It is widely accepted that trustworthy repositories are key to build persistent and functional data infrastructures. At the plenary P6 in Paris we differentiated between 3 types of repository registries:
- Collection Registries that easily indicate human users what kinds of collections are being stored by a certain data centre
- Human-readable registries of (trustworthy) repositories that store relevant data (here re3data should be supported)
- Machine (and human) readable registries that store all kinds of detailed important information about repositories and their data services
It was also suggested to have one schema that can be used by repositories to generate the required information allowing various services providers to extract the information they need to maintain their services.
This session will focus on discussing the WG proposal (needs to be submitted in January/February) on Machine Readable Repository Registries (M3R) for automatic processing in large federations and its major characteristics.
It was obvious that this discussion should not be mixed with well-known registries aggregating metadata of data objects and collections, since different aspects need to be taken care of. However, repository registries will specify information about the ports that can be used to harvest metadata by services providers.