Saturday, February 12, 2011

RE: RDA and MARC (was Linked data)

Posting to RDA-L

Karen Coyle wrote:
<snip>
Quoting Weinheimer Jim:
> But I wonder if what you point out is a genuine problem, especially in an RDA/FRBR universe. The user tasks are to find, identify, yadda -- works, expressions, manifestations, and *items*. Not sub-items.
Jim, I think you're at the wrong end of the WEMI continuum -- what this record lacks is better access to *Works* contained in the manifestation/item. Items are the physical items, the thing you have in hand. The added entries in this record represent persons and works.
</snip>
Well, it is still a *valid* way of looking at it. The purpose of the traditional unit record is to describe the *thing* you are cataloging as a whole--after you decide what constitutes that "whole", and then link it to other records in various ways. The analytic (essentially an extra card--under main entry mostly--in the catalog) was to supplement matters to a point. The problem is and always been: how do you look at the "whole"? I'm sure the problem was always in the backs of our minds when we cataloged one of those single volume complete works of Shakespeare. ("But nobody can find Romeo and Juliet by Shakespeare with this record for his complete works!" This is another of those problems that worked more or less in the card catalog and does *not* work in the OPAC) But ultimately more important: if you are cataloging a conference, do you catalog each paper separately? If you are cataloging a serial, do you catalog each issue separately, or each article in each issue separately? I have done all of that.

These are some of the most important issues in cataloging because if you decide to, e.g. catalog each conference paper separately, your work may increase by 10 times or more and it is a huge responsibility to keep from crashing and burning. Promising something like that means that you have the resources to achieve it. Of course, we don't even try with journal articles.

But in today's world even that is not enough. Let's look again at those examples. In the catalog record for Boykan's Flume, we can see the problems clearly:

245 10 |a Flume |h [sound recording] : |b selected chamber works / |c Martin Boykan.

505 0\ |a Sonata for violin and piano (17:54) -- A packet for Susan (19:59) -- Flume : fantasy for clarinet and piano (10:40) -- String quartet no. 1 (18:58)

700 12 |a Boykan, Martin. |t Sonatas, |m violin, piano.
700 12 |a Boykan, Martin. |t Packet for Susan.
700 12 |a Boykan, Martin. |t Flume.
700 12 |a Boykan, Martin. |t Quartets, |m strings, |n no. 1.
The cataloger could have considered the 245 adequate, but leaving that kind of access is seen as kind of useless for music cataloging and they are expected to do much deeper analysis than most others. (In book cataloging, this is not seen as such a bad problem) Still, note that the 505 note is not enhanced since there is access through the 700s.

But compare this with the Amazon record http://amazon.libcat.org/cgi-bin/az2marc.pl?kw=B00006IZMR
505 0 |t Sonata for violin & piano: molto moderato -- |t Sonata for violin & piano: alla marcia -- |t Sonata for violin & piano: variations -- |t A packet for susan, for mezzo-soprano & piano: it often comes into my head -- |t A packet for susan, for mezzo-soprano & piano: the good-morrow -- |t A packet for susan, for mezzo-soprano & piano: bright star -- |t A packet for susan, for mezzo-soprano & piano: the owl and the pussycat -- |t A packet for susan, for mezzo-soprano & piano: well I remember -- |t Flume, fantasy for clarinet & piano -- |t String quartet no. 1: sostenuto -- |t String quartet no. 1: allegro -- |t String quartet no. 1: interlude, adagio espressivo -- |t String quartet no. 1: leggiero.
which has been parsed into 505$t and the "works" have changed. "Sonata for violin and piano" has been split into "molto moderato, alla marcia, variations." (separated only by that horrid punctuation mark!) The other pieces are split in the same way. So someone could say, "If I search in the catalog for "packet susan owl pussycat" I won't get it, but I will in Amazon. Therefore the Amazon record is superior".

But this still isn't the end. If you look at the public lecture of Jonathan Zittrain "The Future of the Internet" http://fora.tv/2008/05/15/Jonathan_Zittrain_The_Future_of_the_Internet (an important talk), it is a compete talk but is split into 17 parts, each with a different theme.

But even this is not the end. I have seen some tools such as HRAF that index each paragraph! http://en.wikipedia.org/wiki/Human_Relations_Area_Files. With full-text tools such as Google Books, it is hard to predict what is going to happen. In any case, the very idea of "what is a work"? is far from solved today.

The modern information environment is fragmenting. It is being sliced and diced into a bunch of little pieces, while others take what they want and make jambolayas of their own personal recipe, sharing and mixing, changing and discarding. To believe that the FRBR framework of work-expression-manifestation-item encompasses this new world, describes this new world, makes sense of this new world, or even helps to comprehend this new world, takes nothing less than--in my opinion--a tremendous leap of faith. And I have said before that I lost my faith a long time ago.

I am sure the Googley guys would never even consider thinking in FRBR terms. They will look at the situation as it stands and try "something". If that doesn't work, they'll try another one and another....

I think we need a new set of eyes.

No comments:

Post a Comment