Time/Place

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

Attendees 

Agenda

  1. Star master
  2. Announcements
    1. Holiday Party follow up meeting :  Monday January 8th at 3PM Eastern 
      1. https://docs.google.com/document/d/10TpyqAVgy110Zph_aYAH-bJIfNZr4Vj1e3-vhTfUv4U/edit#
      2. Shall we put it out there.
  3. 4.7.5 release - Planning for week of January,15th 2018
    1. Shall we push it out a week?  
      1. Announcement needs to be sent
    2. Release manager - Danny Bernstein
    3. Volunteers
      1. Testers
        1. Danny Bernstein
        2. Joshua Westgard
        3. Peter Eichman
        4. ?
      2. Someone to review 4.7.5 commit message for signs of missing documentation?
      3. Preparers of Module Release Candidates

        ModuleWho?
        fcrepo-module-auth-rbacl
        fcrepo-module-auth-xacml
        fcrepo-module-auth-webac
        fcrepo-mint
        fcrepo-audit
        fcrepo-webapp-plus
        fcrepo4-vagrant


    4. Resources:
      1. component release process tracker: https://docs.google.com/spreadsheets/d/1I_zTMxh2l2rf2wpafoTwhSTR5GZuEoaTcZmTKCI3xT4/edit#gid=1769378986
      2. Release Testing - 4.7.5
  4. Fedora API Test Suite... needing:
    1. Try the tool against an API implementation
    2. Code reviewing the tool... lots of low-hanging fruit
  5. Simple, synchronous query in Fedora
    1. What will it take to make this happen?
    2. Prior art
    3. Queries to support
      1. select ?s where {?s ?p ?o}
      2. select ?s where {?s <some-pred> ?o}
      3. select ?s where {?s <some-pred> <some-object>}
    4. Michael Durbin:  do we have a link to the document tracking  known limitations of modeshape implementation and requirements?
  6. Tickets requiring attention
    1. - Bethany Seeger
      1. Any more discussion needed here?  If a mimetype goes in, it should come out at the very least. 
      2. Bethany Seeger suggested an alternative implementation: https://github.com/fcrepo4/fcrepo4/pull/1272#issuecomment-353173508 that would move the check closer to the HTTP layer
      3. Ralf Claussnitzer: Sounds reasonable. HTTP servers should respond with BAD REQUEST if given an unparsable mime type string.  Should the repository layer check again?
    2. - Ralf Claussnitzer to explore?
    3. - Ben Pennell to explore?

  7. 5.0.0 release
    1. API Alignment
    2. Pairtrees?
      1. Esmé Cowles : status of proposal and/or any feedback from community? 
  8. Beyond 5.0.0 - Areas of improvement
    1. Persistence?
    2. Journaling?
    3. Simple, synchronous query?
  9. ...
  10. Tickets In-Review


Ticket Summaries

  1. Please squash a bug!


  2. Tickets resolved this week:


  3. Tickets created this week:


Minutes

  1. There will be a Fedora Committers meeting next week
  2. Fedora 4.7.5 release
    1. Osman will not be able to be the release manager, so Danny Bernstein will be the release manager
    2. The release will be delayed a week or two
  3. Fedora API test suite
    1. Outside developer working on a test compatibility kit for verifying whether an implementation implements the Fedora API spec
    2. Lots of opportunities for code review and helping
  4. Synchronous Query
    1. Document for tracking known limitations of Modeshape implementation for query
    2. There was a query service at one point, but it was removed because of complications of LDP container implementation with dynamic triple generation, and lack of testing/verification
  5. Ticket review
    1. 2520: have reached general agreement here
  6. API alignment and Fedora 5.0
    1. Danny will work on Memento generation next week
    2. AppleTrees: not a ton of feedback, but generally not in favor
      1. Randall: Don't want to discourage progress, and 5.0 seems like a good place to integrate new breaking features
        1. But we don't see a lot of benefit from shortening URIs, and would rather avoid data migration issues
      2. Peter: Andrew suggested we could make the default minter config flatter, but document configuration options for PairTrees or AppleTrees, etc. for performance
      3. Aaron: PairTrees partition the repository with containment relationships that are different from most other containment relationships, longer URIs aren't the main issue
      4. Peter: we used the partitioning to make it easier to list all objects in the repository node-at-a-time when getting timeouts trying to list all at once
        1. Paging or other approaches might work
    3. We should assess the impact of making changes, and how disruptive they might be
    4. Esme: With the versioning implementation changes coming in 5.0, we may be able to revisit some of the identifer-translation changes that Benjamin Armintor and Unknown User (acoburn) explored, and we may be able to address the underlying performance issues that lead to PairTrees and other approaches.