Repository Preservation Infrastructure (REPRISE)

While object semantics are essential for interoperability within and between information environments, the community has developed various mechanisms to semantically annotate digital objects and deal with heterogeneity in these semantic annotations. The workshop therefore took these mechanisms for object semantics as a given starting point, and focused on the semantics of preservation services, architectures and organisations.

The workshop explored various approaches to design preservation environments such that they can deal with heterogeneous material, evolve over time, and be open to share services with other preservation environments: the focus on human actions ("verbs") rather than technical services in ANDS (Treloar), the micro-services in UC3 (Kunze) focussing on the robust basics in file systems, as well as a Java-based, cross-repository interoperability framework by Planets (King). A coherent and trustworthy preservation system emerges from these building blocks through horizontal mechanisms that are both technical (such as provenance, Missier) and organisational (audit and outsourcing, Ashley).

 One of the discussion points that remained unresolved addressed the balance between an open architectural design and external guidance (standards, audit) to achieve interoperability and trust. Essentially this boils down to the question whether information environments and collaborations between information environments require a minimum level of hierarchy and control to ensure quality, or whether the model of the web should lead us to decentralised environments that are able to scale and evolve over time. This and other issues that span technical issues as well as local and community organisations remain to be discussed in the future.