INTRODUCTION
The Orlando Project at the Universities of Alberta and Guelph aims to produce
the first full scholarly account of women’s writing in the British Isles in
a mode of literary history designed to take advantage of new technological
capabilities. To enable researchers to discover the sophisticated and
nuanced interconnections among this complex mass of material, Orlando has
produced a custom-designed SGML text encoding system capable of reflecting
literary and historical interpretation.
Aspects of this encoding scheme have been discussed in other forums°,° and the ways in which the encoding scheme has, as
planned, turned out to be effective in supporting the identification of
novel and significant interrelationships in literary history has been
presented°,°,°. During its development, the use of the
Orlando textbase and its encoding scheme to explore such relationships has
been limited to the immediate research team, using the in-house Orlando
development tools. But the vision of the Orlando project is, of course, to
present its encoded textbase to the larger research community; i.e. to
create a web-based delivery system that researchers world-wide can use to
explore Orlando content. This presentation will take a systems view of the
Orlando Project, focusing on issues and methods that have arisen in
producing the first version of this delivery system
The three key aspects of this systems view, which will be outlined here, are:
(a) the in-house Orlando development environment, its tools, and the
decisions made in crafting this environment for the construction of the
textbase; (b) the definition and design of a prototype system to achieve
certain core, albeit limited, functionalities without precluding the later
design and implementation of a more powerful system; (c) general issues in
designing an interface that will lead novice users step by step to
exploiting at first a manageable selection from an integrated set of
literary historical materials and complex underlying encoding scheme, and so
by stages to fuller use of the potentialities of such a scheme. A
demonstration of the first version delivery system will both clarify and
concretise the issues discussed in the presentation.
IN-HOUSE SGML DEVELOPMENT ENVIRONMENT
In brief, the development environment supports the creation and ongoing
revision of the three components to the Orlando system. The first of these
is a textbase containing SGML documents for the biography and writing career
of each individual writer, and for historical topics and issues. The
biographical documents contain newly researched material on the lives,
backgrounds, and activities of women writers. The writing documents contain
newly researched material on literary careers, and the production, textual
features, and reception of texts. The topic documents contain newly
researched material under a range of headings deemed crucial for coverage.
The second component to the Orlando system is an Oracle database which holds
information on historical context: brief accounts of historical events and
processes chosen to reflect each period’s literary, cultural, and social
concerns, anchored to a date or date-range. The goal of this material is to
enable a user to produce a chronology—an ordered set of events—relating to
any particular time period, writer, word or concept. Sorting by event-type
or level of priority is an obvious use of this database. The third component
of the Orlando system is another Oracle database, which holds full
bibliographic details of all primary texts and secondary sources referred to
in the electronic text.
The investment in SGML necessitated using SGML-compatible software such as
Oracle in the development environment. The way in which this development
environment has in turn affected the development of the delivery system will
be outlined in a detailed chart.
DEFINING AND IMPLEMENTING A PROTOTYPE DELIVERY SYSTEM
The first Orlando delivery system had several goals. The primary goal was to
provide web-based access to most of the core Orlando materials (writing
documents, biographical documents, and events) to create chronologies and to
provide the first automated hyperlinks.
The SGML encoding scheme includes a number of tags by which materials in each
of these separate databases are related. For example, a
<BIBCIT> (bibliographic citation) tag that appears in a
writing, biography, topic or event prose refers, by means of unique
identifiers of records in the Oracle database, to complete bibliographic
material. Also in development is an SGML-tagged name authority list, which
not only ensures accurate hyperlinking (via the <NAME> tag)
across the various databases, but which also aids in search and
retrieval.
The prototype delivery system allows a user to access the Orlando material in
three primary ways: through a writer’s name, through event chronologies, and
through several thematic entry points. Given a writer’s name, the writer’s
biographical and writing documents are retrieved, and a chronology of events
mentioning that writer’s name is computed and displayed. Certain “core tags”
(name, place, date, title, organization name) act as hyperlinks to other
areas of the textbase. Users can also access Orlando material through event
chronologies generated by means of a freetext search on a given word/phrase,
or a search by tagged name, genre, title, place, or organization name. (For
these categories users are free to type in a search term or select from a
list of possibilities). Users can also specify a given time range, and/or
specify particular event types and/or priority levels. Thematic entry points
are a means of entering into Orlando by way of material that cuts across the
textbase in ways that highlight certain themes. “People” allows users to
find people by name, historical period, occupation, or what they wrote”;
“Texts” allows users to discover texts by title, subjects, or types of
writing”; “Contexts” allows users to search on various topics,
organizations, and places in women’s literary history; “Networks” allows
users to investigate literary, social or family connections, organizational
links or intertextual relations; and “Identities and Politics” allows users
to investigate cultural and political issues.
The full power of the Orlando system, of course, comes with sophisticated use
of its interpretative tags for information exploration, i.e., those tags
such as <INTERTEXTUALITY>, <CULTURAL FORMATION>,
<RELATIONS WITH PUBLISHER>, and <POLITICS>,
which the authors of documents have used to mark up text for literary
historical interpretation. However, for the first delivery system, issues
taking precedence were those related to working with extant web browsers and
XML software, those related to automating the transformation of SGML into
XML for web delivery as materials are moved from the development to the
delivery environment, and those concerning a first pass at a user interface
to what will ultimately be a powerful information exploration system (but
that will always need simpler ways in to be available). Orlando has taken a
modular approach to the definition, design, and implementation of its
delivery system and hence this first version did not fully exploit these
interpretative tags.
ON USER INTERFACE ISSUES FOR INFORMATION EXPLORATION
In the first version of a web-based delivery system, we aimed to follow good
user interface design principles. This involved, firstly, developing a
powerful yet usable interface that would allow both novice and expert users
to access Orlando materials. (For example, a novice user may choose to enter
directly through a writer’s name, whereas an expert user may choose to enter
through a more complex chronological search.) Secondly, this involved
revealing to the user a portion of the underlying interpretive scheme to
allow some exploitation of it. (For example, thematic entry points reveal to
the user some of the complexity of the Orlando interpretive scheme.)
Thirdly, this involved arriving at design principles and choices to create a
coherent display of the material. The paper will outline the ways in which
this process served as a foundation for the next stage of delivery work,
currently underway, which is focused on representing the interpretative
markup, and ensuring that the interface allows for maximum exploitation of
it. By conference time this next phase of delivery work will be completed in
its first instantiation and ready for demonstration.
CONCLUSION
This poster presentation will introduce the first version of the Orlando
delivery system, discuss in detail the progress that was made in the
creation of this system, and outline the ways in which it provided a basis
for subsequent delivery development.