Time/Place
This meeting is a hybrid teleconference and slack chat. Anyone is welcome to join...here's the info:
- Time: 11:00am Eastern Daylight Time US (UTC-4)
- Audio/Video Conference Link: https://lyrasis.zoom.us/my/fedora
- Dial-in:
+1 408 638 0968
+1 646 876 9923
+1 669 900 6833
Meeting ID:
812 835 3771
- Dial-in:
Join fedora-project.slack.com on the "tech" channel
Attendees
Danny Bernstein- Jared Whiklo
- Ben Pennell
Ben Cail- Andrew Woods
Thomas Bernhart- Bethany Seeger
- David Wilcox
- Aaron Birkland
Peter Eichman- Peter Winckles
Daniel Lamb
Agenda
- Announcements
- Fedora Demo
- ..maybe wait on versioning? - or not. Auto-versioning works
- Backwards compatibility document
- Road to Fedora 6
- Milestone list
- Getting sprint dates on the calendar
- Priority tasks (Development Roadmap : 2020 Sprint Roadmap)
- To review:
- <your topic here>
Tickets
In Review
Please squash a bug!
Tickets resolved this week:
Tickets created this week:
Notes
- Announcements
- David Wilcox mentioned applying for grant from IMLS. Currently working on a full proposal that would be a follow on grant to the migration grant from last year. Looking for institutions to work with (as a pilot). Contact David if interested. Goal is to proof out fedora 6 and build a bunch of documentation / training to create more of a path for migration for the community at large. https://drive.google.com/file/d/1vbgn_OWChV1u6oYdzQ3-yT4Chwv0nmgc/view?usp=sharing
- Will grant consider Fedora 3 → OCFL (outside of Fedora)? Some might use the migration to get to OCFL w/ or w/o Fedora. Work will be valuable to anyone going to OCFL. Would be good to know if Fedora3 users who are interested in a more generic OCFL migration tool that's not Fedora specific. The migration tool does create generic OCFL - it be good to make sure the Fedora code that rebuilds the index from OCFL doesn't depend on .fcrepo directory. Feasibility of that? Feasible, but a slightly different start point - one with fedora data (.fcrepo) and one with out. Right now rebuild logic is planning on .fcrepo being there, so it'd be a shift. There would have to be some assumptions made to read that OCFL data. Would be good to infer different characteristics from generic OCFL.
- A separate, stand alone tool would be useful, as it might be really hard w/o input for Fedora by itself to reason over data. Concern is if we load the generic OCFL data wrong, one can't change the interaction model. Maybe some sort of migration tool that looks through the data, that catches issues and asks user, but that could be tedious. Maybe build up a decision tree for their data, so user can have some input. Build some sort of reasoning over a corpus of their OCFL data. Find a way to tell fedora how to do it's decision making over the users data.
- The migration tool takes F3 → Generic OCFL. We could use this output to test out what issues Fedora 6 falls into when rebuilding from that OCFL to get a better sense of what the issues we might run into are. We need to learn more about what these issues might be.
- First combined Islandora 8 / Fedora camp next week at ASU (Feb 24 - 26).
- David Wilcox mentioned applying for grant from IMLS. Currently working on a full proposal that would be a follow on grant to the migration grant from last year. Looking for institutions to work with (as a pilot). Contact David if interested. Goal is to proof out fedora 6 and build a bunch of documentation / training to create more of a path for migration for the community at large. https://drive.google.com/file/d/1vbgn_OWChV1u6oYdzQ3-yT4Chwv0nmgc/view?usp=sharing
- Fedora Demo
- Auto-versioning is working - it's a switch you can run with. `-Dfcrepo.autoversioning.enabled=true`
- Should take about 5 minutes and should have enough content
- Backwards compatibility document
- Looked at Option 5
- ACLs - will these work the same? ACLs go by containment so that shouldn't change.
- In past versions you could interact with intermediate nodes, so this is a change. We should capture some of these differences from past behaviors.
- With #5 affords some flexibility in how we deal with these nodes. If at some point a decision is made that ghost nodes are untenable and we need pair trees, we could create a way to "fill in the blanks". These ghost nodes exist implicitly and we could create them in the future if need be. This minimally commits us to behaviors that we could change down the road.
- We should bring this to community to get feedback (after Danny Bernstein gets back).
- Looked at Option 5
- Road to Fedora 6
- Sprint plans - any planning being done? Not yet.
- Demos can be contextualized with the Fedora 6 Feature Tracking.
- Would be helpful to keep this Feature Tracking list up to date as things get worked on.
- Do some of these have contingencies - ie, w/o auto versioning turned on, you can create binaries, but they are not an actual OCFL object because they are in mutable head.
- Some bullets need to be fleshed out more - ie, Documentation.
- Need to get the integration tests turned back on as features come online again.
- Priority tasks (Development Roadmap : 2020 Sprint Roadmap)
- As you work on things, keep an eye on this ticket and re-enable tests as you go.
- Other priority issues?
- Containment Index. - in review. Builds for Jared Whiklo but not working for travis.
- Lots of tickets in Review... if you have some time, looking at these would be helpful.
- Goal of having Fedora 6 beta release done this calendar year.
- A doodle poll will be created for organizing sprints.
Actions
- Clarify in in documentation that multiple simultaneous writes to OCFL are not supported
- After team has a chance to comment, send Ghost Node idea to general community for feedback
