Tuesday, March 2, 2010

Eliminating the Rule of Three

Posting to alcts-eforum
I don't understand why the question should be reduced to all or nothing. If we want to simplify the current rule of three to, for instance, to trace only the first three, or four, or five, that's fine. But tracing every name for all materials is absolutely unworkable in the foreseeable future. Look at a dictionary. Right now, it's pretty easy and doesn't take much time, but I don't think too many people would want to trace every single name for those who helped to write the dictionary. It would help absolutely no one find anything.

So, if we want to do away with the rule of three for the sake of simplification, we immediately are forced to complicate that same rule when we decide: "Do not apply" in the case of dictionaries, in the case serials, in the case of wikis, in the case of ... The number of exceptions to the rule would doubtless skyrocket and the practical cataloger would probably rely on the rule of "do as many as I feel like doing today, based on the amount of other work waiting on my desk and, let's face it, I had a hard night last night."

I think that if we are going to design a product that others want, we must focus on creating *standards that others can rely on,* (this is absolutely *not* the case that we have today, where we claim to follow standards but we don't) and that we can genuinely fulfill in a practical, everyday kind of way. Based on my own experience of seeing records that purport to follow AACR2/MARC21/LCSH/NAF/LCC and so on, I think that even the current standards may be set much too high. I don't know why this is so: is it because catalogers do not have enough time? Or that the standards are unreachable in a collective, sustained sense? Or that the catalogers have not been trained well enough in the first place? I think all are at fault, but especially the last reason makes sense to me. Morale may be an issue as well.

I don't think that libraries and other institutions are going to devote a larger chunk of their budgets to training and retraining. (What does this mean for training for RDA??) So, I think that eliminating the rule of three would ignore the reality of the working cataloger who has to deal with production quotas and is yet another example of that maxim, "Perfecting the irrelevant."


  1. Hi Jim,
    If one thinks of the current structure of cataloging workflow, your concerns make sense (as does another's more trivial but real concern about the time involved in spelling out abbreviations like Minn.->Minnesota). However, RDA is more about restructuring bibliographic description on a higher level. The publisher has all of the names associated with a work at time of publication, correct? For many of those publishers, that information is coded in metadata, ONIX or otherwise. RDA is reworking the metadata standards of the library world (as well as relationships with other entities including publishers) so that this metadata flows into the library bibliographic universe. Catalogers won't have to enter all of this data for that portion of works where metadata already exists. Add in the Semantic Web layer on this and URIs for the names that link to a data set shared among entities across libraries, publishing, and portions of the Web further lessen the work of individual cataloging institutions. Similarly, if the metadata comes from the publisher with Minnesota spelled out, catalogers don't need to repeat that work. And, if silo-cataloging slowly falls away, the overall impact on the cataloging world is lessened by many magnitudes. Kind regards.

  2. Allen,
    Thanks for the comments and you make some good points. At least some of them I answer in the posting to the JCS Secretary available at http://catalogingmatters.blogspot.com/2010/03/re-acat-rda-aacr2-and-simple.html

    Again, if we can get URIs and linked data to work, that would be great, but RDA does not deal with this: it deals with rules for inputting text. For example, if we could get publishers to add URIs to the respective name headings, that would be absolutely great, but the library community hasn't built the required tools yet.

    Publishers might be persuaded to use e.g. http://dbpedia.org/page/Lev_Lvovich_Tolstoy, and I might agree to this, but that would be a huge admission on the part of the library community that the library-created tools to do this do not exist and even then may not be enough.