Data processing for digital libraries: the experience of the BnF with Europeana Sounds project

Abstract

As one of the 24 partners of the project Europeana Sounds, the BnF provides to Europeana metadata related to digitized sound recordings from its cultural heritage collection. This article describes the BnF experience with the Europeana Sounds project by taking as its point of departure the multiple challenges related to data transformation from base material, the MARC XML records that comprise also metadata related to digital objects, into the RDF based Europeana Data Model (EDM). By way of this experience the BnF is exploring technical methods for exposing and sharing in the best possible way the richness and granularity of its metadata, as well sharing its areas of expertise, such as controlled vocabularies and best practices with persistent identifiers. The array of issues addressed involves instructions that combine analysis of traditional bibliographic description, digital library metadata curation, critical look on the extant relationships between various data repositories within the BnF, current and future role of library catalogues, metadata production workflows, role of metadata quality, licencing issues, and more. The article discusses the organization of work so that skills and responsibilities are shared in the most efficient way. It also discusses lessons learned and perspectives regarding the evolution of the profession of the metadata librarian, including the need for a better understanding of the potential for the metadata to be reused and reprocessed, and the necessity to develop modeling and technical skills. Finally, in exploratory mode, the article highlights the possibilities offered by the Linked Open Data technologies for metadata reprocessing in innovative ways by taking down the data silos borders.