The Development of the Poetry Portal at the Beck Center, Woodruff Library, Emory University

The Development of the Poetry Portal at the Beck
Center, Woodruff Library, Emory University

Alice
Hickcox

Emory University
ahickco@emory.edu

Julia
Leon

Emory University
jleon@emory.edu

2003

University of Georgia

Athens, Georgia

ACH/ALLC 2003

editor

Eric
Rochester

William
A.
Kretzschmar, Jr.

encoder

Sara
A.
Schmidt

Emory University Library's electronic text center, known as the Beck Center for
Electronic Collections and Services, holds a number of poetry collections that
are served out on the web as separate databases. In order to span these
collections of more than 200,000 poems we set out to develop a portal that would
allow searching across several different databases. The portal provides students
and faculty a tool for retrieving and reading individual poems for personal and
classroom teaching and research. The portal was implemented using XML
technology.
This is a story of the development of the site, explaining why the project was
conceived, and how the library, Information Technology Division and faculty
collaborated to design and implement the portal. The web-interface design and
technical architecture will also be described.
Since 1995 the Beck Center has built a collection of texts that are tagged in
SGML; many of these texts are poetry databases. Both commercially and locally
produced texts comprise the electronic poetry holdings of the Beck Center. The
texts were served separately out in a number of discrete poetry databases, each
with its own search interface.
Emory's Irish Poets collection is one example of a locally produced digital
collection. A portion of Emory's literary archives of Ireland's leading poets
was digitized. The archives were developed over the course of the last twenty
years, and includes worksheets of the poets affiliated with the Belfast
Group-including Seamus Heaney, Michael Longley, and Paul Muldoon.
In addition to the substantial body of electronic texts that existed, campus-wide
interest in poetry was building, demonstrated by the establishment of a Poetry
Council which sponsors periodic poetry readings.
These forces converged to raise questions about how to make the use of on-line
poetry resources easier to access and more logical in organization and use. A
group of people from the library and the Information Technology support group
for faculty came together to plan for a poetry portal. The Beck Center provided
data management. ITD provided analysis and programming support. An informal
focus group of Emory faculty who teach poetry provided subject matter
expertise.
The Beck Center put forth the original vision of the project: to present poetry
across multiple collections through a common web portal. A prototype of the
portal was built as a proof of concept of the use of metadata and XML
technology. It also served as the launching point for further design.
The unifying element of the disparate collections is the metadata, which contains
information that was deemed useful for cataloging, indexing, and referencing
items.
The portal was built entirely with open-source software, with the one exception
of the search engine. The process of building the portal may be divided in to
two sub-applications: data preparation and the web interface.
Data preparation involved converting the source collections from SGML to XML.
From the XML metadata was created using XSLT and SAX software. The metadata was
stored in the Dublin Core format.
The web interface for the poetry portal employs a suite of XML technologies.
Specifically, Tomcat and Cocoon, from the Apache Software Foundation, serve up
XML source files on the web. XSLT (XML Style Sheet Language Transformation) is
used to transform XML to HTML.
The prototype application was demonstrated to English faculty involved in
teaching poetry. A blue-sky discussion of how a portal could be useful and
interesting in teaching and research formed the basis of the application
requirements.
The interface provides browse and search capabilities. The collection may be
browsed by author, title, first line, collection, or date. The user may search
for poems, either by word used anywhere in the poem, by author, date or title.
Either browsing or searching will take the user to a poem window or a series of
poem windows.
From a poem window the user may also search for other poems by the same author,
other poems of the same date, other poems with the same title or other poems in
the same volume or collection. The user may also execute a simple search in the
Oxford English Dictionary from the poem window, and
bring up results in another window.
Additional features that we hope to implement include user-defined "poetry
notebooks," which allow users to save links to particular poems or searches, a
search history for registered users, and the ability to save searches.
Some associated features that were not part of the XML search-retrieval
programming also emerged as special projects. One such project would be to
collect manuscript and published versions of certain poems of poets that were in
their collections in various forms, and to present them for study. Another
specialized application is to have poems accompanied by audio files so that
students can hear the poems read aloud. For some of the poems in their Special
Collections, tapes were available of poets reading their own work. Other
possibilities exist for audio performance as well.