Imagine
a desktop on which computational tasks can be accomplished via the movement of
physical objects. Imagine identification tags that can communicate to
electronic “blackboards” and to electronic writing pads. Imagine an answering
machine that works by associating messages with marbles. Messages are replayed
by choosing the appropriate marble and dropping it in the appropriate slot;
they are deleted by putting it in a different slot.

These
engineering phenomena, and others like them, are the central players in Paul
Dourish’s Where the Action Is.
Dourish, once a researcher for Xerox, provides readers with a look at research
into human-computer interaction (HCI). His purpose is to articulate a
foundation for such an approach to designing computational systems. He does
this by looking at several trends in current software engineering, noting both how
they differ from prior approaches and how they are similar to each other, then
by augmenting such observations with work from prominent phenomenologists. This
culminates in the formulation of six principles for software design, analysis,
and evaluation. These principles explore the implications of general features
of the stance that has been articulated through the body of the book.

Very broadly,
HCI is a domain of research that studies and develops the ways that people
interact with computer systems. Dourish’s first chapter provides a history of
such interaction, revealing a broadening of the range of ways people interact
with computers. We have come a long way from the days when human-computer
interaction occurred directly through electrical circuits. More interestingly,
this range of ways of interacting has broadened to incorporate skills we
already have, such as the use of text and pictures to accomplish computational
tasks. Current HCI research is devoted to extending the range of interaction
even further by exploiting still more skills we already have.

Dourish
calls the stance he develops the “embodied interaction” approach to HCI. The
notion of embodiment must be treated carefully to avoid misunderstanding. Although
the implications of physical features of objects is part of the territory
explored here, physicality is not really what Dourish means by embodiment. The
central idea is “embeddedness”. Hence he also draws attention to social aspects
of the use of computers. His point is that the ways we accomplish many
important and everyday tasks are embedded in physical and social contexts that both
constrain and structure these tasks. Hence one way for HCI researchers to
exploit a greater range of skills is to develop systems that explicitly utilize
these physical and social contexts.

Dourish
looks at two trends in current engineering that instantiate the embodied
interaction approach. The first is “tangible computing” (chapter two). Research
projects that fall under this term develop human-computer interaction via novel
ways of using our physical environments. The examples we have already seen
exemplify this approach. Some tangible computing projects envision computation
occurring amongst a variety of objects, perhaps much more task-specific than
familiar PCs, that are spread throughout a given environment and that
communicate with each other. Other tangible computing research projects are
attempts to give familiar objects new computational abilities, thus aiding our
accomplishment of tasks with these objects, and perhaps making possible other
uses. Still other tangible computing projects design whole environments, such
as rooms for certain purposes, in which computational power is available to us
directly through the physical objects in these rooms.

The
other trend that exemplifies the embodied interaction approach is “social
computing” (chapter three). These research projects are devoted to incorporating
work from the social sciences that makes explicit certain features of social
reality into the design of software. Such social science work is important
given a perspective on software that emphasizes both that software users and
designers communicate and interact through the products of design processes,
and that various users interact through software, not just to communicate
directly, but also to accomplish shared tasks. The work of ethnomethodologists
is particularly important to Dourish’s explanation of social computing. Their
emphasis of making clear how social orderliness of action is achieved,
maintained, and intersubjectively understood is enlightening for software
engineers interested in using social skills to facilitate human interaction with
computers.

There is
much for philosophers to quibble with in this book. The admittedly whirlwind
examination of phenomenology (103-117) is the clearest matter. Other
philosophers will be repelled by the very brief and far from unified
discussions of topics in philosophical psychology, such as intentionality.
Other people will find the early (4) characterization of interaction
misleading. Here Dourish emphasizes the interaction of relatively special
purpose computational devices to accomplish tasks, in contrast to the rigidly sequential
approach implemented in algorithm-driven general-purpose computers. This topic
is clearly important to the rest of the book, yet human-computer interaction is the central topic. I expect many
philosophers to be bothered by the general lack of precision in the use of
language. Another possible problem is the lack of discussion of the work of
Hubert Dreyfuss, except incidentally.

It would
be a mistake, however, to dwell on such shortcomings. The reason is Dourish’s
ultimate purpose. He is not primarily interested in theory, nor in the nuances
of philosophical debates about the topics he raises. Instead, his aims are
ultimately practical. Given that he wants to guide HCI research projects, my
impression is that the small tastes he provides of matters close to the hearts
of philosophers will suffice. I have no doubt that these matters could be tied
to practice more satisfactorily, but for a short, general, programmatic book,
Dourish’s approach is enough.

The
first three chapters, containing the history of human-computer interaction and
descriptions of recent HCI research, will be of the most general interest. This
material will be useful for anyone whose work involves ideas about what
computers can and cannot do. It might also stimulate the imagination of those
in a place to speculate about ways to use computational systems. This is the
portion of the book most likely of relevance to thought about the therapeutic use
of computational systems. Finally, chapter three is somewhat important for
people studying human action. The material discussed here provides a useful
corrective to overly “intellectualist” views (in Ryle’s sense) of the nature
and production of action.

© Andrew Sneddon, 2002

Andrew Sneddon
is currently a Killam postdoctoral fellow in philosophy at the University
of Calgary, where he is working on a book in philosophy of action. His other
interests include philosophy of mind and moral philosophy, especially moral
psychology.

Categories: Philosophical

Tags: Behavioral and Cognitive Sciences