NEH Digital Humanities Start-up Grant, Level 1 Proposal Narrative
A unified approach to preserving cultural software objects and their development histories
Enhancing the humanities through innovation. We will design, and create a pilot implementation
of, the first full archival methodology based on the needs of digital humanists, as exemplified by
software and game studies. This methodology will integrate the approaches of the Joint Committee
on Archives of Science and Technology with the recommendations (and ongoing work) of the
Preserving Virtual Worlds project, as described below. This work is urgently needed, because there is
very little archival work being done in software preservation that focuses on the needs of future
academic inquiry — and all cultural software products will benefit from an approach that focuses not
just on the objects but also their development history and context. In looking for an available
software object that allows for inquiry from a general software development standpoint as well as
that of software and game studies (our collective areas of humanities focus), we have chosen the UC
Santa Cruz game Prom Week. The game's development has produced numerous research
publications in the field of artificial intelligence and the game itself has received nominations for
international game design awards. It represents a serious computational research object and an
innovative step forward in interactive narrative games. We will use both these contexts to pilot a mix
of archival strategies never before combined, which we believe holds significant future promise.
This work innovates simultaneously in the archival processes of appraisal and retention.
Appraisal is the means through which archivists collect and analyze the future preservation value of
an object or collection. Generally with the help of a domain specialist, the archival team aggregates
all the work related to the object of preservation that is potentially significant and then winnows it
down before retaining it (storing it safely and securely). Because software objects are relatively
recent phenomena there is no generally agreed upon strategy for either appraisal or retention. We
want to change this by using archival methodologies developed for scientific research to boot strap
the creation of archival solutions for software and computer games, combining them with the best
practices identified by the major investigation of preserving games and virtual worlds thus far.
In 1983 the Joint Committee on Archives of Science and Technology (J-CAST) released a
report detailing the issues concerning appraisal of scientific research. They emphasized the necessity
of preserving the process of scientific research as well as its results. Science is a messy business and
the details of how scientific discoveries progressed through fits, starts, and flashes of insight are
considered as important to future technology historians and students as the published results. The
report recommended that when archiving scientific research every effort be made to gather all
material related to the process of discovery — future researchers could then learn from the missteps
and methods to supplement their understanding. We are asserting that this approach to the process of
development is also applicable to cultural software, and games in particular, as they are potentially
more opaque than most projects due to the inherent complexity of their computational systems,
tendency to display this complexity only partially at the interface level, and highly iterative
development processes. By revealing the development history and the inner workings of Prom
Week's systems we hope to provide a clearer picture to those wishing to understand all software
development and its historical significance.
In developing an appraisal strategy for Prom Week the team will consult with the project's
original developers and collect early prototypes, source code, correspondence and other research
detritus in line with J-CAST's recommendations. We will then sort, cull, and organize this
information for long term archival storage. Along the way the J-CAST guidelines for process
documentation will help us form a picture of the game that can be used by future historians, game
developers and students to understand the inherent complexity of the game's artificial intelligence
and narrative generation systems. This work will result in a more detailed methodology for future