No earlier issue with same topic
Issue
Previous article
Article
SIGCHI Bulletin
Vol.30 No.3, July 1998
Next article
Article
No later issue with same topic
Issue

As We May Communicate

Carson Reynolds

Abstract

The purpose of this article is to critique and reshape one of the fundamental paradigms of Human-Computer Interaction: the workspace. This treatise argues that the concept of a workspace---as an interaction metaphor---has certain intrinsic defects. As an alternative, a new interaction model, the communication space is offered in the hope that it will bring user interfaces closer to the ideal of human-computer symbiosis.

Keywords: Workspace, Communication Space, Human-Computer Interaction

Our computer systems and corresponding interfaces have come quite a long way in recent years. We no longer patiently punch cards or type obscure and unintelligible commands to interact with our computers. However, out current graphical user interfaces, for all of their advantages, still have shortcomings. It is the purpose of this paper to attempt to deduce these flaws by carefully examining our earliest and most basic formulation of what a computer should be: a workspace.

The History of the Workspace

The modern computerized workspace has its beginning in Vannevar Bush's landmark article, "As We May Think." Bush presented the MEMEX: his vision of an ideal workspace for researchers and scholars that was capable of retrieving and managing information. Bush thought that machines capable of manipulating information could transform the way that humans think. What did Bush's idealized workspace involve?

It consists of a desk, and while it can presumably be operated from a distance, it is primarily the piece of furniture at which he works. On top are slanting translucent screens, on which material can be projected for convenient reading. There is a keyboard and a set of buttons and levers. Otherwise it looks like an ordinary desk. (Bush, 1945, p.107)

Our modern computing systems have grown eerily close to Bush's vision, although they rely principally on semiconductors as a medium instead of the film that Bush thought would be used. Many of the ideas Bush forwarded in "As We May Think," have indeed become parts of our most basic presumptions about what a computer should be: a workspace, a desktop that retrieves and manipulates information. Computing pioneers at the Stanford Research Institute (SRI) and Xerox's Palo Alto Research Center (PARC) adopted many of Bush's ideas as they developed the first incarnations of the modern graphical user interface. As these researchers struggled to develop the first windowing systems, they turned to the theoretical MEMEX and the physical workspaces around them for ideas. In particular, Alan Kay and other research scientists began "to regard the screen as a desk, and each project or piece of a project, as paper on the desk" (Levy, 1994, p.61). This was to be momentous, for the research done at SRI and PARC was used as the basis for the first workstation computers: the Xerox Alto and Star. The Xerox Star system adopted a desktop metaphor and used office articles to help form a familiar conceptual model for its users. Bush's influence is clear; the designers of the Xerox Star system cited Bush as providing a vision of a "desktop computer" and placed the MEMEX atop a family tree of interfaces (Johnson, et al.).

The Xerox Star system, in turn, had a tremendous influence on modern graphical user interface (GUI) design. "It is said that Steve Jobs, Apple's founder and chairman, visited Xerox PARC in 1979 and was incredibly impressed with the bit-mapped displays and applications running on them" (Baeker, et al., 1995, p.50). After seeing a graphical system representing a workspace, Jobs set about implementing a commercially viable system that incorporated a desktop and workspace metaphor; what became Apple's Lisa and Macintosh computers. The success of Apple's Macintosh in turn caused Microsoft and IBM to begin development of their own workspace environments, which later became Windows and OS/2. In this way, the workspace---in its various incarnations---came to be the standard GUI among modern personal computers.

The Ethos of the Workspace

The workspace, however, has many shortcomings. Early computer workspaces were modeled after the real workspaces of the time, desks replete with paper, folders, in-boxes and out-boxes. In the office of the 1960's and 1970's, a "user" interacted with peers through artifacts like files, folders, and letters. Our current interfaces still reflect many of the attributes of traditional workspaces: the emphasis is on communication through artifacts like files, folders, web pages, and email. While there has been some experimentation with anthropomorphic interfaces, most interaction and communication occurs through digital representations of artifacts. To encounter some sort of abstraction for another human being the user must first venture into an artifact, an "Address Book," for instance, and then to the digital artifacts relating to that person. Jo Ann Orvec---in Virtual Individuals, Virtual Groups---describes our current virtual representations of people in the following manner:

A virtual individual is a selection or compilation of various traces, records, imprints, photographs, profiles, and statistical information that pertain (or could reasonably be said to pertain) to an individual---along with writing done, images produced, sounds associated with, and impressions managed by the individual. (1996)

Note Orvec's emphasis on the artifacts "that pertain" to people. In choosing the workspace as an interaction metaphor, we have chosen to abstract electrons into artifacts, as opposed to communicative entities with which we can socially interact. In order to communicate with other individuals or with the system itself, a user must first interact with impersonal objects.

