Attendees

• Andrew Woods
• Esme Cowles  
• Longshou Situ
• Michael Durbin 
• frank asseg
• Kevin S. Clarke 

General

• Indicates who took minutes -
• Call-in: Google-hangout at:
  
  • https://plus.google.com/hangouts/_/event/c1glu6soq43r1rr6ou17qtobug8

Goals

1. Finalize versioning
2. Triplestore/Solr
   1. Rebuild
   2. Support multiple repositories
3. Clustering potential: HA and Scalability
4. Performance: hierarchy
5. Resource locking

Minutes

• High level: Honing in and focusing on feature completeness for Fedora 4.0 Feature Set
• Loose ends: versioning, triple store, and performance
• Introduction for new sprint developers:
  • Bug tracking done in Pivotal (documentation for it is on the wiki at Story Management)
    • Tickets for the current sprint are listed in "Current" (their order indicates priority)
    • You're the "Owner" for your tickets; click "Start" when you start a ticket and "Finish" when you finish
    • Put a link to a pull request into the ticket, and wiki documentation if applicable, before you click "Finish"
    • Finished tickets that need more work go temporarily into "Rejected" state
    • Developers can click "Finish" again after additional work has been done
    • Tickets for features are assigned points to indicate level of effort
      • 3 points equals an ideal workday
      • 1 point equals a trivial amount of work
      • 5 points equals a multi-day task
      • 8 points means a ticket needs to be broken into smaller tasks
  • GitHub
    • You can create a branch in the fcrepo4 repo, but it's probably best to fork and create branches in your fork for the work you do
    • If, after submitting a pull request (PR), additional work in the branch is done, the PR is automatically updated with the new work
    • Generally, all commits on a branch are squashed before merged to master; talk to Andrew if there is value in keeping multiple commits
  • Unit tests and displayed in Sonar http://sonar.fcrepo.org/dashboard/index/org.fcrepo:fcrepo
    • New code should contain unit tests (examples are in the code base)
    • Would be nice to have integration tests as well, if applicable (examples are in the code base)
  • IRC channel is used for ongoing/daily project communication (#duraspace-ff on irc.freenode.net (IRC logs))
• Ticket Highlights
  • Index rebuilding (esc) https://www.pivotaltracker.com/story/show/65168574 (need to be able to rebuild indexes both for Solr and indexes like Sesame, etc.)
  • Rollback versions (md) https://www.pivotaltracker.com/story/show/60883346 (allow the REST API to support version related activities)
  • Node level and graph level locking (md) https://www.pivotaltracker.com/story/show/66093788 (picking up earlier investigations into this)
  • Getting a sense of underlying system and development process (ksclarke and lsitu); there are various smaller tickets for accomplishing this
  • Performance testing for various backends; check Mode's support for large numbers of child nodes (ksclarke) https://www.pivotaltracker.com/story/show/68662964
  • Underlying sparql functionality for multiple Fedoras/consumers feeding a single triplestore (lsitu) https://www.pivotaltracker.com/story/show/68853354