Paper Reading #12: Disappearing Mobile Devices

Comments
Cindy Skach
Evin Schuchardt

Reference Information
Title: Disappearing Mobile Devices

Authors: Tao Ni Hasso Plattner Institute & Virginia Tech, Potsdam, Germany

              Patrick Baudisch Hasso Plattner Institute, Potsdam, Germany

Presentation Venue: UIST 2009: 22nd annual ACM  symposium on User interface software and technology;
Date: 2009;
Location: New York, NY, USA

Summary

In this paper, the authors try to find out the technology that would suit implementing invisible devices. The challenge is in implementing a solution small enough that it "blends" with the user, but at the same time having functionality. They call the process of making devices smaller and smaller "miniaturization". Miniaturization has progressed from notebook computers and PDAs to increasingly smaller devices.

What hinders miniaturization, say the authors, is the device's user interface hardware, the size of which is linked to human constraints. Screens have to be large enough to be seen, keyboards large enough to be typed on. So one solution is to make non-user interface hardware. This requires removing all interface hardware that is linked to the user's finger size, eyesight, or other human constraints.

The main focus in this paper is on methods of input.

Some of the technologies the authors have considered are:

• Touch: a device can tell whether it is being touched or not with the aid of capacitive sensors and threshold light sensors.
• Pressure: humans are able to operate buttons of about 1 mm in diagonal.
• Motion: not a popular solution because when devices are scaled down, motion sensors don't always continue to work, and may require the inclusion of a camera

Because of the limitation of input by pressure, the authors have decided to focus on touch and motion. 

They go on to describe three classes of devices: 

• a motion scanner: consisting in a touch screen combined with with a motion sensor. 
• a touch scanner: consists of only a single touch sensor.
• a direction scanner: consists of three collocated touch sensors. 

In addition to these hardware considerations, the input language also has to be considered. The authors explain that one of the input language of touch scanners is Morse code. The input language for direction scanners is Marking. Marking consists in setting up a timeout as a delimiter to gestures. Finally, the input to motion scanners is unistroke. Motion scanners have the task of recognizing the path of a gesture, and should be able to distinguish gestures such as dialing, or entering a character.

The authors then go on to talk about some of the prototypes they have made, and user studies they have conducted. The first user study they did was on the "marking" method of input. In it, they asked users to enter marks by swiping with their index finger. They could be asked to swipe "up", "up-right", "right", etc.. . 
The authors found that users mainly made two types of errors. The first is that users would occasionally rotate the device as they were holding it. The second would occur when participants would stay in the range of the sensor after the gesture was presumably complete. 

 In their second user study, they tested the "unistroke" method of input. This investigated text entry on motion scanners. They found that users would make a lot of mistakes on this type of input method. 

These case studies have allowed the authors to layout a list of things to consider when designing an input method:

• use the entire hand for input
• preferably use marks that do not suffer from the incomplete lift-off problem
• preferably use unistroke gestures
• if orientation matters, provide a frame of reference.
• provide a tactile feature. 
• etc..

Discussion

This paper was interesting, and provided good insight into techniques that could possibly be used for input on invisible devices. The authors also alluded to methods that could be used for output such as visual, tactile, and audio feedback . I personnally don't like the idea of having a device implanted under my skin just for convenience. However, I do find some the methods of input discussed quite interesting.