4.5 Updates and versioning

Sometimes depositors send updated versions of their data. The ingest team in an archive needs clear rules what adaptations have to be made then and how and where these are recorded. Maybe all files need to be exchanged, or maybe it is just the data that needs to be replaced. How to handle this in the OAIS model must be clear so that integrity of data and documentation material can be assured. With regard to the assurance of understandability of the changes, how is this guaranteed? Is an errata/log/curation file necessary?

It is strongly advised that the ingest staff track the provenance of all files that the repository receives and stores in the archive. Re-naming of the files, generation of new files and the means to do the adaptions should be recorded.

The ingest team needs a clear structure on naming of different versions. It must be clear and reproducible for the data curators what are the most recent data in the DIP and also in the SIP to not mix things up by accident. Does your archive implement a “changelog” folder? Or are the folders named after the respective version? Or are they named after the dates?

Versioning of the data entry in the repository solution can differ according to minor or major changes to the archive material. In Dataverse, e.g., small metadata changes lead to minor versions changes whereas substantial metadata changes or data file changes lead to major version changes. Versioning usually is tracked in the used archive system and detailed information on the changes is presented.