CHI 97 Electronic Publications: Late-Breaking/Short Talks
The Strategy for Selecting a Minute Target and the Minute Maximum Value on a Pen-based Computer
Xiangshi Ren and Shinji Moriya
Department of Information and Communication Engineering
Tokyo Denki University
2-2 Kanda-Nishikicho, Chiyoda-ku
Tokyo 101 JAPAN
+81-3-5280-3335
{ren, moriya}@c.dendai.ac.jp
ABSTRACT
This study deals with the
relations between target-pointing strategies and target sizes.
An evaluation experiment was performed in which the experimental
system changed each of five kinds of targets (1, 3, 5, 7 and 9
dots in diameter respectively, 0.36 mm per dot) and eight directions
of pen-movement, while using each of six kinds of strategies of
approaching the target on a pen-based computer. Two results were
obtained: (1) The "Land-on2" strategy was found to be
the best strategy for selecting a minute target among the six
strategies, in terms of error rates, selection time and subjective
evaluation. (2) This study also clarified a boundary value of
target which controlled difficulty of strategy. When a target
is less than 5 dots (1.80 mm), it is necessary to pay attention
to the determination of the strategy in the software design.
Keywords
Pen input, Target-pointing
strategies, Minute targets, The minute maximum value.
© 1997 Copyright on this material is held by the authors.
INTRODUCTION
Computers operated only with
pens are attracting attention in many fields. The necessity for
selecting a minute target is supposed to rise, while targets (menus
and etc.) are getting smaller as the amount of information on
the screen increases in the miniaturized portable pen-input systems.
Up to now, there are a few
leading studies about the selection of a small target [1,2]. Sears
and Shneiderman (1991) focused on testing three selection devices;
a touch screen, a stabilization-touchscreen and mouse. The task
was the selection of rectangular targets 1, 4, 16 and 32 pixels
per side. Their result showed that a stabilized touchscreen was
effective for reducing the error rates when selecting small targets.
Moreover, McClintock and Hoiem (1993) conducted a study to determine
the smallest optimal size for targets in a pen-based system. Eight
kinds of targets were used but the study did not address any pen-pointing
strategies. Still some findings concerning the strategies have
been described (see, for example, [3,4]).
However, the target-pointing
strategies are an imitation of the selection technique for the
mouse and the touch-screen. Also the investigations have not been
conducted so far which focused on the of the strategies suitable
for selecting a minute target and the size of a minute target
in pen-input systems. Base upon our previous experiment [6], this
study focuses on to the relations between target-pointing strategies
and target sizes.
METHOD
The six strategies of the experiment
The six strategies used in
the experiment were as follows:
- Land-on1: the pen approaches
from above. The target is selected only momentarily at the time
of the pen contact with the screen while the pen is within the
target.
- Land-on2: besides the
Land-on1 condition, the target is also selected when the pen touches
the target for the first time after landing on the screen surface.
- Take-off1: the target
is selected only at the moment when the pen is taken off the target
on the screen. The target is highlighted only while the pen touches
the target.
- Take-off2: besides the
Take-off1 condition, the target is also selected according to
whether the pen takes off inside or outside of the target after
contacting the target on the screen. However, the target is highlighted
only while the pen contacts the target.
In the case of the electromagnet
tablet, when the pen is within a given height (1cm) above the
tablet-screen surface, the coordinates (x, y) of the pen-tip were
able to input these data. Therefore, we made following strategies:
- Space1: the pen approaches
from above. The target is selected only when the pen is within
the target momentarily, at the time of contact for the first time
on the screen. However, the target is highlighted only while the
pen is within the 1 cm high cylinder above the target before the
pen touches on the screen.
- Space2: besides the Space1
condition, the target is also selected when the pen passes into
the target from above before the pen lands on the screen for the
first time. However, the target is highlighted only while the
pen is within the 1 cm high cylinder above the target before the
pen touches on the screen.
Subjects
Nineteen subjects (14 male,
5 female), all university students, were tested for the experiment.
Seven people had previously experienced pen-input systems, while
the others had no experience.
