4.1 Description: a cultural and technical process

Bibliographic models and international cataloguing principles have attempted to define bibliographic data (description) and authority data (access) precisely to organise information about the resources described.

Description is the primary procedure (in the sense that it is the procedure that is carried out first) of the cataloguing process and is foremost a conscious interpretation of data. It is analysis, evaluation, and reflects knowledge of what is being represented. Its object, therefore, is the resource being examined, which, in Domanovszky's words, still valid decades after their enunciation, can consist of either an elementary object or a main elementary object and secondary object (e.g. aggregates, attachments, supplements) (Domanovszky, 1975).

The description is aimed at the identification and characterisation of a resource, which is simultaneously a material object and an intellectual content – a dimension that has been referred to by many, including Domanovszky. Its inextricable and fascinating ontological duplicity underlies the question: ‘What exactly does the catalogue describe?’ The question is anything but simple and resolved. Rossella Dini writes, ‘What is the object of a bibliographic description? In other words, what should a distinct bibliographic record describe? No code or cataloguing manual has ever established this.’ (Dini, 1991, 135) The question still awaits a comprehensive answer, although FRBR (and later IFLA LRM) posed the problem in explicit terms and, according to Pino Buizza, studied ‘the object of cataloguing in a new, dynamic way’ (Buizza, 2002).

The record, hence, is a set of data that identifies and characterises an entity. Today, description is supported by automated processes (data capture) and will take the form of an ever-increasing enrichment of information already available on the internet. The issue of ‘whom to trust’ will be of vital importance. With the advancement of methods and tools based on artificial intelligence, it is conceivable that we will see a further evolution also in the field of metadata.

The faithful recording of data, i.e. how they are found on the source of information, should not be confused with a mechanical procedure. It is, rather, a method that respects how the data appear on the resource and are recorded, excluding manipulation: for example, transcribing in Latin characters or translating into English an original title in Japanese (as sometimes occurred in the past).