Our interest is a very neutral, pluggable approach to both identifiers and metadata. There are several aspects worth noting, and our hope is that these activities can proceed in a way that is informed by the work of this group!

Metadata Templates

There is an ongoing working group on metadata templates, and a specification that is being reviewed in the iRODS Consortium based on DFC work. This template mechanism allows structured metadata and validation rules to overlay the unstructured iRODS AVU mechanism. Using this mechanism, elements may be assembled into a template with type information, validation rules, mappings to support the following use cases:

Allow formatted display of metadata records for interfaces
Allow human curation of metadata through responsive forms
Allow export/crosswalk of data managed by the grid to various formats and schema
Allow automatic validation of metadata records at entry, or via policy enforcement

The metadata templates can be bound as required or advisory to various collections, and result in base iRODS AVUs that map to templates, allowing merging, automatic extraction, etc. Treating AVUs attached to files, collections, zones, users as statements about those resources and representing the contents of the grid as an RDF graph has already been demonstrated.

REST and HTTP resolution of resources

Existing and planned development on both REST API, as well as web based representations of catalog contents (collections, users, files) provides HTTP resolvable references to resources on federated grids. These HTTP based representations are prime candidates for the proposed NDS project, and we would like to weave in JSON-LD or microformat representations of linked data

Indexing

Data within the grid may be exposed to indexers at the collection level, and catalog data and metadata can be projected into external indexes (Elastic Search and triple-stores have already been demonstrated). As that work proceeds it will lead us further into the ability to work with linked data representations of DFC catalog holdings.

PIDs

PIDS are being addressed in a regular fashion through pluggable mechanisms, so that identifiers can be assigned and collections can be published as immutable references, exposed through the aforementioned endpoints, and carrying metadata through the metadata templating mechanisms. PIDS are treated as metadata and plug-in mechanisms for Handle, EZID, and other sources of identifiers can provide resolution and assignment services to grid users.

Summary

Our focus is on an extensible framework that can be used in various environments by the policies, metadata templates, and configurable support for identifiers and 'publishing'. We would like our HTTP endpoints to respond to requests via identifiers, and provide catalog metadata in formats that enhance linking and sharing, and look at this subproject as a worthwhile venue to standardization of our architecture!

Browser not supported