Samvera Community Wiki
MODS and RDF Call 2016-05-02
Time: 9am PDT / Noon EDT
Call-In Info: 712-775-7035 (Access Code: 960009)
Homework Reminder:
Subject Individual Mappings: Subject Individual Institution Usage And RDF Conversion
Moderator: Steven Anderson (Boston Public Library)
Primary Notetaker: TBA (etherpad link: https://etherpad.wikimedia.org/p/RDF-MODS-20160502)
Attendees:
@sanderson (BPL)
@Eben English (BPL)
@sonoe (UNC-CH)
@Moira Downey (Deactivated) (UNC-CH)
@Melanie Wacker (Columbia University)
@Danny Pucci (BPL)
@Sara Rubinow (NYPL)
@ksgerrity (Amherst College)
@Rebecca Fraimow (WGBH)
Agenda:
Introductions
Conversion Code Update / Notes Testing
Curation Concerns code up and running at: http://mods2rdf.xyz/
However, while the uploader (http://mods2rdf.xyz/uploads) works for title locally, it fails on the server. Still debugging that but should be up by the end of the day to see some limited conversion functionality.
Steps for usage:
Create an account on the site
Go to http://mods2rdf.xyz/uploads and upload a MODS XML document.
To view the end result, you can do a normal search (but currently on uses fields CurationConcerns knows about). Otherwise, to view in fedora, the credentials are:
username: darthvader
password: tapdancing
Still working on testing performance of storing strings vs. minting objects
Email Steven with any questions
Concerns from recent Applied Linked Data discussion
Using Fedora to store minted objects
Some discussion of whether it might be "cleaner" or less complex to only store finished "records" in Fedora
Use an additional triplestore (Blazegraph, Marmotta, Fuseki, etc.) to manage URIs for minted names, subjects, and other objects
Could potentially improve performance?
We should keep an eye on community best practices with this.
Using a single predicate with both strings and URIs as objects
Example: Dublin Core Elements subject (dc:subject)
Lack of data homogeneity could cause problems for data consumers outside one's own institution
No one on the call doing much work with other people's data, SPARQL endpoints, etc.
We need more information on how big of a problem this is – sharing use cases not well understood
Eben will send out email to Code4Lib and Hydra-Tech lists asking for comments
Subject Element Mappings Continued
Columbia updated modeling for name subjects
class cul:Agent
foaf:name for name value; skos:exactMatch linking to existing vocabularies
Steven will attempt to create collaboration document for subjects based on what people have put forward so far
We will discuss this at the next meeting
Next element: <mods:classification>
Everyone should work on this for next time.
Next meeting: Monday, May 16th, 9:00 AM PST (12:00 PM EST)