Planning Obsolescence: Dynamic Architectures for Digital Scholarship

Digital scholars have discovered the virtues of pastiche; scholarly
projects of the sort developed at the Institute for Advanced Technology in
the Humanities consist not merely of individually authored textual data and
associated media, but large supporting archives, fragments of other
projects, and links to outside materials. Given this shift in the paradigm
of authorship and scholarly production, one might assume that consumers of
scholarly materials will likewise come to view the task of research itself
as bricolage--the act of assembling and recombining digital media into ever
more specific forms. This paper seeks to theorize information retrieval in
light of this fundamental shift, and to present some preliminary research
undertaken as part of the Supporting Digital Scholarship project at the
University of Virginia.

The Granby Suite, an xml-based search and retrieval system for full-text
documents developed by the Virginia Center for Digital History and
Institute for Advanced Technology in the Humanities, employs an extremely
modular architecture with strict separation between tiers devoted to
search, document assembly, and rendering. In this paper, I argue that such
design principles are not merely a matter of good software engineering, but
are in fact a consequence of the new paradigms of digital scholarship. The
digital library of the future should allow users to re-present materials
from diverse media spread across a number of independent projects. Such
re-presentations will require that the underlying mechanics of the search
and delivery system allow for constant remodification. As such, the Granby
Suite demonstrates that the technical exigencies of implementation and the
functional requirements of humanities scholars must give rise to design
methodologies flexible enough to allow for constant and sometimes sudden
changes in the very architecture upon which the system depends.