Five
Styles of Interaction
by
Eric Gould Bear of Monkey Media
Introduction
Numerous kinds of interaction are available to electronic media developers
seeking to engage an audience. Each has a unique feel and communicates
a unique message. Yet despite such variety, the interface elements of
many products fail to complement their content. Conservative development
teams favor tried-and-true interface components borrowed from desktop
computing -- which makes their products feel like productivity applications.
Those willing to step out on the edge have a tendency to use novel techniques
solely to be the first on the block. Both run the risk of mismatching
the message of the content with the style and attitude of the interaction
techniques employed. This pitfall can be avoided and products can be made
to stand out by addressing interface style as an integral part of overall
design.
Interface elements can be divided into five families of interaction -- discrete, continuous, concrete, character and resonant -- based on the involvement they engender between participant and content. (These are broad categories with some overlap, but interface elements falling into more than one family usually weigh most heavily in one.)
Grouping interactions into these five families can prove useful in at least two ways. First it can help select an interface style that reflects the overall attitude we want to convey in particular work. Second, it can help in envisioning alternative styles of interaction when the attitude or form of a piece is yet to be determined. This article looks at each category and how various kinds of interaction affect the participant's relationship to a product.
Discrete
Interaction
Common examples of discrete interface elements are buttons, panels, palettes,
stamps, menus and some cursor-based tools. The interactions in this family
are extremely modal
and work well with media that either comes in fixed chunks, such as pages,
or exists in clearly delineated states, such as playing or stopped.
Discrete
elements exist apart from the media itself and act on it irrespective
of its meaning. Consequently, the participant maintains a distance from
the content, using, for instance, a cursor to act on a device which in
turn acts on the media. In terms of attitude, the participant is invited
to have clear intentions, be logical, and expect results that are implicitly
mapped to their actions. By pressing, clicking, poking, double-clicking,
dragging, scrolling and jumping, users can select, view, search, compare,
document, present and command. The media changes state as a result of
their action. They are in control of the pace. Nothing happens unless
they say so, and when they do say so, they get immediate feedback (if
not the results they're hoping for).
Discrete interactions work well when each button has its own function. Mimicking a physical remote control with a software panel may make for a busy-looking screen, but it has the benefit of communicating clearly individuated functionality.
Continuous
Interaction
Continuous interactions contrast sharply with discrete interactions. Common
interface elements in this family are scrollbars, sliders and animated
cursors for spatial or temporal navigation. They work well with media
that flows, for example 3D panoramas, or is author-paced but elastic,
like a story narrated in person. The family of continuous interactions
can also be considered third-person: the media controllers exist separately
from the media, and the participant respectively maintains a distance
from the media.
In continuous interactions, participants usually use a cursor to act on a device that acts on the media, or use a cursor that, when activated, directly manipulates the media. Because their motions map directly and fluidly to changes in the content, continuous controllers offer more opportunity for emotional and cognitive connection than discrete interactions.
In this style of interaction, the attitude of the participant is rhythmic and tends toward casual or experimental. Participants may not know exactly where they are going or what is happening, but they are able to explore safely and return smoothly to their previous location. They are in constant relationship with the media. By doing "nothing", they exercise as much control over the pace as by moving the input device. By rolling, gliding, moving, pressing and holding, drawing and flicking, participants can browse, explore, visualize, simulate and perform. In fact, the pace of travel through the content or space is primarily dictated by the author, but is flexible enough to be intimately influenced by the participant.
Concrete
Interaction
Concrete interactions provide a first-person perspective. The person becomes
an actor who reaches into the content space with an extension of his or
her body, such as a hand cursor,
to manipulate content as though it were physical. Other common examples
of concrete appendages include tools such as paint brushes, erasers and
smudgy fingers.
The interactions are metaphorically material, physical or corporeal. Animating them not only adds a nice touch but reinforces the physical metaphor. (Don't be misled by the hand cursor in today's web browsers. It's just a pointer cursor in disguise because it doesn't necessarily allow concrete interaction.)
