Tangible interfaces

In summer, I wrote about a framework for reality-based interaction for which I wanted to give a few examples. Although the term reality-based interaction is not mentioned, I think the design ideas presented in two recent publications of the MIT Media Lab team around Hiroshi Ishii fit very well into that category.

The first paper is Simplicity in Interaction Design by Chang et al. which reports on a design exercise conducted at the Media Lab to encourage students to design expressive but simple means for representing information of common devices found in households with the use of a very limited set of interface components. This exercise resulted out of the observation that many interfaces are overloaded with “buttons and blinking lights (B.A.B.L.)” which represent information about complex state machines, such as those found in modern answering machines, in a manner that is too complex for users to understand. The authors argue that this is a result of designers chosing features over usability which, as a consequence, requires ever more complex user interfaces. The primary constraint of the exercise to redesign such a device was therefore set to allow for a maximum of one input and output mechanism, e.g. one button and one LED. With this, the authors state, they hoped to force students to prioritize features and leave out those not representing a core functionality. Before conducting the exercise in a classroom setting, the authors completed the exercise once themselves.

Of the results described in the paper, I found the design of a simple answering machine particularly inspiring: It consists of a bowl-shaped base which is overstretched by a membrane. For every incoming call recorded on the device, the volume covered by the membrane expands slightly more, causing the membrane to bulge outwards. To play back the recorded calls, a user simply applies pressure to push the membrane back down into the bowl. To rewind a recording for a few seconds, it is sufficient to briefly pull the membrane back out. In essence, Chang et al. argue, these restrictions helped them to focus largely on using mechanical change to modify and visualize a device’s state and overall, they conclude that such forced simplicity can in fact encourage finding novel interaction techniques and foster new innovations. Personally, I see significant potential in the marketability of such simple devices and would be happy to see them showing up in stores. It is the instant comprehensibility of the interactions which, in my opinion, would find a big audience of buyers.

No messages, a few messages, several messages, playback of messages

Different states of a tangible answering machine. From left to right: No messages, a few messages, many messages, playback of messages

Continue reading


Reality-based interaction

One of the sessions I attended at this year’s CHI conference was called Post-WIMP (WIMP = Window, Icon, Menu, Pointer). It covered recent trends in user interface design and highlighted attempts to move beyond the typical WIMP concepts we generally use today. In one of the talks, the paper Reality-based interaction: a framework for post-WIMP interfaces by Robert J.K. Jacob et al. was presented which I found particularly interesting. Specifically, the authors developed a framework which allows for the comparison of today’s emerging post-WIMP interfaces according to several aspects, based on the underlying notion of a foundation of these interfaces in the physical, real world.

For the authors, up to four themes of the physical world can be reflected in these new user interfaces:

  • Na├»ve Physics
  • Body Awareness & Skills
  • Environment Awareness & Skills
  • Social Awareness & Skills


While Jacob et al. argue that building user interfaces around these innate human skills can have several positive benefits, such as reduced cognitive workload, they also stress that designers should ideally deviate from these concepts in explicit cases. An example they give is that of walking in a virtual world. The command for walking could then be augmented by a command for flying which should remain as analogous as possible to its realistic sibling. Unless some additional power (such as flying) was gained by a command, the authors state, a designer should always use the command closest to reality.

This rationale has become a central criteria for me when judging novel reality-based interfaces, many of which have been promoted (and hyped) in online videos over the last months. I will soon post a few examples of reality-based interfaces which I have recently discovered and put them into relation with the four themes above.