Design Tidbits

back To Overview of Tidbits

Design Tools

Simplification

Interaction Design

Screen Design

Hierarchies

Visual Design

Design Process

Web, Web Applications, Web Design

Performance

Miscellany

 

 

A Layman's View on Interaction Styles – Part II

By Gerd Waloszek, SAP User Experience, SAP AG – December 2, 2003

In the first article of a series on interaction styles, I presented "a layman's view" on older forms of human-computer interaction (or communication). I also introduced and criticized graphical user interfaces. In this second article on interaction styles, I will sketch a selection of interaction styles for graphical user interfaces, again offering a "layman's view" and using analogies from real life. I will also move forward to some future interaction styles, which are still in their infancy.

 

Interaction Styles for Graphical User Interfaces

"Heap" – (Unattended) Workshop

The "heap" model is, in some respects, the "typical example" for graphical user interfaces. In its essence, it goes like this: There is a room (called "desktop" or "window") where the users have to do something. Someone else (the computer) has piled up quite a number of tools in one or more corners of the room. From the equipment littered around the room, users have to pull out the tools, inspect them, check whether they are applicable, and try them out before they can start working. In other corners of the room, there are raw materials (or just a white canvas), on which users are supposed to use the tools. Nobody tells them how they have to do this. Surely, this is a scenario for daring people who like to experiment a lot. By the way, the computer is a very discreet communication partner in this model – it stays in the background, and waits for users' actions (technically, this is called the main event loop). So, whereas the computer was typically the active and driving force in the man-computer dialog in earlier models, here it adopts a mostly passive role. Maybe, you will find this model somewhat exaggerated – but it highlights some of the issues of this interaction style.

Variant: Drawers and Toolboxes – Tidying Up the Workshop

Not all workshops look that untidy. In a painting workshop called "Dabbler," for example, the tools are neatly arranged in "real" drawers. Many applications arrange their tools (hidden) in menus and a variety of toolbars. This is more or less comparable to tidy drawers and toolboxes that are distributed around the room. Most of these workshops, however, still lack directions for how to proceed. (I do not want to go into help systems, which, in this model, are in "other" rooms.)

Wizard, Assistant – Or, Stepping through an Office

I still do not know why the next interaction style that I want to discuss has been named "wizard." I expect a wizard to fulfill my wishes as soon as I snap my fingers. But wizards and assistants on computers do nothing like that.

The wizards that I encounter on computers behave more like a sort of assembly line – but not like a car assembly line, where one partial job is done at one station; it is a line where you go from assembly station to assembly station to get your task done. Perhaps, an office is a better model for this interaction style. In an office you go from counter to counter (sometimes to the wrong one) to get your form filled out (and in the end, for example, get a passport or driver's license). On the Web, such a form could be an order form in an online shop.

There is little freedom for the user in this model; everything runs "as usual" in a regulated and predefined manner (with some predefined variants). That is why I think that the office model is so appropriate.

 

Future Interaction Styles

At the Dentist – Adaptive Systems

Dentists have their patients lying in the dentist's chair before them, on which they also find the most important tools for treatment. Whenever they need something, they just say "swab," for example, and instantaneously the assiduous dental nurse hands over what they need. In this model, the tools approach the users as needed, while the users can focus on their tasks and have the necessary materials at hand. There is one small problem with this model, though: The user, that is, the dentist, is an expert and knows how to proceed. How can we model a casual user or a beginner? Imagine that there is a supervisor dentist in the room, a sort of prompter, who whispers to the dentist what to do next (and how if needed). Now we have a "dentist-with-prompter model."

This model resembles adaptive systems, that is, systems that adapt to users' ways of doing things and offer only what users need in the current context. Adaptive systems have to model the user, the task, and the current application context in order to be useful and effective. That is not an easy task for computers and calls for "artificial intelligence" methods.

Helpful Souls (Brownies) – Agents

In the fairy tale about brownies an old dream of mankind comes true: Helpful souls do all the humans' work over night. Actually, computers have done such things for a long time. Think of the batch jobs that run over night and, for example, update data. These days, such nightly updates are often no longer sufficient because computer systems for global companies have to be available 24 hours a day. Another problem with this approach is that batch jobs have to be precisely prepared in order to run correctly.