In short, interaction within a workspace is dehumanized; users are immersed in a world of indifferent objects. The computer interface's capacity for polymorphism is one of its strongest advantages, and yet, interfaces have been modeled upon static objects:

"The Macintosh's basic problem is the earthbound metaphor that underlies it---ideas like "file," "folder," "desktop." These ideas are patterned so closely on preexisting physical models that the software is seriously constrained . . ." (Gelernter, 1998, p.64)

Computer interfaces, at the same time, have the capacity to do so much more; they can be designed to provide rich, adapted interaction.

Our workspaces are also generic. While they allow a small degree of personalization like their real world counterparts, at their core workspaces are meant to be neutral in function. Out of this generic conception of interaction arises the myth of the typical user: an abstract representation of the expected population of users. User interface designers are often encouraged by the literature to "know" the typical user that they are designing for:

Much of interface design is built around the premise of "the typical user." As early as Hansen's 1971 dictum "know the user" . . . designer's goals have included learning what typical users needs, what tasks they do and how they do them, and how they react to the software that is implemented. (Baeker, et al., 1995, p.783).

A successful workspace is supposed to provide for the needs of the typical user. I assert that the myth of the typical user has led to poor, depersonalized, one-size-fits-all interfaces. The conclusions of a study conducted to assess the implications of designing for culturally diverse users support this assertion:

While issues in cultural and gender diversity are vitally important to the future of human-computer-interaction, they cannot be addressed by prejudicing the results with overly-generalized characterizations of user populations and vague, unsupported guesswork about what specific user populations find to be an "appealing perceptual experience." (Teasly, et al., p.40)

Our electronic workspaces depend upon "overly-generalized characterizations" of our users. An interface designer forms a model of the typical user, and then constructs an interface for this abstraction. This approach tacitly forces conformity of the user to designer's vision of a typical user. If users drastically deviate form our preconceived notions of what their role should be within the workspace, then they are precluded from effectively using our computers.

The problem is that there is no such thing as a typical user ... Indeed, people's adoption of software is often less a matter of finding a perfect fit with the system than it is of conforming to organizational constraints, finding adequate functionality, and being willing to accept the challenges of learning and using a less than perfect system. (Baeker, et al., 1995, p.783).

This is hardly a symbiotic relationship between human and computer. In the worst situations, the user goes away frustrated and the machine goes unused. Perhaps our quest to design a single workspace that appeals to the greatest possible number of users is a fool's errand.

A Study of Communication Theory

We have decided to call the entire field of control and communication theory, whether in the machine or in the animal, by the name of Cybernetics, which we form from the Greek [for] steersman. -- Norbert Wiener, Cybernetics (1948)

Some individuals involved in forming our conceptions of human-computer interactions observed that people must communicate with their computers. Curiously, communication theorists are rarely cited within the literature of Human-Computer Interaction community. I wish to provide a concise examination of specific communication theories and then assess their implications upon a communications-centric interface. It is clear that for a user to effectively work with a system, that user must first communicate with the system. Moving from this premise I will present a synthesis of communication theory and interface design---a communication space---as a more appropriate model for interaction.

In order to form a cogent understanding of what a communication space might be; we must first attempt to gain an understanding of what communication itself involves. This is, unfortunately, not a simple matter. Several models for communication have been forwarded by thinkers from diverse philosophical schools. In an attempt to provide a concise account of what communication entails, I will examine only a few specific communication theorists.

Roman Jakobson provided one of the first modern theoretical models of what communication involves. Jakobson---a linguist---analyzed communication synchronically in an attempt to ascertain its essential structure. The model (Figure 1) that Jakobson constructed divided communication into the following components:

Figure 1: Jakobson's Analysis of Verbal Communication

While Jakobson's model originally was designed to illustrate "the constitutive factors in any speech event," it was subsequently applied as a framework for analysis of communication in general (Jakobson, 1960). Note the explicit inclusion of the addresser and addressee, in addition to the artifact (the contact) used to convey the message.

