The nature of our project also contributed challenges for
metadata creation. The diverse nature of the items required the
cataloger to quickly gain expertise in cataloging bookplates, fashion
plates, and historical photographs and negatives. Research performed
in one subject area did not usually apply to another; for instance,
knowledge of the history of Pratt Institute, necessary for the prints
and negatives, had no bearing on the fashion plates and bookplates.
As Pratt’s archivist, I was the main source of information for questions
about the historical images, but researching the bookplates and
fashion plates was a bit more complicated. The cataloger had to
consult resources at Columbia University to obtain background
information on the bookplates, their creators, and the historical
techniques used to create them.
The cataloger had to use a slightly different approach for each
type of material. The format differences across collections were a
given, but the interpretation of the creator, description, and subject
fields often varied as well. For example, the description field for a
fashion plate with printed text in French required a different format
and style than that for a photograph of an architectural drawing. In
hindsight, it would have been more efficient and cost-effective had we
focused on the metadata for one type of material at a time.
As they did with color reproduction, the bookplates proved
especially problematic. Many of the artists were identified by only a
monogram or other symbol, and a large number were undated, but the
cataloger simply lacked the time to research and identify each plate
more thoroughly. The techniques used to produce the plates (such as
etching or engraving) might also have been determined for many of
the plates had there been additional time or cataloging resources,
though the cataloger was able to do this for some of the plates. There
were issues at the quality-control stage as well. For example, we
initially entered the bookplate owners’ names in the format Firstname
Lastname. When we realized that this produced idiosyncratic sorting
in our image presentation software, however, we changed the order to
Lastname, Firstname—a relatively small change in the metadata
template that proved time consuming because it had to be
implemented in each individual record.
We also had to modify the Dublin Core metadata template in
order to provide more effective descriptions of the bookplates’ unique
elements. For instance, the person for whom a bookplate was made is
as important as the artist who created it, and the motto, when there is
one, becomes a major means of identification. We had to
accommodate these features by creating the elements Owner and
Motto, and we changed the Creator field to Artist because we felt
many people might misinterpret Creator to refer to the printer of the
physical object rather than the artist who designed the plate.
Additional modifications were also made (Figure PRAT-2).