06.md

Islandora Fedora 4 Interest Group: Meeting 06

April 24, 2015 1:00-2:00 EST

Call details: Skype (add ruestn as a contact)

Attending? Fill out this form.

Agenda

1. Introductions - recorded in the Attendance
2. migration-utils
   • Portland Common Data Model (repo) - past month and a half has been focussed on this area, crossing over with Fedora sprints. Also doing modelling and mapping from Fedora 3 -> 4 to bring us into compliance with PCDM.
   • Hydra Metadata Working Group
     • Not just Hydra! It's just a name. Good work being done here, broken into sub groups. Nick with technical, descriptive , and rights metatdata groups. Work is wrapping up in the next two weeks, where it can be pulled into out project.
   • migration modeling
     • auditTrail mapping
     • Feedback from Fedora tech list here
     • Also looking to make sure we have all fedora 3 audit trails and that the mappings make sense.
     • RELS-EXT Islandora namespace
     • RELS-EXT Fedora namespace
     • Predicates: had a valid use case in Fedora 3. What do they do for Fedora 4? isMemberOf and isPageOf in a Book, for example. How do they differ? Other predicates here. Answer: we use these but we do not yet have a defined ontology. Nick is going to work on pulling together and publishing one in the near future, to go on islandora.ca. Right now they are just out there in the code and usage is based on how they have been used in the past. In Fedora 4 predicates like the examples may merge into "hasParent," simplifying matters. Further discussion on the details may wait for when we put together an ontology.
     • Visualization of the PCDM and its structure. Made with yEd. And here it is exploded.
     • There's an idea to take the DC datastream and apply it as properties to the object. There would be no explicit DC datastream on the object. OR - if the community wants to keep it, it could be an LDP/RDF source in plain xml on the PCDM object.
3. Documentation
   • Services
   • Danny's proposal for how services should work. Not an exhaustive list, because Islandora is meant to be a plugable framework. Instead, this defines the interface for how these plugable bits can be built. Everything is keyed by the Drupal ID, which can collide if you have a multisite (thanks for catching this, Paul Pound!), so it will be changed to UUID.
   • Currently creating CRUD operations for PCDM object types. Following conventions as much as possible. File services make things a little hazy. In Fedora 3 we'd have DSID. Services will be boxed by DSID.
   • Next task is to make this happen in Camel.
   • Derivative services comes next. Making this RESTful is taking some creativity, but it's coming along. Patch will be used to regenerate - Danny is open to suggestions if there's a better way.
   • Asynchronousity for the win.
   • It will be up to the implementation whether derivatives are stored in Fedora or not. This still needs some debate. Danny proposes that anything subject to regeneration should not be in Fedora. Maybe we just make it configurable.
   • Zip ingest is going to be a core function. Other techniques available for files that exceed zip limits.
   • Standardizing on a manifest file format and sticking with it would be great. Perhaps holey bags.
   • For Fedora 4: the default assumption should no longer be "everything on the same server." Horizontal scaling comes out of the box.
4. Audit Service
   • Audit Service Implementation Proposal
   • Using Audit Events Phase 1
   • Fedora Sprints: first phase is over. Testing went smoothly for initial implementation of audit service; standalone triplestore. One of our requirements for migration. We will use PREMIS migration vocabulary to store audit trails.
   • Timeline: next ticket to go in is implementation of audit trail mappings. A few weeks off.
5. Issues
   • Investigate the Xpath Field module
     • From Alex O'Neil, Islandora Developer of Yore. Our fork
     • Danny and Jared have created Drupal Content type for Basic Image that we are experimenting with.
     • Form Builder may have to carry over. We don't have the resources for this right now, but someone is going to have to take this on. This should be a community effort - if the community wants it, we have to come together to gather the resources for it to happen.
   • Investigate XML editors to replace XML Forms
   • Create Drupal Content Type for Basic Image
   • Performance testing triplestores for Islandora community recommendations
     • Mulgrara has burned us a little. We would like to recommend something robust in the future. Diego Pino has put together potential benchmarks. A lot of crossover with Fedora community in general, also DPLA.
     • The role of the triplestore in the stack: things are indexed in Drupal, so no triplestore to view things. Middleware/syncservices will require a triplestore to line things up
6. Install/DevOps
   • PHP Code Sniffer has been added
   • Fuseki 2.0.0 has also been added. Deployable war, so it can go in Tomcat.
7. Use cases
   • Check them out, add yours (there's a template!), bug Nick if you need help.

Attendees

• Nick Ruest (York University)
• Danny Lamb (discoverygarden)
• Melissa Anez (Islandora Foundation) ⭐
• Mark Jordan (SFU)
• Lingling Jiang (USTC)
• Gavin Morris (Common Media)
• Kelli Babcock (UofT)
• Jared Whiklo (University of Manitoba)
• Martin Dow (Acuity Unlimited)
• Mark Cooper (LYRASIS)
• Chuck Schoppet (USDA)

Notes

In-line.