Interaction and the Interface
by Kevin Fox


If I had to choose the most valuable lesson in this seminar it's the difference between interaction and interface.

In most places I've worked, the terms "Interface Design", "Interaction Design", "User Experience Design", and "UI Design" have been used interchangeably. The differences were usually of little semantic consequence, because none of these workplaces utilized more than one of these roles. In practice, the actual task that the Interface/Interaction/Experience Designer was fulfilling was determined more by their scope of vision, and less by the arbitrary job title they themselves didn't fully understand the implications of. Coworkers outside of the discipline fared no better.

It's taken me two months of on and off thought to really understand the difference between Interaction, Interface, and Experience Design and another month to figure out why these differences were important. In truth, unless others interpret these words the same way that I do, then my interpretation is irrelevant in practice. Taxonomy is a democratic science.

Nevertheless, I think that the understanding of the differences will help me be a better designer (interface, interaction, and experience). To not get ahead of myself, I should talk a bit about my conclusions, and what they will mean to me as I face my next design problem.

I began, in true geek fashion, by logically deconstructing the phrase "Interaction Design", to see if elicited any insights.

Starting with "Interaction Design" I switched the terms around to:

The "Design" of "Interactions"

"Interactions" seemed a little vague, so I reduced it to the 'nouning' of the verb "Interact"

The "Design" of Noun("Interact")

Next I needed to find another way to express "Interact" and looking at a few dictionaries and my own experiences, I came up with "Act upon that which acts upon you."

The "Design" of Noun("Act upon that which acts upon you")

This is starting to look a little silly, but let's take out the Noun() and change it to "The union of acts performed by a system upon another system that acts upon the first system." This leaves us with:

The "Design" of "the union of acts performed by a system upon another system that acts upon the first system."

Though the sentence is ungainly, I think we're getting somewhere. It's about feedback loops. Not just about the loops themselves, but about the design of these loops, the balancing, the architecting of the very systems that regulate how things interoperate with other things.

In short, it's about the designing of interactions, or "Interaction Design."

"Great," the reader may be thinking. "I see that Kevin's grasped the art of tautology."

Well, yes and no. While the substitution, expansion and distillation came full-circle, it helped me realize the areas that the path didn't venture near. Specifically, it doesn't talk about interfaces.

Consider nerve cells. They rely heavily on interactions with their neighbor cells in order to fulfill their functions, and their interfaces play a key role, but the two aren't the same thing. "Interaction" implies the entire system, the gaussian surface enclosing both the transmissive cell and the receptive cell, as well as the synaptic gap between them, the re-uptake inhibitors, neurotransmitters, receptors, the whole deal. Taken together, they represent a feedback mechanism where each part affects and is affected by the other part. Taken as a unit, they represent an interactive model.

Dividing the model into component parts, different roles emerge. As long as the model consists of objects that are interpreted as separate and distinct (in this case, the two neural cells), each object will have an interface by which it affects and is affected by outside forces. The transmitting cell has neurotransmitters, ready for release as soon as the electrical impulse comes up its axon, and the receptive cell has a tree of dendrites, waiting for the electrical imbalance that the neurotransmitters will induce as they attach to a dendrite's protein receptors. Split apart, each of these 'surfaces of communication' are interfaces.

Interfaces abound. Found in the lungs, at the ATM, in racks at the department store, in computer peripherals, pencils, milk cartons, and doors, interfaces are the enablers.

But enablers for what? Interactions. To build off a joke already tired in the Design department, when a professor asks a student to design a new keyboard, they turn the student into an Interface Designer. The moment the student asks, "Why does it have to be a keyboard?" they become an Interaction Designer. Interaction implies an increased scope. A caring about the system that includes the user, the system, and all external factors that impact on the way the two interact with each other.

Most of the time when we design 'interactions' we don't have the power to design all the interfaces contained within that interaction. Understanding and accommodating for this fact is probably the single most important lesson that an Interaction Designer must learn. Take, for example, the design of a web site. The 'front-end interface' of the site is a simple enough interface to understand, but that's just half of the system. The interface of the person actually using the site, how they physically access the site, their motivations, goals, interests and emotions, all play vital roles in the interactive system, and all too often designers create unjustified scenarios in their heads, of idealized users with motivations and interests that match perfectly to the designed product interface. This is the primary danger of designing by intuition or designing by persona.

The counter to this problem is to fully understand the 'user interfaces' (meaning, in this case, the interface of the user, not the interface for the user) of the people who will be part of the interaction. Sometimes there are ways that a user's own interface can be bent by framing the product interfaces in different contexts, using metaphor or visual design to 'tell the user how to think'. More often though, the Interaction Designer needs to work from the pre-formed interface of the user, and the overall goal of the interaction, and determine what, exactly the missing piece is, and design the interface of that piece to suit.

An example of how the user's interface can be changed comes from 'Graffiti', the PDA text recognition system. Graffiti first came out as a third-party product for the Newton MessagePad. the MessagePad had a built-in handwriting recognizer, which was widely panned for its frequent, often humorous, misinterpretations. Less than a year after the Newton's release, ParaGraph came out with Graffiti, which offered near perfect recognition, at the cost of learning a modified character alphabet and writing letter-by-letter in a recognizer box. It was instantly ridiculed and dismissed by users and the press.

Given the Newton's built-in ability to recognize natural handwriting, albeit badly, users thought it was preposterous that a recognition system would require the user to learn how to use it. The product did badly in stores, despite performing very well at its intended task. Years later, when ParaGraph created the Pilot (later sold to U.S. Robotics, which was bought by 3Com, which spun off Palm) they created a simple PDA, built from the ground up around the Graffiti character recognition system. The physical recognition area made it clear that this was the way to enter text, though an on-screen keyboard could also be used.

From the moment the original Pilot 1000 was released, it met with rave reviews. The accuracy of the recognition system was acclaimed, and the minority who said it would be a chore to learn were unruffled by converts who assured then that it only took a few hours, and was really easy.

In this situation, ParaGraph's Interaction Designers succeeded in the rare and difficult task of designing the user's interface, while changing the object's interface only slightly. they presented the interface as a requirement, not an option, and overcame the preconceptions of the media and the mainstream user.

This is the true challenge of the Interaction Designer. Intense contemplation of the existing system and behavior, and the ideal behavior must inform the consequent system, including the way that the entire interaction will be framed to the user. By the time the interface comes in to play, much of the game has already been won or lost. There are clear examples of a great user interface saving an otherwise questionable product, but there are far more cases of innovative and intelligent interfaces that never saw the light of day, or died premature deaths, because their place in the greater system was not considered.

The Newton's cautionary tale, and the success of the Palm Pilot illustrate what I perceive to be the nature of Interaction Design: a discipline that begins in strategy, marketing, and judgement and decision-making psychology, before settling down into the nuts and bolts of the actual interface of the product.