2016-09-16 RDRIWG P8 Meeting

WG Research Data Repository Interoperability Session

Date: Friday 16th September, 13:30 - 15:00


  1. Introduction
  2. Presentation of the Case Statement review results
  3. Short talks on state of the art technologies
  4. Discussion about suitability and gaps
  5. Summary of results and next steps





  1. Outcome of the COAR Interoperability Project (Thomas Jejkal)
    • Focus on publication world, partly applicable to research data
    • Good state of the art overview usable for D.1
    • Good opportunity for collaboration, especially in the direction of technical interoperability issues


  • Check overlap regarding common metadata and vocabulary
  • Reference implementation highly appreciated
    1. Refreshing the SWORD Protocol (Dom Fripp)
  • Enable SWORD for (research) data
    • E.g. POST data by reference
  • Long term sustainability plan
    • Community maintenance best options in terms of long term guarantees and initial effort)
    • Alternatives: maintaines by JISC (no guarantee) or IETF/NISO (high initial effort)
  • Take next generation repositories into account
  • ResourceSync and SWORD integration planned
  • Comment at https://goo.gl/8E11wf


  • Reference implementation highly appreciated
    • For version 2 available at swordapp.org
    • Import/Export Standards for Repository Resources (David Wilcox)
  • New server-sided APIs not available in ‚legacy’ instances of repository platforms
    • Standalone tool using public interfaces could be used for old and new repository platform instances
  • Looking into BagIt as container format
  • No information about payload needed
  • BagIt Profiles extension to add payload semantics for later import
  • Apache Camel for RDF serialization
  • See https://wiki.duraspace.org/display/FF/Design+-+Import+-+Export for more details


  • BagIt also used in DataONE, support for OAI-ORE resources planned?
  • What is the effort for creating BagIt profiles?



  • What about domain specific repositories?
    • Outcome of WG should be applicable also for such repositories.
  • Will this WG try to find a consensus between repositories on resolving (persistent) identifiers?
    • No, identifier resolution depends on identifier.
  • Check PIT and Data Fabric output, e.g. to obtain content by identifier in a standardized way.
  • Check Metadata IG for common metadata elements
  • Think about flexible multi-purpose protocol (like HTTP)
  • Additional technologies to investigate/support: HubZero, DataONE API, Globus Publication, Frictionless data, R (as consumer application)