The current issue of the Code4Lib journal contains an article by Jeremy Nelson of Colorado College, “Building a Library App Portfolio with Redis and Django”, that highlights the development of FRBR datastores that run on a NoSQL database server (Redis). More interestingly, perhaps, this platform is based on the BIBFRAME model with four core classes: creative work, instance, annotation and authority (for further details, see the project site on GitHub). To me, such a two-level mapping makes a lot of sense. In fact, I quite like the reduction of the FRBR complexity in the BIBFRAME model, especially with the anticipated re-use by other communities in mind. Jeremy Nelson explains on the BIBFRAME mailing list:
Because of Redis’s flexibility, I’ve been able to use RDA element names as either discrete properties for each BIBFRAME entity or as part of the naming scheme for the BIBFRAME entity’s associated keys. A nice feature of this approach is that we are not restricted to just RDA but we can use other metadata standards (MODS, DC, ONIX, VRACore, etc.) as discrete properties or as part of the Redis key naming schema for the BIBFRAME entities. We are also using a simplistic mapping of FRBR Work and Expressions to BIBFRAME Creative Work, along with FRBR Manifestation and Item to the BIBFRAME Instance…
It was with great interest that I read the paper (PDF) “FRBR and Facets Provide Flexible, Work-Centric Access to Items in Library Collections” (2011) by Kelley McGrath, Bill Kules and Chris Fitzpatrick (mentioned on NGC4Lib) because it modified and enriched my understanding of the relationship between facets and FRBR and the way facets help meet the users’ information needs. Sure, facets are there to help users refine their search and pull out a smaller set of results that match certain attributes, but what is the theoretical underpinning and how does the FRBR model relate to facets?
The paper cited above highlights the authors’ experience with modeling and building a search interface including facets for a moving image collection, and while some of their observations are specific to these resource types and the retrieval requirements that go with them, much is generally applicable. The main point for me being (as alluded to in the paper’s title) that facets are much more flexible than hierarchical FRBR structures through which the user would have to navigate – facets allow the user to combine any number of attributes when limiting the results, without clicking through hierarchies of work, expression etc.
What makes the model and the prototype interface so powerful is the fact that FRBR is not slavishly followed but rather adapted to the specific features of the resources, collapsing the work, expression and manifestation entities into two levels, “movie” and “version/publication”. This helps avoid duplication of information, both regarding display and cataloging, and answer the questions: “what do you want?” and “how/where do you want it?” (probably the most general questions user bring to the catalog).
Through facets, users are offered several pathways into collections: “Patrons can start their search at any point in the FRBR hierarchy, from Item (location) to Work (genre, date), and easily transition between search and browse strategies, using facets to broaden or narrow their results and pivoting on facet values.” (p. 4) – explorations they cannot as easily undertake in a tree-like FRBR representation.
In one of his recent posts, James Weinheimer stated that the possibilities to limit and sort results by facets provided by discovery layers “fulfill the “FRBR user tasks” right now, and even overfulfill[s] them.” I hadn’t yet seen facets as embodiments of FRBR (or at least I hadn’t spelled it out so clearly). But it turns out you can ascribe attributes associated with each FRBR group or entity to one (or more) facets. So I figured it’s worth visualizing this with a concrete example for myself as well as for others (to who this may have been obvious already …). Let’s not argue about whether form/genre belongs to manifestion or expression, all that matters is whether whatever tool or construct we offer will help users find (I take “find” in the broadest sense of the word, i. e. to include identify, obtain etc.) what they are looking for, not so much how we librarians conceptualize or name it.
Since the example I chose is from the Austrian union catalog (ETA: http://search.obvsg.at/OBV), which serves a consortium of academic and administrative libraries, facets pointing to specific libraries or locations (to items in FRBR speak) are included. However, grouping is not done very elegantly – more effective grouping would sort the number of hits in a clearer way. A rather explicit way of grouping together expresssions and list all manifestations under the respective expressions is shown in slides 27 and 28 of Thomas Brenndorfer’s presentation “The FRBR-RDA Puzzle: Putting the Pieces Together” (unfortunately, the catalog he refers to is no longer available at that address, so I wasn’t able to replicate his screen shots). This way of hierarchical grouping makes it easier for the user to see whether a library has, say, different language editions, talking book or film versions of a given work.
Just a quick note to direct your attention to AustLit, a project I ran across in a post by Jim Weinheimer. You have to be a subscriber to access the full database but some examples can be explored here.
I was curious to know what the underlying structure was and what strategy of collocation they use for the FRBR views and found these two pages:
Data models and technology.
It turns out that AustLit draws on the topic maps “topic” and “association” constructs to model the relationships in FRBR. It’s interesting to actually see this in action.
While I am skeptical about the representation of ontologies and human semantics in computer systems, topic maps concepts seem to lend themselves very well to expressing bibliographic relations and collating bibliographic information.
The FRBR display is a choice librarians make in the hope of helping users navigate the bibliographic and resource universe – one example for more “elegant” ways of organization.
ELAG 2010 featured a “Workshop on FRBR and Identifiers”. The presentation gives an overview of which identifiers exist for various forms of resources, with special emphasis on FRBR entities, and including a brief look at the role of identifiers in linked data. Just for completeness’ sake, I won’t talk about URL identifiers for FRBR entities and relationships here – a vast topic in and of itself.
Library-created control numbers identify the metadata about the resource, not the resource itself (like ISBNs). For one resource different institutions (publishers, booksellers, libraries) create different identifiers – but how reliable and consistent are they? One ISBN doesn’t necessarily stand for one book only, undermining uniqueness in many cases. As WorldCat data shows (assuming that catalogers correctly recorded the details available), we have a large number of books without ISBNs (which only came into widespread use in the 1970s). Generally there is a considerable percentage of resources which are not identified in a standard way. So the picture is not uniform at all, and some of the established identifiers will have to be reconsidered: the ISBN system is likely going to reach its limits with the proliferation of e-books, and maybe the library world will sometime stop thinking in terms of “records” (possibly with metadata being assembled just in time instead of just in case) – will the LCCN be obsolete then?
There are many efforts of creating and maintaining identifiers in different domains. Libraries around the world maintain separate authority files (albeit tied together in VIAF) and create separate “records” and thus identifiers for the same resource. It’s important for identifiers to be reused outside their specific areas. Library identifiers have lingered in silos for a long time and are only slowly being adopted by “outside” communities (e.g. German Wikipedia linked identifiers from the National Library’s name authority file with the articles about the respective persons).
A given FRBR work usually has various manifestations which in turn have several identifiers (leaving out the expression level for the moment) – those are the most commonly used (ISBN, LCCN…). OpenLibrary, for one, collocates manifestation identifiers. Topic maps could integrate information from heterogeneous sources on the basis of identifiers. We can probably never achieve global agreement on one unique bibliographic identifier, nor do we have to if we have systems that enable us to consolidate the diversity of identifiers.