For a couple of years now, there has been a new interaction model that neither cares for day nor night and consists of small applications that – after receiving some initial instructions – can act on their own. These applications are called "agents," sometimes even "intelligent" agents. The latter have the ability to adapt and learn what their users (or "masters") need – at least to some degree. Agents are controversial in the UI community because some people feel deprived of control by them. Typically, agents fulfill small tasks. Some agents do not even have a user interface of their own (at least for the actual job). They work in secret, just as the brownies do.

Talking "with Your Hands" – And More ...

If you watch people using cellular phones, you will find that they often talk "with their hands." The problem is, however, that their partners do not see these gestures; neither do they see the mimics nor posture. While this behavior shows that the confrontation of humans with technology can lead to some interesting slips, we nonetheless know that nonverbal cues play an essential role in human communication.

Interaction styles that include gestures or other nonverbal elements are still in their infancy, apart from some more simple gesture-based approaches in handheld computers. I can recognize two research directions: The first focuses on gestures, the second on the recognition and synthesis of emotions as well as physiological states. While it is fairly easy to think of using gesture-based commands for mobile business scenarios, there are far less conclusive ideas on how emotions can be included in business software. Maybe, the detection of interest or exhaustion would be interesting paths to pursue. By the way, the gestures currently used on computers are simple commands; they do not convey any nonverbal information as is the case in human communication.

At Issue – From Human to "Human" (Avatar)

We are also still far away from realizing the next interaction model: humans and computers leading a human-like conversation in order to fulfill a task. Of course, this model is only usable for a subset of tasks that can be carried out with computers. Many tasks are better automated and carried out by the computer alone. Possible candidates for a conversational interaction style are planning and decision processes.

There are already some approaches to realizing this model being developed but these are still in the research stage: Narrative systems, often in conjunction with artificial characters (avatars), are being developed and tested. These systems have, however, a long way to go until they can be used on a broader basis.

 

The Ultimate Vision: Symbiosis – Humans and Computer "Integrate"

Even the most futuristic interaction styles presented here are based on a traditional view of computers: Computers are conceived as boxes standing on a desktop, resting on our laps, or being held in our hands. Visionaries, such as Mark Weiser or Don Norman, have claimed for years that computers will "disappear" or become "invisible," and many others have joined in. Further buzz words for this trend are "the ubiquitous computer" and "pervasive computing." According to this vision, computers will penetrate our environment and become omnipresent. Think of smart dust and similar circuitry that will be built into appliances, rooms, and even clothing. But technology will not stop here. Researchers like Kevin Warwick and artists like Stelarc have been experimenting with implanted electronic circuits for some years now. Their vision is that one day we will communicate with computers simply by exchanging nervous signals originating from movements, thoughts, and emotions. Today, it is not fully understood how this can be done effectively. The brain waves that we detect are overlays of a multitude of signals. But once the correct signals are discovered and isolated, we will be there – technically it is possible today.

This last vision is comparable to symbiosis as found it in many creatures. Think of lichens that are composed of fungi and algae and form a densely knit community that appears as one creature to the outside world. In other forms of symbiosis, the participating creatures keep their individuality. As in nature, the visions outlined above offer different degrees of integration of humans and computers. They add a new meaning to the often-cited "augmentation" (contrary to automation) of humans through computers.

 

References

There are a number of articles on narrative interfaces, emotions and agents on the SAP Design Guild Website. In these articles, you will find links to further information.

Narrative Interfaces

Emotions

Agents

Other information on the Web:

Books

Norman, Donald A. (1998). The Invisible Computer. The MIT Press. ISBN: 0262140659 (Hardcover)

Norman, Donald A. (1999). The Invisible Computer. The MIT Press. ISBN: 0262640414 (Paperback)

Shambhu Upadhyaya (Ed.), Abhijit Chaudhury (Ed.), Kevin Kwiat (Ed.), Mark Weiser (Ed.) (2000). Mobile Computing: Implementing Pervasive Information and Communications Technologies. Kluwer Academic Publishers. ISBN: 140207137X

 

top top