente de la Comisión de catalogación de la FIAF, plantea algunos interrogantes sobre las dificultades de orden metodológico que podrían encontrar los investigadores de la Asociación de Cinematecas de la Comunidad Europea (ACCE) encargados de llevar a cabo la filmografía europea patrocinada por el Programa Lumière.
Los comentarios de Harriet Harrison permitirán a G. N.-S. y a sus colaboradores la verificación de sus hipótesis de trabajo.