Buxton, W. (1994).  Metaphors that keep us on the periphery. Human Computer Interaction,  9(1), 50-52.

Metaphors that Keep Us on the Periphery

Bill Buxton
University of Toronto
Toronto, Ontario
Canada M5S 1A4
tel: (416) 978-1961
email: buxton@dgp.toronto.edu



Brown and Duguid make a strong case for paying more attention to the periphery and to context in design.  One senses in their arguments, however, much deeper roots in philosophy and literature than in design.  As a result, the implications - no, rather the opportunities - for designers are weaker than they might be.

I think that the problem lies in an over-reliance on the literary metaphor that pervades the paper:  text and context, books and publishing.  At first glance, the examples are compelling.  But it is precisely that which makes them problematic.  Because they fit the arguments so well, one tends not to stop and wonder how thus restricting the repertoire of metaphor constrains thinking about the problem.

Let me illustrate this by way of two examples - examples that are based on an assumption that the information and communications   technologies that we are designing are not always best understood by comparing them to what has preceded them.

Books Can't Read

Throughout the essay, I was struck by a strong asymmetry.  The authors argue cogently about the importance of the periphery and context in fostering communication.  On this point, I am in total agreement.  Books, computers and telecommunications equipment are similar in that context and the periphery have a large influence on the sense of meaning derived from them;  however, the nature of our communication with them is radically different.  Simply put, the channel of communication is unidirectional in the former, and bi-directional in the latter two.  We don't have interactive dialogues with books.  We do with computers.

From this asymmetry emerges a missed opportunity.  The problem is this:  if knowledge of context and the periphery is important for communication, then that should be true in both directions.  Not only should design consider context from the human's perspective of the technology, so should it do so from the technology's perspective of the human.

Whereas a book can't read, computer and telecommunications technologies do have the potential to sense and react to the periphery (physical and social). When confronted with foreground communications, with appropriate design, these technologies have the potential to interpret these communications according to the background context within which they are made.

As an example, imagine a VCR in a conference room that is sometimes used for videoconferencing.  One can imagine instrumenting the room such that what happens when the "record" button is pushed depends on the context:  if in a video conference, both sides of the conference are recorded, otherwise, just what is occurring in the local room.  The user need only initiate recording.  The system senses the context in which the record request was made, and configures the room's A/V network accordingly.

Such sensing of the periphery by technology is not generally done.  But it can be, as is illustrated by examples such as security systems and Xerox PARC's "Reactive Environment" project [Elrod, et. al, 1993].  It is a relatively small step to employ this background sensing to provide the peripheral awareness that would enable foreground interactions to be interpreted in context.  This is directly supported by the authors' arguments.

As we add cameras, microphones and other sensing mechanisms to our systems, this becomes an increasingly practical option.  I would argue that this property does distinguish these artifacts from those which preceded them, and by its virtue, one way to compensate for the "loss of continuity" embodied in them is provided.

Resolving one Paradox with Another

What the authors describe as the "paradox of demassification" drives the second half of the paper.  Central to this is the view that physical demassification works at counter purposes to social demassification:  technology breaks down social networks just as they are needed to provide the support to deal with other more specialized technologies.

However, another observation is omitted from the essay:  how the technology can be turned back upon itself to address problems that it helped create.  Perhaps the best example of this is how those same technologies that bring about physical demassification afford the provision of a "social prosthesis" that can counteract social demassification.  These affordances emerge from so-called "mediaspace" and "telepresence" technologies [Bly et. al, 1993].  These increasingly have the capability to effectively support tight social networks among communities which are otherwise dispersed.


Integrating "participant observation" and other anthropological techniques into the design process is to be encouraged.  Clearly design must equally reflect our motor-sensory, cognitive and social skills and mores [Buxton, in press].  To date, the social side has been greatly neglected.  But lest the pendulum swing to the other extreme, let us be clear that a holistic approach that incorporates all three is essential.

Design must be holistic.  And, I believe, the approaches discussed in the Brown and Duguid essay must be matched with an approach that reflects an intimate familiarity with a suitable repertoire of potentially applicable technologies and their affordances.  From such an outside-in approach, perhaps the social prostheses that all of us are striving for will emerge.


Bly, S., Harrison, S. & Irwin, S. (1993).  Media Spaces:  bringing people together in a video, audio and computing environment.  Communications of the ACM,  36(1), 28-47.

Buxton, W. (in press).  The three mirrors of interaction:  a holistic approach to user interfaces.  In L.W. MacDonald & J. Vince (1994).  Interacting with virtual environments.  New York:  Wiley.

Elrod, S., Hall, G., Costanza, R., Dixon, M. & Des Rivieres, J. (1993) Responsive office environments.  Communications of the ACM,  36(7), 84-85.

William Buxton is an Associate Professor of Computer Science and Scientific Director of the Ontario Telepresence Project at the University of Toronto.  He is also a consulting reearch scientist at Xerox PARC.