Time/Place
This meeting is a hybrid teleconference and slack chat. Anyone is welcome to join...here's the info:
Attendee
- Danny Bernstein
- Andrew Woods
- Jared Whiklo
- Peter Eichman
- Aaron Birkland
- Daniel Lamb
Agenda
- Announcements
Barriers to Migration Survey
- Key take aways
- Fedora 3 File System Layout is popular
- People like Fedora's linked data support
- API is popular with users
- Lack of staff time and good migration tooling present biggest barriers for migration
- 2019-05 Fedora Camp Atlanta
- Fedora 6 Sprint Doodle
- Fedora 5.1.0 Release
- 5.1.0 Open (is triage complete for inclusion in 5.1.0?):
- 5.1.0 Closed
- 5.1.0 Open (need to re-evaluate for inclusion in 5.1.0):
- Import Export
- Fedora 6.0.0 Sprint planning
- High-level roadmap
- Tactics
- Development Roadmap
- Pilot Program
- Contracting
- <Your agenda item here>
Tickets
In Review
Please squash a bug!
Tickets resolved this week:
Tickets created this week:
Notes
- announcements
- barriers to migration survey
- why are people still on F3
- people like the idea of linked data
- sharing data is valuable, people recognize that
- we do hear a lot about how people don't like linked data
- main pushback sources:
- people designing against the API (e.g. Samvera), may not be as easy as other models
- people who want to have a non-exposed/buried Fedora
- linked data is more useful/popular with users than developers
- general approach: emphasize linked data is optional
- there is value in showing how to replicate F3 models in F4+
- Jared has a presentation for OR about Fedora without linked data
- missing features in F4
- query support
- validation in the repository
- field ordering; difficult in RDF
- file system layout in F3 is very possible
- Fedora Camp
- there is still room
- encourage people to come if it would benefit them
- Fedora 6 sprint Doodle
- dates when Danny and/or Andrew are available
- ideally 3 sets of 2 week sprints
- please fill out the dates when you think you could be available
- once we have dates, come up with broad goals for each sprint
- Fedora 5.1.0 release
- Jared is release manager
- raised and resolved a couple issues regarding binary descriptions (FCREPO-2996, FCREPO-2997)
- state token issue merged
- 3 outstanding bugs
- been open a while
- may be difficult to fix under Modeshape
- Andrew will have PRs for FCREPO-2782 with DuraSpace checkstyle rules
- contains suppressions for new DuraSpace rules Fedora doesn't conform to
- future conversation about whether to adopt these
- release buildtools with updates
- then update the maven stuff to use the new release of buildtools
- 5.1.0 RC once checkstyle updates are in place
- goal is RC this week, 3 weeks before OR
- 2 week window for community review
- import/export
- Mohamed started testing at UMD
- there might be a way to import a F4 export to F5 if we can ignore ACLs and versions
- UMD is not using versions, and only has a tiny number of ACLs
- no work happening on Danny's end next week (Fedora camp)
- maybe target week after next
- process:
- export with F4 tool
- remove ACLs (by hand) from exported files
- import with F5 tool
- not a formal release per se
- wait for F5-to-F5 round tripping to release
- resumability is a huge feature for UMD
- borrow features from plastron, UMD's batch tool
- use an embedded SQL database to track progress?
- Fedora 6 sprint planning
- David put together a public high-level roadmap
- there is also a committer-generated development roadmap
- F6 high-level roadmap
- synchronous query, i.e., synchronously updating index for querying
- is this for Samvera?
- synchronous is not in the design document
- maybe it means that they want a bundled query service
- want a query service that talks to the same store as persistence; guarenteed consistency
- F3 had a simple synchronous query service
- indexed a defined subset of content
- F3 also had integration with triplestore, could be either synchronous or asynchronous
- need to pilot these query services with folks who have these use cases
- determine to what extent synchronicity is required
- could there be an optional way to enable synchronous indexing? header? config setting?
Actions