It’s here! It’s finally here! While we were all out shopping or eating our Thanksgiving leftovers, LOC published the long awaited data model for bibliographic linked data on Black Friday, November 23, 2012! In a document entitled, “Bibliographic Framework as a Web of Data: Linked Data Model and Supporting Services,” the contracted team of Zepheira outlined how we might move from our static MARC record silos to open linked data networks. I read through it today and here are my thoughts:
AACR2, RDA, DACS, CCO, MARC, MODS, MADS, METS, DC, PBCore, VRACore, EAD, CDWA, CDWA-lite, LCSH, LCNAF, MESH, AAT, LCC, DDC, FRBR, FRAD…and list continues…welcome BIBFRAME! Another acronym you’ll use to bore your friends and confuse non-cataloging coworkers! BIBFRAME is short for Bibliographic Framework and is the name of the new model that will take our bibliographic descriptions into the 21st century. Here are a few of my takeaways:
“The goal of this initial draft is to provide a pattern for modeling both future resources and bibliographic assets traditionally encoded in MARC21.” (p. 6)
While reading this document, I had to remind myself that this is a draft of a conceptual model. LOC acknowledges this and even wants feedback from the library community. This is a work in progress that hopes to establish a data model and develop an encoding standard for expressing bibliographic metadata based on MARC21 records. Thank goodness!
Goodbye MARC…Hello RDF!
BIBFRAME moves from flat bibliographic descriptions that use controlled vocabularies, to dynamic linked RDF triples that bring together the inherent relationships of our resources. The model looks like this:
(My apologies for the poor quality of the snagged image from the report – I’m trying to publish this before the end of the work day) From this model we can see the relationships between Works to their creators and subjects; Works to Instances; and Instances to publishers, locations, and formats. Creators, subjects, publishers, locations, and formats are all considered Authorities.
BIBFRAME then uses Annotations that relate to either Works or Instances to include local holdings data or other linked data services such as reviews, book covers, etc.
but what about FRBR and RDA you say?! Not much mention of them in the draft, except for a broad overview at the end. Instead BIBFRAME re-conceptualizes the FRBR WEMI (Work, Expression, Manifestation, Item) model into a less hierarchical structure. It uses network graph relationships between Works, Instances, Authorities, and Annotations.
I have mixed feeling about this new WIAA model. Just when I was starting to see how FRBR’s WEMI model could be expressed through RDA, BIBFRAME throws curve ball. While I was reading the draft, I kept thinking where are the Expressions? Just Works and Instances? Then I have to remind myself – it’s all just semantics with relational data. Here’s how I broke it down:
- Work = Works
- Instances = Expressions + Manifestations
- Authorities = Authorities
- Annotations = Holdings + Other linked data stuff
“Formally reconciling the BIBFRAME modelling effort with an RDA-Lite set of cataloging rules is a logical next step.” (p. 15)
Really?? An RDA-Lite?! We’ve just spent 3 years debating whether or not to adopt RDA in the first place and we finally have the green light from central command. Implementation is March 31, 2013. Period. Now they want a “lite” version to express in linked data relationships, rather than hierarchical descriptions!
Content Standards are guidelines that have to account for all circumstances of resource description. If you want a “lite” version, use what you need from the standard and ignore the rest! There’s not need to develop a new version of the standard to accommodate a new model.
A Centralized Approach
The BIBFRAME model calls for a centralized namespace for all Works, Instances, Authorities, and Annotations. I’m
very confused by this approach. How I am supposed to adopt a new model and standard of Bibliographic description at my local institution if it’s all being managed by LOC? Does LOC really care that my institution has three copies of the Games of Thrones? Or will they only be implementing this for their collections and descriptions. (see comments below) I want to research how MARC was implemented now.
So LOC will maintain the framework centrally, like it does with the MODS family and other structure standards. But why maintain a new separate model for authority data that links to existing linked data service? The report gives the following example:
There must be a reason for creating a new authority record that references an existing authority record, but it just seems redundant. If I link my BIBFRAME Work and Instance record to BIBFRAME authority records, the BIBFRAME authority record continues to link out to another linked data service. It feels like an unnecessary mediating link.
More to Come
Ok – that enough for now. Please watch out for more over the next couple days as I process this document further and continue my research. I promise to post more often too. I’ll be reviewing RIMMF soon!