CCO In Action Ð Work Context
Video Transcript

In this series, we will explore what constitutes a work, how different works can be
related, and how collections, both physical and virtual, can be represented. One of the
basic tenants of CCO is that both a work and an image record is required. VRA Core
4 is a standard which uses this structure, and so our examples will use VRA Core 4
as a starting point.

Defining what constitutes a single work record, and when it needs to be split into
multiple records with part/whole relationships is a common task. When you find that
different components have different time frames, creators, titles or are in different
repositories, then you may require more than one work record, plus a relationship.
Cataloging always represents choices based on your local needs and users. These
are mitigating factors, but the goal is always that the asset can be easily found by
users of differing levels of expertise.

Using standards is highly desirable, and it is always possible to create minimum
records within a given standard framework. Adopting a relatively rich standard
framework, even if you are not currently using all the elements or fields, allows your
data to be expanded over time, if that becomes desirable.

There is no one ÒrightÓ way to catalog, rather, best practices are about efficiency.
Consistency in data entry and field values through the use of standards means you
will be able to automate processes, reliably manipulate data for different uses, and
more easily remediate and migrate your data. In other words, make sure your data
has continued value over time.

WeÕre going to look at the problem of cataloging plates or illustrations in books or
periodicals, as an example of thinking through cataloging decision making and
testing user access.

In the case on the left, we have plates in a rare book. They are all by the same
artist or author. They are of different subjects. On the right we have plates in a
fashion periodical, all by different artists, representing different fashion designers,
of different types of clothing.

LetÕs look at a decision process for the book, thinking about access points for the
user. This book happens to be in a rare book collection, with a MARC record, but
with no online images.

The images are obviously part of the book. How could we represent this in an
image delivery system? The MARC record can be used as a reference, or a link as
well.

All of the plates reference works by Palladio. Because all of the images are also by
the same author or artist, done at the same time, they can be represented by a
single work record, with more specific information in each image record. Note that
with this approach it is vital to standardize that information, in this case building
name, and plate number in each Image Title (the caption).

Think through the test of this structure; will it be discoverable? Will this work well
with other cataloging solutions you have used? With other images of buildings by
Palladio? In this case will it come up with the building name, architectÕs name as
well as the book title and author?

It is always a good idea to run a test, if only a single record, loaded into your local
delivery system. Searching on Òvilla rotundaÓ shows that images of the actual
building do come up, along with the plates from the bookÑit passes the test as a
viable cataloging solution.

Note that although both CCO and VRA Core 4 require image and work records, in
a display system these are mapped and flattened into one display record. If you do
not have a relational database, you can still construct a flat record template that
corresponds to standards; while not recommended, this may work for some
collections, or as a short term solution.

In our second solution, each plate must be cataloged as a separate work, because
of different dates, fashion designers, and illustrators. The plates can be linked to a
work record for the entire run of the periodical, or by using the periodical as the
Source element for the plate.

This happens to match the MARC record, because in this library, the plates have
been removed and are accessed as separate sheets. But, again, no image is
available in the OPAC.

Using a place holder image, a work record can be created for the whole run of the
periodical, this can link to a MARC record. The relationship to the illustration work
record is Òpublished in.Ó

Here, what is represented in the Image Title is the Plate number, Volume, Issue
and Date. The title, illustrator, and fashion designer appears in the Work record.
Again, loading a test in a delivery system and searching by ÒGazette du Bon TonÓ
turns up all the plates, and searching on individual artists and designers also
works. It passes the test.

Here we will look at a literal ÒcomplexÓ record. What is key with this decision
process is thinking about the possibility of adding more images in the future, and
whether our decision will be adaptable over time. Here we are looking at multiple
buildings at the Taj Mahal site.

Decisions with this material might be made by local curricula and graduate or
undergraduate users. A major issue with cataloging for all delivery systems is
whether data can be updated and refreshed, as new images are added to existing
work records over time. This has to be factored into decision trees.

Many of us catalog the Òwork in handÓÑpreparing groups of images as needed
immediately in teaching. In other words, it will affect our decision today if we only
have two images of the gatehouse and 20 of the Taj itself. Here a pragmatic
solution is to create a record with the title ÒTaj Mahal [Complex]Ó and the
gatehouse information in the Image Title for now. It will be possible to have this
[Complex] record carry forward in time as the Parent record with the ability to
create Child records for the other buildings in the future, as Òpart ofÓ the complex.