2018-12-04 Meeting notes

ETD / Publications / Data Sets Review Updates - https://docs.google.com/spreadsheets/d/1GH5ulJJavsNjdsZcae5Y7t9EOb2_nyarJj9C11nl4Vk/edit?usp=sharing

• Should keyword be required field?

  • Possibly with local keyword authority, e.g., https://wiki.duraspace.org/display/ld4lLABS/Keywords+(Agrovoc)+-+Lookup+with+Context; free-text entry by submitters may have limited use / cause more problems than it helps

  • What is intended outcome of keywords in aiding indexing/retrieval? Would we want users to add keyword synonyms/alt names for words already present in document (e.g., doc refers to cardiac arrest and never mentions heart attack, user/submitter should put ‘heart attack’ in keyword field?)

• Question re: “Advisor” - comments seem to indicate discussion re: creating CSU authorities; where if at all would ORCID IDs come into play for this?

• Re: Controlled vocabs (majors, institutions, degrees) - does this IG need/want to take a stab at creating those (would that be helpful @Kevin Cloud (Unlicensed) ?)

Linked Data for Languages Recommendation

Working on the San Marcos instance, recently there was a question on whether we could have an authority for our language field. The use case is to have a drop down or set of requirements to check the field against a controlled vocabulary. The linked data representation in the application is structurally DC.11.language, however, we do not currently have a defined controlled vocabulary.

In order to do the technical work for this, I have two options -- either to check the field against a linked data authority such as http://id.loc.gov/ or to rely upon a local authority file. I'm happy to consider either (or both) but a bit of that requires some understanding of what standards are available/preferable for this particular field as well as what published linked data authorities are available to validate against. For instance, our recommendation for ETDs (see the screenshot attached, or https://calstate.atlassian.net/wiki/spaces/SCHOL/pages/79732530/ScholarWorks+Data+Models) in Hyrax/Samvera is to use a controlled vocabulary based on ISO 639-3. It doesn't appear that LOC publishes ISO 639-3 identities as they do for ISO 639-1 (http://id.loc.gov/vocabulary/iso639-1.html).

Even then, it may not be a superb UI decision to have a drop down list of all of the languages within the standard. For starters though I wanted to check in to see what insight you have on linked data for languages.

Seems like we should be using ISO 639-2 as those are the basis of MARC language codes and could facilitate crosswalking where/if needed (http://id.loc.gov/vocabulary/iso639-2.html). ISO 639-2 is very comprehensive so I’m not sure we would need a local authority for this except for cases where the language of the item is not known or invented/so obscure it’s not on the list of ISO 639-2 (could put zxx / no linguistic content / not applicable for these I suppose).

To handle length of drop-down; is it possible to pre-populate with list of top 5-10 languages (English, Spanish, French, German, No Linguistic Content/Unknown) and in link fields, link out to guidance for selecting appropriate ISO code? Maybe pre-populate English by default so would only need to be changed/updated if something was non-English? (do we have much in ScholarWorks not in English currently?)