A Data-First Preservation Strategy: Data Management In SPAR

The Bibliothèque nationale de France has developed its trusted digital repository, SPAR (Scalable Preservation and Archiving Repository), as a data-first system. This implies having fully described collections, through use of metadata standards in the information packages, such as METS, PREMIS, MIX or textMD, in a way that will make sense given the diversity of our documents. The need for full documentation also applies to the system itself. On the one hand, SPAR is self-describing in order to ensure its durability. On the other hand, all the information that is ingested into the system contributes to determine its settings and its behavior. The Data Management module is at the heart of these information flows. We expect to push this data-first objective ahead by using RDF technology, based on existing and trusted information models and ontologies, such as OAIS and PREMIS. The challenges and successes we encounter all serve the greater goal of having a unique and versatile data model for every user of the system, whether collection curator or system manager.