the continuing crusade to devalue cataloging: 6th suggestion

Posting to Radcat

On 4/18/2016 4:23 AM, James Agenbroad wrote:
6. On Bibframe

For several years there have been efforts to define a replacement for MARC (MAchine Readable Cataloging) with something called Bibframe (short for Bibliographic Framework Initiative). Having retired from LC over ten years ago my grasp of it is severely limited. If 1. Bibframe is concerned exclusively or mainly with improving connections between library cataloging data (now mainly in MARC) and non-library Webbased data users such as Google, AND 2. Bibframe will not make it harder for libraries to acquire or use MARC data needed for OPACs and other tasks, then Bibframe’s impact might be limited to competition for programming resources to implement Bibframe or more library oriented asks such as OPAC improvements. One might ask, “Should libraries or non-library organizations who want to use MARC data pay pay to gain easier access to MARC data?” To whom do the expected benefits accrue–libraries and their users or external Web-based organizations and individuals? I cannot say.

Thanks for sharing. I’ll respond to this concerning Bibframe.

I also am not sure who it is being made for. One thing though: Bibframe will not allow librarians to do anything with their own data that they cannot do right now. Let’s face it: we understand our own data and we can already do anything with it that we want with it–even turn it into Bibframe!

The main purpose of putting your data into RDF (which is what Bibframe is doing with our bibliographic data) is so that others can use your data in more useful ways, because otherwise outsiders cannot understand your data structures. For instance, a non-librarian should not be expected to know that 245$a is the title proper (or even what that technical term means) and how it is related to a 240 or 246 or 260. Putting your own data into RDF is supposed to make these things clearer to outsiders, so that they can use your data to build their own tools. This is how and why the World Bank puts up their information in RDF so that some person can make a tool using World Bank data with Google Maps that lets people see how incomes are structured in Africa (one example). It makes World Bank data more useful.

I personally don’t know how well this works out in reality for the vast majority of linked data sites. So, I don’t know how many webmasters who might want to use our data could understand the incredibly complex Bibframe terminologies and structures, even though they are in RDF. But, I’ll give everyone the benefit of a doubt: at least it’s a try.

It seems as if many people believe that Bibframe must be stored in RDF/XML or JSON or some other format, but it doesn’t have to. Within our current library catalogs, MARC data may be stored as MARC records but it is rarely used by the catalog. (By this, I mean the raw data of MARC with all of the numbers at the beginning, e.g. seen at might be stored in the database somewhere, but the database does not use it) When a record is transferred into a catalog, it deconstructs the MARC record into all of the tables and cells of the relational database, and this is what is used by the catalog. Each database can store this information in completely different ways. Today the raw MARC records are used only in the few instants it takes to transfer records from one library catalog to another. As a result, a Millennium catalog has a different relational database structure from a Koha catalog, but they can transfer records from and to each other. After the transfer, each catalog converts the MARC records to its own internal structure.

Bibframe records can also be stored as they are, but they probably won’t because each catalog will still have its own needs, just as our catalogs do today, and the RDF will be generated in various ways as needed.

Also, you can be a linked data “consumer” without being a linked data “provider”. Bibframe is designed to make us into linked data providers and that means it will allow others to use our records. For free? you ask. Good question! I cannot believe that many people will be willing to pay libraries for their records, especially when they can get e.g. Amazon records and others for free–and if somebody buys a book they might even get paid out of it! With that option, why would anybody want to use our records? Especially in that incredibly complex Bibframe? But I am sure that somebody will want to use ours.

So, consuming linked data has been a possibility for a long time and there is no need for Bibframe if libraries want to build something that uses linked data. Still, how all of this is going to work, especially with the growing popularity of the library “discovery layers” which brings together all kinds of stuff into a single search and a single result: the proprietary journal indexes, separate local catalogs, open-access databases, ebooks and so on and so on, some even in full-text, with numbers of results that can quite literally overwhelm the much smaller, anacronistically-named “library OPAC”–I have no idea how there could be a coherent result.

Perhaps there can be, but try as I might, I cannot imagine it.