Posting to RDA-L
On 21/09/2012 00:07, Kelley McGrath wrote:
I sometimes wonder what the silent majority on lists thinks. There are definitely people interested in trying to insert this kind of data into existing records. Many moving image (and music) catalogers are very interested in relator terms and codes because our materials include people performing many different kinds of roles and users want to know who is doing what. This doesn’t mean just when they’re looking at a single record. They might want a list of the movies directed by Clint Eastwood or the directors of recent French comedies or they might want to slice and dice the data some other way.
I am involved in a project that is trying, among other things, to retrospectively add role information to authorized names in records for moving image materials. We have an article in the Code4Lib Journal about our preliminary test, which including figuring out which 700(s), if any, were for the director: http://journal.code4lib.org/articles/775.
I also did a presentation at ALA in June about our current work to do this in a more sustainable, scalable way: http://pages.uoregon.edu/kelleym/publications/CCIRG_FRBRinMARC_DatainText.pdf or http://goo.gl/pFvFV. It’s not a trivial problem and we can’t get 100%, but we can do far better than 0%. My goal is to convert what we can to a machine-actionable form, identify and fix erroneously-converted info where practical, triage the rest and move forward.</snip>
This is very interesting, but how will it work in the real world? Let’s assume that this has all been done with an “acceptable” percentage of the records: 60%? 70%? 80%? You are working as a reference librarian and a senior faculty member on the library committee of your institution comes up to you and explains that he or she is writing an article and needs a list of the movies directed by Clint Eastwood. (Yes, the faculty member would have this information in other ways, but I am positing a reference question, and variations of this kind of question come up all the time). We also assume the reference librarian fully understands the issues in the catalog and knows that it will be 20%, 30% or 40% wrong.
What does the reference librarian do? The only responsible answer is that if you decide to point the faculty member to the catalog, you must explain the problems, which as I have shown, are very complex. To knowingly point someone to a tool that you know will not produce the correct answer without providing an explanation is–at the very least–unprofessional. In the case of film directors, luckily there are other options, such as the IMDB and (I am sure) other tools. At least I know what I would choose to point the faculty member toward if I were the reference librarian, if nothing else but to shield myself from future blame.
Now, to make this scenario even more realistic, let us assume that the faculty member avoids the reference librarian altogether and goes straight to the catalog. He or she sees an option to search by “film director” but how is this person to know that the result will be off by 20, 30 or 40 percent? They must know this, otherwise they will write the paper and risk looking like an incompetent idiot. The faculty member will be very angry. And who will be to blame?
It seems to me that there are certain tools for certain tasks. A screwdriver is not a chisel. A Smart car (are there Smart cars in the US?) is not a tow truck. Sure, you can try using the wrong tool for the task but the results may turn out to be very bad. The library catalog is a tool designed for certain tasks. To ask it to do more than it is designed to do will have consequences.
I don’t believe that bringing this up is being negative or part of the “profession of no”. Instead, it is positing a normal and realistic situation seen in the library everyday and admits that there will very definitely be consequences to human beings and to the library profession as well.