Digital Curation Day 2 – Jane Kwok and SCOPE
For me the highlight of the late morning session was Jane Hunter and colleagues (Queensland) describing their SCOPE system of managing Compound Document Objects (CDO’s). Jane is a materials scientist turned informaticist (hope that’s fair) and we’d already been partners on a grant application (didn’t get it, but …) – so this was the first time we’d met Jane did a double act with her colleague Kwok Cheung and inter alia gave a demo of their SCOPE system – which manages CDOs while preserving an imporeesive metadata structure in RDF. Some bullets: Increasing pressure to share and publish data while maintaining competitiveness. Main problem lack of simple tools for recording, publishing, standards. What is the incentive to curate and deposit? What granularity? concern for IP and ownership Current problems with traditional systems – little semantic relationship, little provenance, little selectivity, interactivity , flexibility and often fixed rendering and interfaces. no multilevel access. either all open or all restricted usually hardwired presentation Capture scientific provenance through RDF (and can capture events in physical and digital domain) Compound Digital Objects – variable semantics, media, etc. Typed relationships within the CDOs. (this is critical) CDO has Identifier URI – can sit in different places. SCOPE has a simplified tool for authoring these objects. Can create provenance graphs. Infer types as much as possible. RSS notification. Comes with a graphical provenance explorer. Have developed an ontology for Relationships. Rules for inferences are coded in SWRL and run with Algernon. (PMR: Note, must look into SWRL seriously – we should/could use it). Jane had picked up on the idea of datuments that Henry Rzepa and I coined a little while back and has suggested we expand these with semantics. Basically a datument is a compound document in XML, often with several namespaces and a separate rendering approach. She suggests a semantic datument – presumaby XML with embedded RDF. Sounds great.