Descriptive Metadata Call 2015-09-30

Time: 1pm EDT / 10am PDT

Call-In Info: Google Hangout: 

Moderator: Carolyn Hansen (U of Cincinnati)

Notetaker: Julie Hardesty (Indiana University)



  1. Metadata Survey
    1. Update from HydraConnect
      i.  Slides from Metadata WG:
    2. Publication of results?
  2. Best Practices for Descriptive Metadata
    1. Discussion of proposed base/if applicable elements (see list below)
  3. MODS To RDF update

Other items?

Survey Results:

Element Spreadsheet:

Proposed Best Practice Elements



Title / Label




Content Type ()

Subject (at least 1)


If applicable:





  1. Metadata Survey
    1. Not much time for response/feedback in session, but there was indication of interest in accessing results
    2. Can just publish results on wiki or present compiled stats
    3. No objections to just publishing results on wiki
  2. Best Practices for Descriptive Metadata
    1. Discussion of proposed base/if applicable elements
      1. List comes from session at HydraConnect
      2. Notes from session are not posted yet
      3. Need recommendation that could map to DPLA and won’t interfere with Hydra-in-a-box
      4. Feedback from group on listed elements
        1. Steven - title elements from MODS have to be represented differently than what most folks are doing now
        2. Structural metadata might get mixed with descriptive when talking about data sets
          1. think this list should just be focused on descriptive metadata only
          2. Julie - structural metadata subgroup is focusing only on how to express segments within a file, no plans at this point for larger structural metadata work
        3. Resource type, format, description are all missing
          1. Difference between format and content type? content type is closer to genre (like dc:type vs dc:format)
        4. Best practices needed for vocabularies to use as well?
        5. Survey showed lots of local fields as well as vocabs being used so need to make this as easy as possible
        6. Is it helpful to have smaller group look at this over next 2 weeks?
        7. DLF Aquifer guidelines or other already-established guidelines for descriptive metadata might help guide recommended elements -
        8. DPLAFest session about metadata practice 
          1. Metadata completeness score from Europeana -
          2. Date and place elements are incredibly important to DPLA and Europeana
          3. Metadata Quality Research Brainstorming: particularly section on Research, readings and existing work:
          4. Mark Phillips blog post -
        9. Recommendation from our group could be that it’s not time yet to make a recommendation
        10. MODS/RDF subgroup should maybe drive this conversation? Steven doesn’t think that’s necessary, Julie notes that the group is not looking at all MODS elements so maybe there is a "core" set of elements there
        11. Take this to larger Metadata WG? crickets
        12. Start with this list as a draft and see how it goes? not going to be one solution that will work for everyone
        13. Work on this for the next 2-4 weeks
          1. Review resources shared in meeting - Everyone
          2. Create spreadsheet of proposed elements, effects on developer/end user/aggregators/other audiences
          3. Start spreadsheet and share on wiki - Carolyn
  3. MODS To RDF update
    1. Draft complete for MODS title into RDF - need feedback from Descriptive Metadata group -
      1. (Note: "Complex" tab shows that titles are their own objects in Fedora, under Custom Minted Resolutions column; there is also a "Simple" tab with simpler recommendations)
      2. Modeling this level of complexity (the Complex tab of Draft) renders it unshareable; no reason to use RDF if this is what it looks like
      3. It is more complicated to use this as a SPARQL endpoint but it’s valid RDF so it is still parse-able
      4. Reason to use more complex recommendation - don’t want to end up with shell of former MODS XML record
      5. Example: Modeling a person as a separate object (30,000 objects in one place), other place that also has objects by that person can use this object as authority record
      6. Object model similar to what UCSD has now and they modeled that way because they didn't want to lose any fidelity
      7. Goal of MODS/RDF group is that there will be transformation provided for MODS XML to RDF and also from RDF back to MODS XML
      8. UCSD has revised data model -
        1. Talking about moving away from LCSH and towards FAST to simplify away from complex subjects
      9. Maybe BIBFRAME is better place to look for way to express this info?
      10. There is response to survey about a place using BIBFRAME in Fedora (not sure if it’s 3 or 4 though) so there might be helpful info available regarding BIBFRAME in Fedora
      11. BIBFRAME variant title example - - might be incorporating blank nodes but might also be possible to get around it
      12. Carolyn is co-chair for a BIBFRAME group within ALA so would like to see this in Fedora (no pressure, though)
    2. Working on MODS name element now, to be done in next month; 
      1. Issue: list of author triples, how to order in Fedora 4? 
        1. UCSD not doing author ordering, can’t automatically generate preferred citation in this case and have to make those citations as text, hoping affiliation info will be stored in author profile systems but not there yet
        2. relator:auth with order elements provided as possible example
      2. Issue: Name with an affiliation that now has different affiliation from when object was created? 
        1. Affiliation doesn’t really have relation to object, but it’s an affiliation with date ranges associated with people
        2. Problem with that if object doesn’t have a date and start/end dates of affiliations are not being stored (which is often the case)
      3. Working on solution to this at next MODS/RDF group meeting
  4. Other items?
    1. No