Image indexing in the Bodleian ballads project
| DOI | https://doi.org/10.1108/eb040644 |
| Date | 01 March 1997 |
| Pages | 51-57 |
| Published date | 01 March 1997 |
| Author | Alexandra Franklin |
| Subject Matter | Information & knowledge management |
Image indexing
in the Bodleian
ballads project
by Alexandra Franklin, Special
Collections, Bodleian Library,
University of Oxford
The creation of a digital database of
broadside ballads at the Bodleian Library
presents the challenge of cataloguing and
indexing a large collection of documents
containing both textual and visual material.
The woodcut illustrations of
the
ballads will
be indexed by subject using ICONCLASS,
but it is not clear how this indexing should be
integrated with that of
the
textual material.
How does the cataloguing of images differ
from the cataloguing of texts? And how
successful is ICONCLASS in establishing
communication between the indexer and the
user?
Introduction: history of the
project
The Bodleian Broadside Ballads project is funded
by the higher education funding councils as part of
a programme to preserve and make more widely
available specialised research collections in the
humanities. With over 30,000 titles in five major
collections of
ballads,
the Bodleian owns an
important resource for researchers in popular
literature between the 16th and 19th centuries. As
single sheet items, the ballads are ideal candidates
for preservation on microfilm and by digitisation;
they are relatively easy to photograph but also, in
many cases, relatively fragile items which would
not withstand a much greater level of physical
handling. The task undertaken in this project was
to produce an online catalogue allowing users to
search through this vast body of
material.
The
funds provided for the conversion to machine-
readable form of existing handlists noting the
titles,
first lines, and publishers of the ballads;
iconographic indexing of the illustrations to the
ballads, and the creation of a computer package
incorporating the enhanced online catalogue and
scanned images of the entirety of the ballad
collections are included in the funded project to
further enhance the utility of the online catalogue
to users from various disciplines. Creation of the
catalogue was planned in conjunction with a
project to microfilm and scan the ballads, for
the purposes of allowing wider access to the
collections and to ensure the preservation of these
materials, some of which date from the sixteenth
century. In accordance with the terms of the grant,
which demand that the material be made generally
accessible, the computer package is being planned
as a Web site.
Since 1980 a handlist of the ballad collections had
been available in the library, but computerisation
has made this handlist searchable across many
more fields than first line and shelfmark, as had
been the case with the hand-written cards. It was
also intended to extend the utility of the catalogue
both within and beyond the Bodleian's walls. The
pre-existing data consisted of two files: one re-
cording titles and first lines of ballads, together
with author and subject information, the other
consisting of publisher information and other
information relating to the physical item. A com-
puter listing existed for about one third of the titles
file. The data was not in AACR format, nor the
computer file in MARC. The level of bibliographic
information and the depth of cataloguing under-
taken in the present project were determined partly
by the nature of the material and partly by the
constraints of the funding, which provided one
data-entry/cataloguing position for
3
years and one
computer programming position for 18 months. A
full-text markup of the ballad texts was not
planned as part of this project, because of the time
and cost involved and because the ballad collec-
tions contain many similar, and sometimes
identical, texts, many of them already well known
to the potential users of the database. Cataloguing
work has focused on the accurate entry of the titles
and first lines, by which ballads are usually identi-
fied, to the entry of multiple subject fields, and to
the entry of publication data which facilitates the
dating of
imprints.
Names associated with the
texts,
and all publication data, have been entered in
MARC-compatible format, and further editing is
planned to maximise the MARC compatibility of
VINE
107
—
51
To continue reading
Request your trial