Descriptive Metadata Call 2015-05-13

Time: 1pm EST / 10am PST

Call-In Info: Google Hangout: 

Moderator: Carolyn Hansen (U of Cincinnati)

Notetaker: Jenn Colt (Cornell U)



  1. Subgroup review
    1. introductions
      i. Chrissy @ UC Santa Barbara - interested in standards, areas of trouble with mapping from other schemas to f4
      ii. Corey @ NYU - interested in different data models in PCDM and f4
      iii. Danny @ Boston Public - not in f4 yet, here to figure out best practices
      iv. Eric @ Yale - interested in f4, linked data, RDF, standards
      v. Julie @ IU - Avalon team member, interested in PCDM and Fedora 4 for upcoming issues, deal with MODS and how that will work
      vi. Nick @ York University - part Islandora teams, how to handle f4 descriptive metadata, mostly MODS, mapping
      vii. Jenn @ Cornell - hydra, blacklight
    2. meeting schedule
      i. current time is good; will continue meeting bi-weekly
  2. Review subgroup goals
    1. Collect user stories (goal of having something to present at Hydra Connect)
      i. Want to collect from hydra users; What are people actually doing, what are the roadblocks, how to get people working together
      ii. Workflow? Collect user stories from our home organizations. Carolyn will compile list of organizations using hydra to divide up work of collecting user stories from those organizations
      iii. Collect user stories from both end users and administrators 
      iv. What granularity of user stories? Some are very specific and small, but not sure if they are wanted. We’ll figure it out as we go along.
      v. Start with learning more about what people are doing before jumping to recommendations
      vi. Perhaps a survey asking if people use RDF, XML, etc, what is it based on MODS, DC etc so we can see where people are at present
      vii. There is likely a hierarchical xml/mods contingent and a flatter RDF/triples group, have to balance both those interests
    2. Best practices on handling blank nodes and nested attributes (see also Fedora-tech discussions)
      i. Workflow will be creating from user stories 
    3. Best practices on using multiple schemas/vocabularies
      i. Workflow will be creating from user stories 
    4. Base descriptive metadata application profile with mappings to Shared Data Model and community standards such as Dublin Core and/or DPLA Metadata Application Profile
      i. Workflow will be creating from user stories 
    5. Sufia sprint
      i. What to do with descriptive metadata? Something has to show in the user interface
      ii. Dealing with only required fields in Sufia right now, 4 fields: creator, title, rights, keywords. All dublin core.
      iii. Where desc metadata shows up is still in progress
      iv. Right now it’s at the work level not the file level, that’s as much as they are doing for this sprint
      v. Two questions from the sprint will be posted to the wiki for discussion
      vi. Need to get more details for future Sufia work, set for this sprint
  3. Next Steps 
    a. Carolyn will check with Karen about scope questions, application vs users; will send out Hydra institution list and Sufia questions

Questions from Sufia Sprint

  • What descriptive fields should be at work level and what should be at file level?
  • Should work and file have different list of resource types? Files could be things like image or poster, but genre-based (conference proceeding, journal article) might be better at work level.

Notes from Sufia Sprint meeting on descriptive metadata (2015-05-05):

  • For sprint, stick with Sufia required fields:
    • Title - require title at work level and allow file title to be editable, but file name will be default file title
    • Creator - creator required at file but not at work; if work has creator, then any subsequently added files could have that creator offered as suggestion
    • Keywords - keywords required on work only; files will only be uploaded as part of work for this sprint
    • Rights - rights are assigned at work level and files below use those same rights
    • Resource Type - just at file level but later there might be separate resource type at work level
  • Metadata Standard - keep with DC as Sufia does now, knowing that people will want to customize that later and Hydra Descriptive Metadata Subgroup might recommend a different model (based on DPLA or some other major profile)
  • If anything in a file is edited/moved, the work record is shown as step in workflow to complete file edits