To create concrete interactions, the media must be objectified and directly manipulable. Either it is malleable, like a paint-type picture, or it affords interaction by being virtually tangible. It may be an object in the content space, such as a switch, a thing or a container for things. Or it may be part of an object: a handle, knob, tab, dog-ear, corner, edge, surface or hole. When offering concrete interactions to an audience, it helps a great deal to let them know what they can grab. In Real Time Knowledge Systems' market positioning tool CRUSH, grippy affordances (like the little rubber bumps on Braun electric razors) were placed on everything that could be grabbed and moved [SEE figure 3]. The cursor turns into a hand when the user rolls the cursor in proximity of the grippies.
As in discrete engagements, concrete interface elements invite the participant to approach the content with clear intentions, but act on them in an explicit and kinesthetic fashion. Again, participants are in complete control of the pace. By touching, pulling, pushing, panning, tearing, stretching, smudging, grabbing, dropping and throwing, they can construct, manipulate, organize and design. Concrete interactions can be extremely satisfying because the actors affect their world directly.
Although the MacOS Finder uses a standard pointer cursor, it embodies one of the earliest popular uses of a concrete interface with its drag-and-drop interactions for arranging file and folder icons. With smarter software, one can support richer experiences via proximity sensitivity, edge and surface detection and simulations of friction and gravity.
Character
Interaction
In character interactions the participant is a player experiencing the
world through, or with, a character who lives in the content space, often
by manipulating the actions of that character. Thus, character interactions
are literally second-person. This style is often aimed towards children,
in which case it relies on simple actions carried out with convincing
animation. We may find, though, that a combination of concrete and character
interactions proves useful in virtual spaces for all ages. The content
space in which character interactions take place -- which consists of
a storyline, other characters, objects and environments -- makes interactions
of this type inherently more variable than those of other families. Unlike
the discrete family, in which a given interface element has the same effect
regardless of the meaning of the content it manipulates, character interactions
vary dynamically with the character's situation. The specific attributes
of characters and the things they meet determine the nature of their interactions.
For example, the same D-pad controls for drawing a sword on an opponent
may bring forth a flower in an encounter with a friend. Further, neither
the sword nor the flower could be revealed had the character not previously
picked them up.
Character interactions work well when the material is energetic, active and relational, enabling the participant to become involved personally and intimately. Not only do children learn more readily in educational titles that employ character interactions, but they characterize their experiences as first-person rather than second. By running, walking, jumping, sitting, standing, following, listening, teaching, picking up, carrying and using, people are encouraged to play, explore, learn and express themselves. A storyline supports them in developing their character and playing out specific roles, possibly fighting, negotiating or collaborating with other players.
Resonant
Interaction
This family is algorithmic and can be called resonant. An early example
of this style was Laurie Spiegel's musical improvisation application,
Music Mouse, marketed during the mid-1980's. In Music Mouse, the player
performs by rolling the mouse around on their desk, affecting a limited
set of variables in the music. Another fun example is David Zicarelli's
OvalTune, from the same era, in which the performer creates graphics and
music simultaneously.
We've seen how discrete and continuous interactions maintain a distinct separation between interface and content. Likewise, in the concrete family, interactions and content can be isolated from one another. With character interactions the possibilities of engagement are tied to larger, more dynamic issues of theme and history. But in all these cases, the specific interaction techniques can be added after the media is created. In resonant interactions, on the other hand, the technique of interaction is the framework for the whole experience. The interaction scheme must presuppose the production of both the media and the information architecture that holds it.
In resonant interactions, the content changes with the participant's movement or stillness (i.e., the amount of time spent in a particular locale). And manipulations of the media aren't necessarily associated with the visible interface elements. So if there is no visible cursor or control device, the author may choose to supply additional feedback in the form of audio, an animated status area or even text. These interactions are inherently more complex than those in the continuous family because the media changes through time.
Reversible flow -- the ability to return to the way things were -- isn't required. Pacing is determined by a balance between author and reader, who is invited to be perceptive, mindful and acting on intuition. By listening, attending, waiting, rolling, pressing, expressing and capturing, they may read, watch, hear, experience, express, feel and share. Author and participant are both acting and communicating through the form of the content itself.
Eric
Gould Bear, whom Communication Arts recently described as "one of
the
most thoughtful, provocative interface and interaction designers working
in the
field", launched MONKEYmedia in San Francisco in 1994 to provide
the
desktop computing, new media and consumer electronics industries a design
and technology resource focused specifically on human-computer interaction.
They are now based in Austin TX.
