Design Tidbits

back Tidbits Overview

Tools for Design and Visualization


Interaction Design

Screen Design


Visual Design

Design Process

Web, Web Applications, Web Design





A Layman's View on Interaction Styles – Part I

By Gerd Waloszek, SAP User Experience, SAP AG – November 11, 2003

In a short series of two articles, I will describe ways in which humans interact with computers (read follow-up article). For this purpose, I will adopt a "layman's view" and use analogies taken from real life, hoping that these analogies make the characteristics of the interaction styles more evident. Professionals would call these ways "interaction styles" or "paradigms." I will also use the term "model" because analogies can serve as mental models for the interaction styles. The terms "communication" or "dialog" are also used for the interaction between humans and computers. Therefore, it can be useful at times to also speak of "dialog models" or "communication models."


Early Interaction Styles

Parcel Mail

I have been dealing with computers for a long time. Our first dialogs were based on five channel tapes. This was like talking at breakneck speed. So I quickly learned to find the line ends (carriage return linefeeds) in the tapes to have at least some orientation in those "tapeworms." Using punchcards meant a major step forward for me. One card typically contained one computer instruction (in FORTRAN) that was also printed on the upper border of the card. I delivered my programs at the computer center as packs of cards in a plastic box. There, I pushed the box through a counter window onto a conveyor belt. The belt moved the box to the card reader, which finally read them into the computer. The next day, I found a pack of paper with the program results somewhere on a huge shelf.

This kind of human-machine communication, whether using tapes or cards, has a lot in common with sending parcels by mail: I send a pack of cards to the computer, and the computer sends me a pack of paper in return (plus the box with the cards – at some point any analogy will break down...). This kind of interaction was on the one hand quite leisurely but also sluggish and inflexible on the other – not well suited to rapid prototyping... It also was very "basic" with respect to communication.

Chat (or Intercom)

Next, an interaction style followed that started with the use of teletypes and was carried over to the first computer terminals: The user and the computer exchanged text messages in order to issue directives and to keep each other up-to-date. Both sides often used cryptic tokens, that is, a secret language which ordinary people would not understand, in order to make the dialog more elitist. This sort of dialog reminds me of the messages that people exchange in chat rooms these days. There, people also type in messages to each other, often containing terms and abbreviations that only insiders understand. Again, this is a very restricted communication style. It is based solely on text strings with no resort to visual, contextual, or any other additional information. Initially, I also considered the intercom or walkie-talkie analogy, but this analogy is wrong: When people are talking to each other, their voices carry a large amount of contextual information, which helps them to interpret the messages; this is lacking in written messages.

Card Games, Questionnaire

There is a later variant of this interaction style in which programs require users to enter certain data before they set to work. In the times of operating systems, such as CP/M and DOS, most programs followed this pattern. And a lot of today's business applications are still somewhat reminiscent of this style.

When thinking of a model for this interaction style, card games come to mind – for example, the game of patience. There is a deck of cards in front of the user, who turns over one card after the other and acts according to the card. You could also think of a questionnaire that is read aloud by an interviewer. There are also questionnaires that you have to fill out yourself. This does, however, not make a big difference, apart from the fact that you can look a few questions ahead,which is similar to computer-based forms in more recent applications.


And Today?

Today, the dialog cycles have shrunk from days to fractions of a second. The amount of information that is exchanged between computers and humans has risen immensely and comprises much more than pure text messages. Modern graphical user interfaces (GUIs) create a virtual world (or representation) in which users communicate and interact with computers. Below are some buzz-words that are associated with graphical user interfaces:

  • Direct manipulation: Users interact with objects on the screen, that is, with virtual objects, as if these were "real" physical objects. As an example, a user may move an icon (a graphical symbol on the screen), which stands for a file, to another icon, which stands for a folder, in order to move a file to a certain directory; this is done using the mouse, commanding a mouse pointer on the screen. For most users, this equivalence is so realistic that they do not make any difference between the virtual (the icons) and the real objects (the files, directories, and so on) that they represent.
  • WYSIWYG: The computer provides users with a reliable preview of what they will get as the final result. For example, users are given a preview of a word processing document showing how the document will look when printed.
  • Metaphors: Computers create an environment that is familiar to users, allowing them to transfer their knowledge from the real to the virtual world and to act in both worlds in the same way.

That at least is the theory. In practice, these principles do not always work ideally. Much research has been done and a lot of papers written about these problems – I do not want to add more to this.

When graphical user interfaces were new and exciting, companies like Apple Computers conceptualized an image of computer users that was way too optimistic. They propagandized the curious, exploring, daring, and emancipated user, an ideal that had probably been inspired by U.S. college students. For sure, no one would like to constrain the freedom of and the possibilities for such users. Think of the motto "Don't mode me in" that banned modes and all modal interactions. (That is, they demanded that users were not restricted in a certain context to one single, possibly pre-thought action.)

These ideas started the boon but also the curse and distress that are connected with graphical user interfaces. As these interfaces were declared intuitive "by definition" and easy to learn and use, the masses started to buy computers and turned into "users." But quite a number of them failed and still fail because some of the premises in this "formula" seem to be wrong...

As a general interaction model, I offer the "real world" analogy for graphical user interfaces. In the next article, I will, however, complement this model with more specific and "realistic" analogies.


When you have been using a computer for a long time, you easily forget that interaction with computers is based on a huge amount of conventions to which users have to conform – just as in a Japanese greeting ceremony, including the correct exchange of business cards. These conventions are by no means as intuitive as is often purported. When I unpacked my first Macintosh computer and wanted to start working with it, I did not know that I had to double-click a program icon in order to start a program – I had to consult the manual (and deferred my encounter with the Macintosh to the next day).

Many of the interaction conventions in graphical user interfaces are simple, much simpler than the cryptic commands of earlier command-line interfaces. But intuitive and simple are two different things. Simplicity helps us to remember conventions until they become our second nature. Simplicity alone, however, does not guarantee that we can remember conventions. Problems arise whenever there are many similar conventions, as is, for example, true for keyboard shortcuts. Like most users, I can recall only a couple of frequently used shortcuts, most of them are offered system-wide. Usually, I make the effort to learn specific shortcuts only if I have to perform a task with high frequency and if mouse commands would be too cumbersome.


What's Next?

In the follow-up article, I will sketch a selection of interaction styles for graphical user interfaces. I will also move forward to some future interaction styles that are still in their infancy.


top top