Advice on handling multiple copies of an item + creator roles

We’re a non-profit association looking to provide an open digital catalogue of books, media and some other objects that currently exist in the personal libraries of various experts in our field. Several of these experts have expressed an interest in bequeathing their libraries to the association at a future date.

While Omeka S is conceived for digital collections, I think it could work well in this context and is not quite as challenging for the cataloguing end users as Koha for example. Some manual cataloguing has already taken place a simple form in doc files and I have extracted data from that in a structured form for importing via CSV.

I’ve read around the docs and forum but still have a couple of questions…

  1. Sometimes several physical copies of the same item exist in different personal collections. Each copy has been assigned an own signature-id but otherwise the item details are identical. I’m unsure what the best approach is here in the context of Omeka:
  • 1 work (book/object/…) = 1 Omeka item – i.e. store a single item with a property that could hold multiple signature identifiers (or repeat multiple times). A public visitor would get a single search match with multiple signatures in the detail view.
    Here, I’m not sure which dc:property (or other vocabulary property) might be well-suited for recording multiple signature identifiers (text-string)? For example, the Library of Congress’ modsrdf vocabulary has Location with sub-property “LocationCopy”. Is there something more suitable? Presumably, one could use the multivalue separator of the CSV Import module to split a CSV column holding one or more signatures into multiple instances of a property (analog to dc:creator).
  • 1 physical item = 1 Omeka item – i.e. create multiple items for the same work but with individual signature identifiers, and then somehow interlink these. A search would then return multiple identical entries but with different signatures. Presumably the cataloguing users could use the ItemCopy Module to duplicate items simply. How could I automatically interlink the parallel items in Omeka? And do I have a chance of making the connection between items already clear as part of CSV import?
  1. Many of the items already recorded have multiple authors. It should be possible to split these out into multiple dc:creator entries using the multivalue separator of the CSV Import module. In several cases, one of the authors is the book “editor” in the sense of a person who collates a volume by one or more creators (as per the LOC edc relator). How could I additionally designate this role for one or more creators in the dc:creator instances. I’ve seen the Relators but the associated forum thread suggested it’s not compatible with the CSV Import module.

Many thanks for any assistance!

Items in Omeka S (and item sets, and media) can link to each other to show relationships. Any Dublin Core or other metadata field can be filled out on one item with another resource.

For your case, you could use the Dublin Core field “Is Version Of” to link two items that are near-identical (other than the identifier, the item set / collection, and maybe some condition or provenance notes, I’m guessing). There may be other vocabularies, such as Bibliographic Ontology, that provide a better field for this information.

These links by default work one direction only (X is a version of Y, but Y is not automatically a version of X). You will need to go to each item and define its link to another explicitly. But, you can do that with CSV Import, as long as item X can point to a unique field of item Y, to find the right item to link up with (your identifiers, surely, would do the trick).

So, I would suggest one physical item = 1 Omeka item.

Remember that you could make other Omeka “items” that function like other nodes of information, such as people who are your authors or publishers, or the publishing companies, or locations in the world. Items in Omeka S are very flexible that way, nonhierarchical. You can choose one kind of division for item sets (say, the personal collections are separate item sets) and then define other commonalities using “items”.

As for question 2, you may find Dublin Core limiting for your purposes in book description. A lot of people find dc:contributor and dc:creator quite limiting. We offer value annotations that allow you to specify, say, the kind of contribution each dc:contributor has, but this isn’t very clear data structuring. And CSV Import isn’t currently set up to bulk-import value annotations, though I believe it’s on the to-do list.

I would suggest looking at other vocabularies online and loading them into your installation to pick and choose a custom resource template. Follow the instructions on the Vocabularies page linked above to add other vocabularies. You could easily add BIBFRAME, which I think comes with the relator roles you’re looking for:

If you’re planning on loading everything in using CSV Import, that’s one thing; if you are going to have staff doing manual item-by-item cataloguing later, you may want to add Value Suggest or other shortcuts to choosing the right values without entering in typos:

Using Omeka for books shouldn’t be too hard; I’d be interested to see what you come up with. The pieces you assemble into your data model might make a great case study for other users.