2019.02.13 Community Meeting - OCFL/spec GitHub Wiki

Call-in Details

Attendees

  1. Andrew Woods
  2. Daniel Bernstein
  3. Julian Morley
  4. Peter Cornwell
  5. Aaron Birkland
  6. Andrew Hankinson
  7. Ben Cail
  8. Dan Granville
  9. David Wilcox
  10. Doron Shalvi
  11. Neil Jefferies
  12. Rosalyn Metz
  13. Simeon Warner
  14. Tom Wrobel

Agenda

  1. Community updates
  2. Beta updates
  3. Client libraries
  4. Related initiatives
  5. Next meeting: Wednesday, Mar 13th @11am ET

Notes

1. Community updates

Fedora leadership group has expressed an interest in exploring Fodora-over-OCFL where OCFL is the persistence layer, see Fedora 6. Currently exploring a set of questions including how to map Fedora 4/5 and/or Fedora 3 models onto OCFL. There will be a face-to-face meeting of committers at the end of February to explore a way forward. Also exploring the idea of dropping some future Fedora over a set of generic OCFL resources to expose them in some way. See Fcrepo-OCFL Mapping: Background and Open Questions.

Oxford ORA expecting to start work on OCFL implementation end-March or early-April.

Data Futures (Peter, Dan) exploring manual creation of OCFL objects for about a dozen objects (scrolls with thousands of Open Annotation data model annotations), have looked at existing validation work. Find that OCFL structure is similar to previous work on proprietary filestore. Data available at https://ocfl-dev.data-futures.org/exports/hachiman/. Open questions of how to integrate archival store with live support for use of data and annotations. Have not yet explored assembling OCFL objects under a storage root. Comments and questions to [email protected] & [email protected].

2. Beta updates

Discussion of unresolved Beta issues and in particular unassigned issues.

Issue 281 - should be broken up into two tickets. Rosy will take the "Add 'base-ten' clarification", someone else needs to take on the "Add Unix ls discussion to implementation-notes" part.

Issue 275 needs discussion and agreement at next editors' meeting

Action on current assignees to resolve these tickets.

3. Client libraries

Question from Dan of whether Python code (currently zimeon-ocfl-py) will be fleshed out to implement more verification (such as content)? Oxford team not expecting to work on implementation before end-March but do want a Python implementation. Simeon says PRs welcome on his code. OCFL Slack channel is the best way to communicate about this.

Aaron's Go implementation birkland-ocfl was done to evaluate OCFL for Fedora work and in JHU context. Focus is process of creation of OCFL objects and to look at any performance implications. Has code for listing objects in OCFL root, files within object, and building OCFL objects. Question about ids and resolution: there is an expectation to have a mapping from an id (URI or any string) which is currently implemented using a simple URI encoding library in Go.

Andrew Hankinson has started some work on a Go implementation that focuses on validation: https://github.com/ahankinson/ocfl.

4. Related initiatives

NDSA Levels of Preservation

  • NDSA is currently out for consultation now so we can feed in
  • Rosy is on two committees (assessment and curatorial), comments that users are picking and choosing elements rather than implementing whole levels.
  • Neil & Julian at PASIG and will discuss with Sybil who is on core revision groups. Question of where OCFL matches (or not) some of the levels.

OAIS-IF

  • Initiative to find common technology for interacting OAIS systems.
  • p4-5 discussed different areas of the framework: one part is UI, then lower level abstractions
  • Might OCFL fit in at the bindings stage? Not clear from diagrams.
  • Andrew Woods will reach out to better understand possible connections.

eArchiving specification

  • See "specifications for review"
  • Not sure of possible connections to OCFL, consensus on call is to watch this work.

Audio recording

Action Items

  • ...

Previous Action Items

  • Neil to contact NDSA group regarding openness of preservation storage in the "Levels of Preservation"
  • Neil to contact Data-Futures regarding engagement