Lisa Ede and Andrea Lunsford in their explication of the role of audience within discourse first examined several models for written communication and then developed their own model. Two communication models that they consider closely are Mitchell and Taylor's "general model of writing" and Corbett's model of "The Rhetorical Interrelationships." Mitchell and Taylor's model consists of "Writer," "Written Product," "Audience," "Response," and the process by which one entity feeds into another (Ede and Lunsford, 1984). By comparison, Corbett's model involves "Writer," "Universe," "Audience," and "Message" interconnected to each other (Ede and Lunsford, 1984). Both models, keeping with Jakobson's, contain representations for the sender and receiver of information.

Rhetorical analysis of communication typically goes further to divide the sender into an author and narrator and the receiver into an actual reader and "mock reader." Ede and Lunsford make this rhetorically grounded delineation in their view of the receiver of information as both "Audience Addressed" and "Audience Invoked" (1985). In Coney's analysis of reader-writer relationships, similar divisions are made: the sender is portrayed as an "Author" and "Implied Author" and the receiver is divided into a "Reader" and "Mock Reader" (1984).

Within the literature of Human-Computer Interaction, Terry Winograd's "language/action approach" is one of the few methods rooted in communication theory. Based on the philosophy of language and the work of Austin and Searle, the language/action approach treats the interaction between speaker and listener as a series of actions by both parties (Winograd, 1988). Winograd's approach is most commonly used in support of the design of groupware or Computer Supported Cooperative Work (CSCW) applications.

Figure 2: Communication Theorists Compared
Theorist(s) Sender Receiver
Jakobson Addresser Addressee
Mitchell and Taylor Writer Audience
Corbett Writer Audience
Ede and -Lunsford Writer Personal Audience Addressed Audience Invoked
Coney Author Implied Author Reader Mock Reader
Winograd Speaker Listener

What we see is that each of these communication models agrees that communication involves some sort of sender and receiver (Figure 2) in addition to the artifacts that are used to transmit information. It is my intent in the next section to use this commonality among these different models as the basis for a revised interface paradigm.

The Communication Space

While these communication theories all specifically include senders and receivers of information as basic constituents of any communication, our workspaces rarely include a representation of a sender or receiver of information. How would our workspaces appear if instead of abstracting information into workspaces and files we abstracted information into communicative senders and receivers?

Imagine that instead of seeing a quaint desktop when using a computer that we see a portal into a world of animate senders and receivers of information: avatars for agents, programs, operating systems, distributed systems, and, of course, other humans. Ideally, we could exchange and distribute information with a natural language system, handwriting recognition, or more traditional input systems. When we wish to gain information about a computer, software, or another person, we communicate directly with the representation for that object. Instead of forming inaccurate schemas of what the software designer's model of a system is (and our role within the system), we could ask the software to relate itself to us and perhaps (as with human communication) it could adapt its message to us. As in a conversation, we could be simply introduced to a communicative entity at first, but over time we could gradually be given a more complete picture of exactly what "it" is and what "it" can do. Instead of overwhelming users with a dizzying array of unfamiliar features and commands, a communicative entity could gradually ease a user into a more comfortable relationship with itself.

A communication space is an environment in which all constituents are capable of interacting with represented users and with each other in a manner modeled after human communication. The users themselves must also be represented. A synthesis of a computer-compiled user model and Orvec's conception of a virtual individual (see quotation above) could serve as one possible representation. User models could be extended to include an external representation as well as the artifacts that a user creates while collaborating with other communicative entities (software or human).

One possible objection to this approach is the distasteful, Orwellian aspects of computers collecting sensitive information about their users. There are ways to alleviate this concern though:

User will be even less likely to trust a system that continuously records sensitive information about their interests, tastes, and comings and going. All of that information must be kept private if users are to trust the user modeling system, resulting in a truly personal computer that knows everything about you---including your privacy boundaries---so that it can gate the flow of information between it and the outside would according to your desires. (Orwant, 1996)

Perhaps users could be portrayed in a communication space as public (visible to others) and private selves. Users could browse another person's public appearance or "introspect," that is, view their own private and public representations.

It would be difficult to completely break from the use of artifacts. However, in a communication space artifacts would be made secondary in stature to communicative entities. Artifacts would be presented as the product of collaboration among communicative entities. For instance, this treatise would be presented in my communication space as the product of my interaction with the communicative entity that provided me with word processing facilitates. When another user wishes to view this document, they would view my "public self" and see all of the artifacts that are associated with me. Likewise, when I view the communicative entity with which I collaborated to create this document, I would see an artifact that represents this document. Organization of documents would not necessarily have to be spatial, as it commonly is today; documents could be organized relationally, that is, artifacts would be identified with the entities which created them.

