...
- Announcements
- OCFL Community Call Update
- Landing the Sprint
- Set sprint closing meeting time
- Issues in progress
- Issues in review (PRs)
- Status of Search / Containment Index Design
- Handling RDF in Fedora 6
- Decisions and Open Questions
- Establishing clear description of F6's RDF (on disk, from non-Fedora OCFL, what changes, etc)
- Trim `ldp:contains` and other SMTs
- Normalize URLs to internal URIs
- Normalize serialization
- Establishing clear description of F6's RDF (on disk, from non-Fedora OCFL, what changes, etc)
- Draft functional requirements for repositories based on NDSA Levels of Preservation 2.0 (update)
- <Your question/topic here>
...
- Announcements
- OCFL call - discussed a couple issues, planning to add some language to OCFL spec or implementation notes. Stanford has a new validator (Ruby gem).
- Desire to make sure Fedora 6 runs properly in the cloud. Can test S3 on OpenStack. Need details on what exactly people want to do in the cloud.
- ocfl-java client - starting to add support for S3 (incomplete). More complicated, may have more requirements (eg. DB).
- Landing the Sprint
- sprint closing meeting - 1pm EST (Friday)
- issues in progress/review
- get to the point of running Fedora 6 - put something in, get it back out
- should make some good progress on PRs today. Things should be pretty clear. Get PRs merged even with some cleanup to do afterward.
- 3129 - implement get headers. Put together a proposal for fcrepo sidecar, then work on ticket.
- 3119 - try to get that in, base 3129 on 3119.
- migration-utils - things are progressing, single-stream.
- transaction piece -should be wrapped up this afternoon
- search/containment index design
- realigned with proposed interface, writing code to retrieve data from DB - WIP PR. Discuss later today.
- handling RDF in Fedora 6
- shared doc about handling server-managed triples, and OCFL objects from outside Fedora.
- 4 approaches: ignore SMT, how much to interpret RDF/directories on OCFL objects. How Fedora would behave in different cases.
- leaning toward ignoring SMTs, but still discussing.
- working on follow-up document.
- discuss more in Part 2.
- is it possible to push json-ld in, and get it exactly back out? Agreement to trim SMTs.
- info:fedora URI from Fedora 3 - could make it configurable to support multiple use-cases. Optionally translate URI prefix.
- roll out simplest things first, then add more features.
- should RDF files be containers? what should migration-utils do? Discuss in doc.
- shared doc about handling server-managed triples, and OCFL objects from outside Fedora.
- design/open questions
- object deletions - root of Archive Group - add a tombstone, then delete ocfl object when tombstone is deleted?
- F4/5 - when object is deleted, you get a tombstone, and all the info is all lost
- what would other backends support? It's not part of the API. Could be specific to OCFL backend to allow bringing back an object that was deleted.
- could provide access to version history of tombstones, or not.
- Are people OK with the memory of a deleted object still being there?
- Could fit somewhat with Fedora 3 'D' state.
- object deletions - root of Archive Group - add a tombstone, then delete ocfl object when tombstone is deleted?
Actions
- Aaron Birkland to look explore notion of OCFL client with database as authoritative metadata source + asynchronous writing of the inventory.json file
- David Wilcox will review the NDSA matrix and pull out the concrete technical requirements that could be considered during the Fedora 6 development.
- Call for comments on https://docs.google.com/document/d/18rSFqqoxixmozZrgPKON59Ojpg5iOu4lsHDNMm23soY/edit# till tuesday
- Clarify in in documentation that multiple simultaneous writes to OCFL are not supported
...