Posting to Autocat
On 4/18/2014 11:10 PM, MULLEN Allen wrote:
It is unfortunate that I, a largely ignorant cataloger on the sidelines in a small city public library, feel a need to explain the goals and possibilities of RDA/Bibframe. It is likely I’m off-base in one or several respects, but this is my best take on it, and I really have not seen a clear, easy overview of this for catalogers to digest anywhere else, so I offer it. It is partial, but my time of this public service desk is up and I need to return to cataloging.
RDA needs improvement and RDA needs further development. But it’s a pretty decent foray if one believes in the utility of library metadata. Whether it is sufficiently useful to stand up in a world of sophisticated algorithms, predictive results, user profiles, data mining, etc. is open to the future to answer. Those aren’t our strengths though.
The problem is not that I do not understand the goals and possibilities of RDA/Bibframe. I have gone to great lengths to show that I do understand them. I am guilty of an even more heinous crime: understanding and disbelieving–or at least not allowing myself to put my faith where there is no evidence, and my experience tells me that the goals of RDA/Bibframe do not deal with the problems I have seen–with my own eyes–that people have when they use the catalog, and with the problems I experience myself. I have worked with scholars and researchers and students from around the world, and done so in international institutions. I have even built tools using different technologies, some completely on my own that I realize were not very good, and better ones where I have played a supporting role.
Therefore, my skepticism does not come from lack of understanding, but seeing that RDA/FRBR/Bibframe do not address the very real problems that the public is facing when they use the tools we make. And yet, I still believe in the utility of library (more specifically: cataloging) metadata. RDA concentrates on changing individual records, e.g. including an author’s degrees and affiliations in the SR, spelling out abbreviations, adding the relationship info, without any concern for the trade-offs involved, the costs, or even to find out if this will make any real different to the public. It is my belief, based on my experience, reading some of the relevant literature (not only from the library world), and discussions with others, that the main problems people have when using a catalog is not with individual records, but with how the catalog itself works and how it is structured. Some of this is not new at all, and we can find evidence in very early writings on the catalog. Other problems are new.
The public is being bombarded with outrageous amounts of information every, single day (I keep referring to that talk about the IT person who is adding 8000+ records to his catalog every day, but this represents only a fraction of what the public experiences), and libraries feel under pressure to provide people with a single search box to search “everything” in one search (whatever the word “everything” means to a particular library). The single search can be achieved in different ways, by adding records to the catalog, like the IT person in the talk is doing, or it can be done through federated searching, but what is important is that for the user, the result will ultimately be the same. They will see all kinds of records thrown into the same pot and the very concept of authority control begins to disintegrate since 99% of everything the user sees will be made by non-library organizations.
These organizations will–at best–follow their own internal rules but not by any means AACR2 or RDA or even ISBD. In such a metadata mash, how can a well-experienced and knowledgeable librarian tell someone how best to search the writings by Mark Twain? Adding the relationship information becomes completely random as well. In that metadata soup with all kinds of records, standards and non-standards, what can you tell someone who wants Vladimir Nabokov only as a translator? It will all be hit and miss.
In such a situation–which is the situation our users face now every day–discussing the minutiae of RDA, e.g. relator codes, is as useful as washing your windows when a tsunami is approaching. This is only one problem that people experience with our catalogs. I have gone into long discussions of some of the other problems, but what I am trying to say is that the problems the public experience with the catalog are not with the cataloging rules. We can change them in any ways we want and they will still experience the same problems.
These problems will also not be solved by changes in format, since that is also not the problem. Changing from ISO2709 record transfer should have been done long ago, but there is no “miracle format” that will change anything for the user. With a more web-friendly format (i.e. a variant of XML), other entities will be able to take our records more easily, and while that is primarily a good, the underlying problems still remain. Strange headings that would never enter a user’s head, such as “Characters and characteristics in literature” will be even stranger outside the library catalog environment.
Changing to an FRBR-type format will also not change the user’s experience of the catalog. As I keep pointing out, anybody can do the FRBR user tasks now, and nobody cares. The XML format right now is so powerful that we can make these search results look and operate any way we would want to. Why can another format make any difference? In the card catalog, people had no choice except to do the FRBR user tasks, and when other tools arrived, they discarded the old ways immediately and never complained.
Implementing linked data may very well change the user’s experience, but first we have to figure out what to link to (id.loc.gov? VIAF? dbpedia? Something new?) and by doing so we may be able to add information from Wikipedia or whatever, and that may or may not be useful to the searcher. In any case, it remains an article of faith that linked data will be the solution, and in spite of the hype, from what I have seen of linked data so far, I have seen nothing so incredible. Anyway, RDA and FRBR are irrelevant to linked data. All you need for linked data is a change of format and Bibframe should allow it, but the change in format should have been the first step and done long ago, by the early 1990s. I feel that linked data will be a part of the solution, but only a part.
I have explained it all yet again to show that I do have an understanding of matters, and why I cannot hold the faith that RDA and Bibframe solve the problems that people experience with the catalog. And yet, I say again that catalog information can be very useful for the public, but we must create tools that focus on their problems. We can only do that by learning more clearly what those problems are, and build tools that serve that purpose. That can be done only through studies of user behavior, and building tools whose success will be based on trial and error.
Unfortunately, I don’t think that is going to happen and that we are experiencing “RDA, FRBR, Bibframe! Damn the torpedoes! Full speed ahead!”