At first, a communication space might seem similar to Microsoft Bob, Magic Cap, or any number of other failed social metaphors. It is my contention that previous social interfaces failed because they did not provide an adequate way to communicate with the other entities, most importantly, other humans:

The other problem with the domesticity of Microsoft's Bob is that the imagined space is a profoundly anti-social one. It conceptualizes the infosphere as a private home sequestered from the outside world. The only contact with other "people" comes in the form of those ridiculous cartoon characters, those agents and info-butlers . . . This might have been reasonable in the old days of stand-alone desktop computers, but in the age of the Internet, using an interface that doesn't offer some vision of public life can seem less like a cutting-edge exploration through information space and more like a visit to Miss Havisham's. (Johnson, p.62, 1997)

However, if care is taken to assess communication, an interface that contains social entities might be more successful. A medium that ingeniously improves our ability to communicate with each other like the written word, telephone, or web seems bound for success.

In summary, if we abstract information into humanized, communicative entities (instead of mere artifacts) we might be able to move closer to the distant goal of human-computer symbiosis. This brief and incomplete vision of a communication space is meant to be a call to arms; our metaphors are faltering and it time to repair them. Computer users should not be expected to undergo continual frustration due to outdated and constrained metaphors, we need to find something new.

References

Baeker, R.M., Grudin, J., Buxton, W.A.S., Greenberg, S. (eds.). (1995). Readings in Human-Computer Interaction: Toward the Year 2000, San Francisco: Morgan Kaufmann Publishing, Inc.

Bush, V. (1945) As We May Think. In: The Atlantic Monthly, 176, 202-208.

Coney, M.B. (1984). The Implied Author in Technical Discourse. In: Journal of Advanced Composition, 35, 163-171.

Ede, L. Lunsford, A. (1984). Audience Addressed/Audience Invoked: The Role of Audience in Composition Theory and Pedagogy. In: College Composition and Communication, 35, p.104-154.

Gelernter, D. (1998). Machine Beauty: Elegance and the Heart of Technology. New York: Basic Books.

Jakobson, R. (1960). Closing Statement: Linguistics and Poetics. In Sebeok, A. (ed.) Style in Language. Cambridge, Mass.: The MIT Press.

Johnson, J., Roberts, T., Verplank, W., Smith, D., Irby, C., Beard, M., and Mackey, K. (1989). The Xerox Star: A Retrospective. In: IEEE Computer, 22(9), 11-29.

Johnson, S. (1997). Interface Culture: How New Technology Transforms the Way We Create and Communicate. San Francisco: HarperSanFrancisco.

Levy, S. (1994). Insanely Great, New York: Penguine..

Orvec, J.A. (1996). Virtual Individuals, Virtual Groups: Human Dimensions of Groupware and Computer Networking. New York: Cambridge University Press.

Orwant, J. (1996). For Want of a Bit the User Was Lost: Cheap User Modeling. In: IBM Systems Journal, 35(3&4), 398-416.

Teasly, B., Leventhal, L. Blumenthal, B., Instone, K., Stone, D. (1994). Cultural Diversity in User Interface Design: Are Intuitions Enough?. SIGCHI Bulletin, 26(1). 36-40.

Winograd, T., (1988). A Language/Action Perspective on the Design of Cooperative Work. In Human-Computer Interaction, 3, 3-30.

Acknowledgments

This article is dedicated to Susan Pearlman, as my discussions with her provided the basis for this article. Innumerable thanks to Mary Coney and David Farkas for reviewing and commenting on the manuscript for this article. Thanks also to Dad, Mom, and Marshall Smith who provided comments on later drafts. Lastly thanks to Roger Squire, Daryl Carlson, and Dan Johnson for listening and commenting upon various ideas associated with this article.

About the Author

Carson Reynolds is a student of Technical Communication and Philosophy at the University of Washington. He is working to implement biologically-based systems for adaptive and evolving interfaces.

Contact Information

Carson Reynolds
Department of Technical Communication
University of Washington
4210 Brooklyn Ave. NE #208
Seattle, WA 98105, USA

Telephone: +1-206-292-2888 x138
Email: creynold@acm.org
http://weber.u.washington.edu/~creynold/

No earlier issue with same topic
Issue
Previous article
Article
SIGCHI Bulletin
Vol.30 No.3, July 1998
Next article
Article
No later issue with same topic
Issue