Time/Place

This meeting is a hybrid teleconference and IRC chat. Anyone is welcome to join...here's the info:

Attendees 

Agenda

  1. Current F4 implementation status, and steps towards production release
  2. Application Messaging and existing header format
  3. Revisit ordered-list ontology discussion
    1. Design page: Design - Ordered Lists

Minutes

 

Current F4 implementation status, and steps towards production release

  • Need to wrap-up before Thanksgiving
  • Lots of dead code in the codebase that needs to be removed
  • Lots of work went into supporting LDP
  • Possibility of versioning the API separately from the application
  • Core API and extension APIs
  • Performance testing framework (Grinder)
  • Clean up docs
  • Authorization: attaching ACLs needs to be flexible rather than creating them as children of protected nodes
  • Need line-by-line code review and feature/capability review
  • Shooting for 75% test coverage (unit, integration, cumulative)
    • Chris: Many of our unit tests only prove that the code gets called; not that it does what it is supposed to do
    • Andrew: Some value in looking at our unit test process to see if we can improve (e.g. dependency injection)
    • Andrew: Integration tests show that everything works; unit tests are more useful when actually structuring and writing code
  • What else do we need?
    • Chris: we’re promising RDF support and need to make sure it works in the real world with real data
      • There is an integration test module to start this effort
      • Further tasks will emerge with real world data
      • We should add integration tests to support real-world needs
    • Esme: Message consumer needs to be reviewed
      • Most refactoring has occurred in the main codebase rather than the plug-in projects
    • Esme: We have tests that target objects but not datastreams in some cases
    • Stefano: How are the external modules aligned with the core release?
      • Andrew: External modules should be part of the release, but they are not blockers if it turns out there are any issues
    • Stefano: SPARQL queries fail on persisted properties

Fedora 3.8.0

  • There are some issues but it needs to be released
    • Are these issues blockers?

Application Messaging and existing header format

  • STOMP assumes header names and values are separated by colons
  • Maybe use org.fcrepo.jms?
  • Should be dereferencible - maybe using a mapping?
  • Aaron has a pull request; will also patch message consumer

Ordered lists ontology

  • Could be accomplished using a CND file
  • Mike: As a second step, we could write some code that interprets the ordered list ontology and determines how to persist the ordered structure in the repository
  • Performance can degrade with many ordered nodes at the same level in the hierarchy