Saturday, 24 July 2010

JSON for linked data - migrating to RDF


Project partners meeting - 17-Jun-2010 (delayed post)

This entry was mistakenly posted to the wrong blog, so I'm belatedly moving it to it's rightful home, for the record.

We held a project partners meeting on 17 June, notes from which can be found at

The main focus for this meeting was discussion of our initial experiences with the OULS data repository, and much of the meeting was taken up with quite arcane technical discussion. Overall, it's looking very promising: we have been able to create test submissions through the provided web interface, and the API looks very straightforward for us to use by other means.

Monday, 19 July 2010

ADMIRAL Sprint 9 review and Sprint 10 planning

The review of sprint 9 has been published at The main achievements for this sprint were:
  • a very useful project partner meeting
  • substantially completed the Databank test suite, except for tests dealing with metadata merging
  • started work on RDF serialization for Shuffl, via JRON
  • initial page to display RDFDatabank contents - using jQuery and rdfQuery to create a user display from RDF data
  • brief project description published in D-Lib
We start sprint 10 with the unfortunate news that our second developer is resigning with immediate effect to attend to family commitments.  This will certainly impact achievements in the coming weeks, though we hope to maintain forward progress while we investigate and put in place alternative arrangements. This will surely test our risk assessment claim that an agile approach to development helps to mitigate staffing-related risks!

The planning notes and plan for sprint 10 are at:
Out focus for the next sprint will be:
  • RDF Databank test cases for metatada merging
  • finish RDF Databank web page to display dataset details
  • progress deployment of ADMIRAL data store with other research groups
  • progress RDF/XML serialization for Shuffl data

Wednesday, 7 July 2010

ADMIRAL Sprint 9 plan

The ADMIRAL sprint 9 plan has been posted at

This was a very simplified planning meeting for a short sprint, with a focus on scheduling priority tasks from those already planned.

For the next sprint, we aim to:

  • Complete test suite for accessing ADMIRAL databank
  • Progress RDF creation using Shuffl
  • Arrange meetings with Development and Elephant groups

With a view to working towards a front-to-back, researcher-to-repository, test deployment, including:

  • capturing dataset descriptions as RDF
  • packaging data+description and submiting to the RDFDatabank test system

Tuesday, 6 July 2010

ADMIRAL Sprint 8 Review

Project activity has been impacted by unplanned staff absences, but useful progress has been made on a number of technical fronts. Interaction with Library Services to develop the data repository service has been going well, and we look forward to rapid progress in this area. Work to add RDF serialization to Shuffl is under way. Rolling out ADMIRAL data stores to additional research groups has been stalled.

This review is being posted nearly a week later than planned, so includes a period not covered by the original sprint plan.

Summary of achievements this sprint:
  • project partner and OULS meetings
  • attendance at JISC TransferSummit meeting
  • brief project description for D-Lib magazine
  • Shuffl WebDAV file browsing implemention complete
  • RDFDatabank access
  • RDFDatabank initial test suite
  • negotiating revisions to RDFDatabank API in light of experience
  • started on RDF serialization for Shuffl

SWORD white paper: relevant to ADMIRAL?

I've just read through a white paper about directions for the SWORD deposit protocol:

I'm recognizing many of the discussion points we've been having about the submission API for ADMIRAL to the library service RDF Databank appearing here:

  • submitting datasets as packages of files
  • selective updating within dataset packages
  • accessing manifest and content
  • accessing metadata about the package
  • etc.
I'm not advocating at this stage that we should be trying to track the SWORD word, but I do think we should try to ensure that noting prevents us from creating a full SWORD interface to RDF databank at some stage in  the future.

Friday, 2 July 2010

RDF implementation for Shuffl, via RDF-in-JSON

I'm throwing my hat into the RDF-in-JSON ring, planning an implementation of RDF/XML serialization for Shuffl workspaces and cards using RDFQuery [1] and JRON [2].  My initial mdesign notes are at