Hydra Tech Call 2016-04-06

Owned by Bill Branan
Jun 16, 2017
3 min read

Time: 9:00am PDT / Noon EDT

Call-In Info: 1-641-715-3660, access code 651025

Moderator: justin

Notetaker: cam156

Attendees:

  • Justin Coyne (Stanford)
  • Carolyn Cole (Penn State)
  • Lakeisha Robinson (Yale)
  • Lynette Rayle (Cornell)
  • Steven Ng (Temple)
  • Adam Wead (PSU)
  • Esme Cowles (Princeton)
  • Mark Bussey (DCE)
  • Gordon Leacock (Michigan) 
  • Hector Correa (PSU)
  • Mike Giarlo (Stanford)
  • Trey Pendragon (Princeton) 
  • Andrew Myers (WGBH)
  • Peter Binkley (University of Alberta)

Agenda:

  1. Call for additional agenda items
  2. Reconciling Hydra and Fedora WebAC implementations
    1. See https://docs.google.com/document/d/1O76Vo1crUqt84cG_V8CGOlcbmiqp0CQKOrzXSYlUZHk/edit
    2. Link from LDCX session on this topic
    3. Is there anyone on the hydra side that will work to implement more like fedora on web acl?
      1. Justin thinks that is a good idea
      2. Should happen along with the PCDM data migration
      3. Should be straight forward to do, but there may be some performance concerns
        1. need to try it out before the whole community adopts
        2. Mark B: Volunteer a DCE project to try it out
        3. Should be a part of hydra-head 10 which includes the access controls
      4. We have no way of looking at fedora to know wich access controls apply, so we should do it the way fedora does it
      5. Individual access control for each item or access as a class, or access via containment section (aka a public section with an acl)
      6. We current can and do have an ACL for each object, but it is not necissary.  This was done for simplicity sake and provided parity for how access controls were done in fedora 3.
      7. Trey may take a few days in a couple weeks to look at making this happen
  3. Sufia 7.0 progress update
    1. 31 open tickets remaining
      1. larger than a few weeks ago
      2. Gone over tickets and made smaller  bite size tickets
      3. People actively working
        1. Stanford Hydra in a box
        2. PSU (part)
      4. Work is accelerating
      5. 7.1.0 are backlog items that we want to attach sooner rather than later
    2. Batch update is broken in sufia 7.0
    3. Browse Everything integration is also broken
    4. will be ready soon for an early beta release
    5. large quatity of rough edges to be fixed, but seems to be functionally working 
  4. Reminder about Michigan Developers Congress
    1. May 4th - 6th @ Univeristy of Michigan`
    2. Google form for registration
  5. Merging hydra-collections into curation_concerns
    1. As part of the work we are trying to find the correct granulairty so that we strike a balance for adoptability vs maintainability
    2. Are there many use cases for hydra-collection outside of sufia and curation concerns stack
    3. Bring concerns together and allow for easier core gem matainenece
    4. Chris Beer will be working on this this week.
    5. Will be keeping the repo around for older installations
    6. What is the plan for cleaning up hydra collections in the future?
      1. Push it into project hydra deprecated?
  6. Resolve the schema clash https://github.com/projecthydra-labs/hydra-works/pull/288
    1. predicates that are clashing
    2. language and creator clash with document schema
      1. Avoid duplication: possible options
        1. characterization on fileset and basic metadata only on works
        2. remove the overlapping predicates and assume they exists in basic metadata and build on top of basic metadata
    3. conversion to DC11 at different times created the clash
    4. Reccomendation is just to remove the overlap from schema
      1. Hector will update his PR
      2. look for file author and language and change them to creator and language
    5. How metadata gets applied to a class is a function of the application
      1. The problem might be that active tripples should error instead of warn when we have an issue
      2. used two terms with different names set to the same name
        1. same title and same predicate should be ignored
        2. same predicate inside two property names should error
      3. AlreadyThereStrategy  will  have a noop is the predicate is defined the same way as it is already there.  Predicate is the same and name is the same.
      4. schemas do not define indexing behaviors.  Do not put indexing into the schemas.

Next Call

  1. Date/Time: 2016-04-13 9:00am PDT / Noon EDT
  2. Moderator: Adam Wead (PSU)
  3. Notetaker: Hector Correa (PSU)