10 1. INTRODUCTION
Revisions to the Anglo-American Cataloguing Rules were underway at the time. Reading through
the proposed rules for cataloging photographs, I was astonished that the notions presented had
little to do with the ways in which photographers, editors, advertisers, and others who worked with
photographs spoke of the or used them. eadora Hodges, asked if I would care to write a paper
on my concerns— “Access to Images: Axes to Grind.” It was here that the idea of using production
concepts and expertise as a foundation for modeling and representing movies began.
In 1981, lm theorist and member of the Berkeley Comparative Literature faculty Bertrand
Augst asked me: “Why can’t we use a computer to measure and speak of lmic structure in the
same way we can for a verbal text?” He had conducted computational analysis of French literature
in the 1960s, so this was an obvious question. e primary answer was that there was not yet any
available system for digitizing. So we worked on modeling lm in hopes of making some progress
so as to be ready when digitizing became available. We dropped the literary metaphor and devised
a time-varying signal set model.
During my research I still showed my lms in classes and in venues such as Canyon Cin-
ema. My portfolio lm on horse pulling generated reactions that provided an informal data set on
structure and meaning. As one might expect, the topic brought out a mix of reactions, but more
intriguing was the near even divide among viewers over the structure. e lm is a seven-minute
visual poem about an event I had attended since early childhood, so I made it an impression of
images and sounds. ere is no explanation of what is happening and only the intriguing, dancerly
movements of the horses and drivers are shown most all of the ordinary explanatory objects and
events are not shown. Also, the camera is always very close to the horses and drivers. Some people
“loved” the “abstractness,” while others “hated” that they couldn’t “tell what was going on”; likewise,
some “loved” being “so close to the action, like I am in the ring,” while a near equal number “hated”
the “claustrophobia.”
Working with information philosopher Patrick Wilson, I formalized my explorations into
a dissertation on lm and representation—how does lm represent the world and how might we
represent lm in such a way as to make it a tool with utility analogous to books and journal articles.
We might re-state this as looking at how lm structures time and space and how might we use
this understanding to generate topographic maps of lms to enable control over locus and depth
of penetration into a lmic text.
1.2.2 RICH MISE EN SCÈNE
Mise en Scène 5: Massaman Curry and Garlic Tofu
My entry into the eld of lm analysis is somewhat serendipitous. I was a doctoral student on the
cusp of a dissertation proposal. I had come into information science from the eld of Behavior
11
Analysis. My day job at the time involved managing the cybersecurity program for a large univer-
sity. I was hoping to write a dissertation that involved these two elds.
On the night my career in lm analysis began, I was having dinner with Brian to discuss
some barriers and frustrations I was having with my dissertation. First, I discovered that another
researcher in the cybersecurity eld who was farther along with his dissertation than I was had
independently pursued a similar idea. Second, I had encountered some bureaucratic resistance to
the idea of using real world cybersecurity data for my research. All in all, this was a distressing
state of aairs.
Our dinner arrived, garlic tofu for Brian and massaman curry for me. Over the course of
dinner, our conversation shifted from the specic problem I was working, nding an eective way
of visualizing and communicating complex and high volume security data (such as that produced by
network intrusion detection systems) to decision makers who likely did not have domain expertise,
to the theoretical underpinnings of my dissertation. e general idea was to transform the data in
a way that the consumer of the data could see everything at once and apply stimulus control prin-
ciples to draw their attention to the areas that needed attention.
As conversations tend to do, we moved from my dissertation to a problem that Brian had
been working on with his mentor, Betrand Augst. Augst had been interested for some time in an-
alyzing lm using computers much in the same way textual documents had been analyzed. Augst
thought the work of Raymond Bellour provided a potential framework for computation lm anal-
ysis. In “System of a Fragment,” Bellour provided a detailed and rigorous analysis of the Bodega
Bay sequence of Alfred Hitchcock’s e Birds. As Brian and I nished dinner, we realized that we
could conduct a functional analysis of Bellour’s work and develop a computer-based heuristic for
analyzing lm with modern technology.
A few days later, we printed 12,084 frames from the Bodega Bay sequence of e Birds on
a 9’×4’ color print and hung it on the wall. It did not take long to identify the signicant discon-
tinuities in the data that correlated with Bellour’s analysis of the same data. In October of that
year, Brian and I presented our initial work at the Document Academy in Berkeley, CA and had
the opportunity to meet with Bertrand Augst, describe our work, and show o our 9’×4’ print of
the frames.
Over coee, I described my approach as “tear apart to reveal structure; deconstruct and syn-
thesize.” My dissertation came about a year later and we have used the general ideas rst discussed
over massaman curry and garlic tofu to create the research described in this volume.
1.2 ORIGIN STORIES
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.