Talk:Examples/mylib

Dividing the Examples Library into multiple libraries?

While I understand the rationale for keeping all of the references within Examples Library, I worry that it will not be helpful to sites that want to implement schema.org structured data for (say) library systems.

For example, at the Work/Expression level, I believe there are very few library catalogues that display the Work or Expression on a page that is distinct from the Manifestation level. My experience with implementing Evergreen at Laurentian University has shown that our users want as much information (in particular, links to electronic resources or listings of available physical resources) to be surfaced as possible at the initial search results level.

Thus, my instinct would be to divide the higher-level Work/Expression level examples (A0 and A1) * into their own "Authoritative Bibliographic Resource" (one could mentally substitute Freebase, OpenLibrary, WorldCat, LibraryThing, IMDB, etc) libraries, and the Manifestation/Instance level examples into "Example Library". The Manifestation/Instance level examples would then link to the "Authoritative Bibliographic Resource" to express their Work/Expression relationships.

I think this would make more sense to the developers of the current generation of library catalogues, while at the same time providing a realistic example of linked data at the higher object level. Certainly it seems much more achievable to graft that structure onto the existing Evergreen catalogue ("all" that I have to do is find authoritative W/E resources to link to for each bibliographic record in the system) than to try to create additional layers of UI that our users don't want to have to wade through anyway.

Ah, and it would remove the need to express the "sameThingAs" relationship that is in the current examples. I suppose for those catalogues that do surface W/E level records, there's a value to expressing one or more "sameThingAs" relationships; but assuming that you can follow your nose, if you're linking to an authoritative resource, then that resource should maintain its own "sameThingAs" relationships in a much more robust fashion than individual libraries are likely to be able to achieve.

• (kcoyle) I would add A6 and A9 to this. I am also unclear as to the intention of "sameThingAs" -- so perhaps that does not belong in the model until clarified.

The "authority file" model

(posted by kcoyle)

When FRBR was first being promoted to the MARC community, Barbara Tillett came to a MARBI meeting and explained the FRBR entities as being approximately equivalent to authority records of various types, and suggested that this would be the design for implementing FRBR. Such a design would provide a number of different options for applications. The "authority records" could be used to populate bibliographic records during cataloging. Depending on the system design, either data or links or both could be stored with the instance records (which would most likely represent FRBR:Manifestations).

I agree that few library systems today have divided the bibliographic universe into the FRBR entities, and I believe it is as yet unclear what the utility of such a division at the user interface might be.

I am also concerned that while one could describe a FRBR:Work using the properties of schema.org's CreativeWork, there is nothing inherent in CreativeWork that would distinguish the FRBR:Work description from one for FRBR:Expression or FRBR:Manifestation, and . Thus I consider a separation of WEM to be outside of the current scope of schema.org. I admit that I am also skeptical that FRBR will ever be in scope in library systems in its current form -- I think it remains to be seen how or if that exact model will be used.