Validating the completeness of an archive with an external reference list. The Publications Office experience

Trustworthy repositories of archival data need to be reliable and ensure the authenticity of their holdings. However, one aspect that is not often addressed is the verification of the completeness of these holdings. With this lightning talk, the Publications Office will share its experience on how it ensures that the collections it archives are complete. The Publications Office of the EU is the central point of access to EU law, as well as to publications, open data, research results, procurement notices and other official information. It plays at the same time the roles of a publishing house, a legal depository and an archive for the publications authored by the EU institutions. Among its digital collections, the ‘Official Journal of the European Union’ occupies a central place, as the digital version is legally binding since July 2013, replacing the paper edition. It thus requires particular attention in terms of authenticity and completeness. Until now, the control of the completeness was ensured by a list established at the entrance of the archive. The filenames and checksums of all the documents ingested in the long-term preservation repository are kept in a list that is regularly compared to the inventory of the assets contained in the repository, and to their index. This way, we check if all ingested documents are still present over time and after preservation actions such as migrations, media refreshes, etc. Recently, we made a significant change to our IT architecture by creating a reference list that is physically and conceptually separate from the archiving workflow. Most importantly, this list is fed at the entrance of the documents into the dissemination repository, i.e. at the moment where they are published (and not when they are archived). With this change, we can carry out a regular check that our archival assets are complete and intact, by comparing the filenames and checksums from the inventory to those of this new reference list. In short, we have established a reference list that tells the archive what it SHOULD contain, and no longer simply what the archive has received.