Print

Print
I'm part of the BIBFRAME testbed group at Princeton, and the question of how to handle bf:Providers was one that came up last week while we doing data-entry exercises with the BF Editor.

The editor's default Monograph profile has a "Publisher" field that lets users create a "Provider" Entity, complete with lookups and associated URIs for bf:providerName and bf:providerPlace. However, because the current lookups only query http://id.loc.gov, it wasn't possible to do anything with commercial publishers, since these don't get established in the Name Authority File.

Compared to the Provider Entity and its lookups, the free-text "Provider statement" field did seem like an odd appendage. Yet this is the field that RDA instructs us to choose, and our RDA-based cataloging only supports names established in the NAF--and since commercial publishers don't get established there anyway, we are stuck with not being able to make our data less "stringy" (the word one participant used to describe RDA).

In this regard, I think it is important to bear in mind the distinction that Kevin Randall drew between description, on the one hand, and relationship, on the other.

To Kevin (Ford's) point:

On Thu, Jul 31, 2014 at 6:06 PM, Ford, Kevin <[log in to unmask]> wrote:
I'm not a fan of the repetition either, but RDA often requires /transcription/ with all the pitfalls that might entail (misspellings, non-standard abbreviations, non-standard spellings) and so I worry about the need to record those details "as they appear on the source of information" and the impact that would have trying to standardize on bf:Organizations and bf:Places.   Functionally, the transcription serves to help a user match that which he/she may hold with what is seen in the record, which is why standardizing abbreviations (Chicago, Ill. becomes Chicago, IL perhaps), for example, can be a problem, since it would no longer match what is on the manifestation.

I would also say that transcription serves to reflect the entity's own self-presentation. We can't establish "authorized" forms unless we have evidence from actual usage of how an entity presents itself. 

Commercial publishers have not traditionally been understood as playing a significant role in a work's intellectual creation, and so even though we have usage for their names, we don't make authority records for them. Nonprofits or government agencies are a different story, because they often are seen as making an intellectual contribution, and so they're names can be established in the NAF and traced from a 7xx field in MARC (with an appropriate relationship designator like "issuing body," etc.).

Kevin's "fantasy" of entity recognition on bf:publicationStatement would be ideal, and one hopes that future revisions of cataloging rules will accommodate this. However, I wouldn't necessarily see it as duplication of data. We need usage in order to establish an authorized form, and evidence of usage comes, in part, from transcription--and transcribed forms can often differ significantly from the form that is eventually established in an authority record. 

Of course, on one level, maybe we don't have to worry so much about the "preferred form" of a name when we are identifying things with URIs, but the preferred form does continues to be an important piece of information when, for example, we are doing BFE-style lookups against sources like id.loc.gov.

Would it be feasible or desirable, then, to create something like "transcription" classes in BIBFRAME that would provide an abstraction layer (similar to bf:Authority) and let us link from a transcribed form to its bf:Authority? I mentioned something along these lines in a previous post on statements of responsibility, where the issue is essentially the same.

All best,
Tim

--
Tim A. Thompson
Metadata Librarian (Spanish/Portuguese Specialty)
Princeton University Library