Posting to RDA-L
On 5/4/2015 7:31 PM, Goldfarb, Kathie wrote:
My concern about the addition of a relationship designation, e.g. Director, expecting that it will find only those records in which Orson Welles is the director, will also result in finding works where Orson Welles has the role of actor, but the director, who is also listed in a 700 field, with the designator $e director, will also be found. (I found this while testing out the ability to use these designation to narrow down the list of records).
That is precisely my point. And the same goes not only for every single name relator including all editors, translators, along with the others http://www.loc.gov/marc/relators/relaterm.html, but also for all of the other relationships we are adding, e.g.
775 08 |i Abridgement of (expression): |a Denney, Robert E. |t Civil War years |d New York : Sterling Publishing Company, ©1992 |h 606 pages ; 26 cm
Of course, our records already have access for abridgements, but the overwhelmingly majority of records encode it using 100/240 or 700$a$t. As a result, any searches for 775 information will have to also search those other fields if the catalog is to provide an even half-way truthful search result, just as searching (or sorting) results for “director” will always give a wildly incorrect result until all–or most–of that “legacy data” is updated.
When people compare our data with other organizations, primarily with how businesses deal with major changes to their data, the differences with library data becomes pretty clear. In a business environment, any information over 10, or often even 5 years old is pretty much obsolete (in a business environment almost nobody cares what was happening 15 or 20 years earlier: often the staff is 75% new, there are all kinds of new machines, new types of data and so on, and they are interested only in the information concerning the immediate future). Although businesses have legacy data like just anybody else, theirs is relatively unimportant for their needs so they can safely archive it, then start with new databases, along with new types of data and structures.
That cannot happen in libraries, where what was acquired 50, 100 or more years ago may still be just as important as ever. So, although businesses can avoid dealing with their legacy data by archiving it in various ways, libraries cannot just put their old records in a zip file and access them only when someone specifically requests it. It must all be merged together into a whole that provides reliable results when you search it.
As a result, with the changes taking place, tens (hundreds?) of millions of headings will need to be updated to add various kinds of relator information, if the catalog is to provide reliable information. That is just a fact. Adding that information will be the greatest retrospective conversion project ever undertaken by the library community. I have never read anything about this project that very obviously must take place: how long it will take, how much it will cost, or who will do it.
But somebody, sometime will have to discuss such practical concerns, that is, IF the catalog is to provide reliable information, which I think we can all agree is by far the most important function of the catalog. If the updates don’t happen, the results can only be very bad for everyone–from librarians to users.
None of this is nay-saying, just stating rather obvious facts.