Procedure
For each strategy, each subject
followed steps (a) to (d). In addition, each subject repeated
these steps for all six strategies on a pen-based computer (HD-640A,WACOM
Co.). (a) Pointing of initial position: the subject pointed to
the initial position displayed at the center of the screen with
the pen. (b) Display of a target: the experimental system changed
each of five kinds of targets (1, 3, 5, 7, 9 dots in diameters
respectively, 0.36 mm per dot) and eight directions of pen-movement
(0, 45, 90, 135, 180, 225, 270 and 315 degrees), but the distance
(131 dots) between the initial position to the target was the
same. (c) Target selecting: The subject would receive a message
on the screen to indicate whether he/she made a successful selection
or not. (d) The subject repeated (a) and (c) as above.
RESULTS
In order to examine the influence
on the strategies by changing target sizes using the experimental
data, an ANOVA with repeated measures was performed to determine
whether there was a significant difference among the six strategies
for each target size and each direction in which the pen moved.
The result showed that the
"Land-on2" strategy was the best one among the six strategies,
in terms of error rates (F5,108 = 9.76, p < .01), selection
time (F5,108 = 3.17, p < .05) and subjective evaluation (F5,30
= 258.5, p <.0001). The Land-on2 was much more suitable for
selecting a minute target.
There were significant differences
for the six strategies in error rates, in each target size of
1, 3 and 5 dots. On the other hand, there were no significant
differences in 7 or 9 dots (p <.05). Thus, we clarified the
boundary value of the target which controlled the difficulty of
strategies. When a target is less than 5 dots (1.80 mm circle
in diameter), it is necessary to pay attention to the determination
of strategies in the software design based upon our experimental
results. We determined 5 dots as "the minute maximum value".
In other words, in case of selecting a target size of more than
5 dots, the selection performance becomes identical even by using
different strategies. In Fitts' law [5], ID=log2(2D/W) where ID
is an abbreviation of Index of Difficulty of pointing,
D is the distance, and W is the width of the target, if D is a
constant, W1, W2,... > 5 dots, then, there are no significant
differences among ID1=log2(2D/W1), ID2=log2(2D/W2), ..., as far
as our results are concerned. In other words, when selecting a
target size of more than 5 dots, the difference of the ID (here,
we consider error rates) will disappear.
DISCUSSION
In this experiment, in order
to investigate a strategy suitable for selecting a minute target,
we used a single target. We used circular targets to keep the
same distance which the pen reaches to the edge of a target displayed
in different direction as the first step of the study. The reason
why 131 dots were chosen as the distance in the study was that
we considered the maximum finger-moving-distance when the wrist
is in a fixed condition in the case of using portable pen-input
systems [6]. It was reported that there was an influence on the
selection time by the differences of the target shapes [7]. It
is necessary to examine the other target shapes in order to compare
them with the results of the circular targets used in this experiment,
to investigate the relations between strategies and target shapes
and to find strategies which are suitable for specific shapes.
It is interesting to pay attention to whether or not a minute
target of those shapes change in maximum value.
REFERENCES
1. Sears A. and Shneiderman
B. (1991). High precision touchscreens: Design strategies and
comparisons with a mouse, International Journal of Man-Machine
Studies, Vol.34, No.4, pp.593-613.
2. McClintock M. and Hoiem
D. (1993). Minimal target size in a pen-based system, In Abridged
Proceedings of 5th International Conference on Human-Computer
Interaction, Elsevier Science Publishers B.V., p.243.
3. Buxton W.(1990). A three
state model of graphical input. In D. Diaper et al. (Eds.), Human-Computer
Interaction ---INTERACT '90, pp.449-456.
4. Sears A., Plaisant C. and
Shneiderman B. (1992). A new era for high precision touchscreens,
Advances in Human-Computer Interaction, Vol.30, Ablex,
Norwood, NJ, pp.1-33.
5. Fitts P.M. (1954). The
information capacity of the human motor system in controlling
amplitude of movement, Journal of Experimental Psychology,
Vol.47, No.6, pp.381-391.
6. Ren X. S. and Moriya S.
(1995). The concept of various pointing strategies on pen-based
computers and their experimental evaluation, Proceedings of
the Eleventh Symposium on Human Interface (Kyoto, Japan),
pp.565-574.
7. Sheikh I. and Hoffmann
E. (1994). Effect of target shape on movement time in a Fitts
task. Ergonomics, Vol.37, No.9, pp.1533-1547.
CHI 97 Electronic Publications: Late-Breaking/Short Talks
