Firm foundation in the main hci principles, the book provides a working

Figure 3.20 Deriving the actual movement time by fitting based on samples of performance 
data. (From MacKenzie, I. S., Human–Computer Interaction, 7(1), 91–139, 1992 [24].)
Target Distance
A
Index of difficulty
ID = log(A/W+1)
Movement time
MT = a + blog ( + 1)
Target Size
W
A
W
A = {2, 4, 6, 8} (mm)
W= {0. 25, 0.5, 1, 2} (mm)
ID = {1, 2, 3, 4, 5} (bits)
Figure 3.19 Illustration of Fitts’s law. (From MacKenzie, I. S., Human–Computer Interaction, 7(1), 
91–139, 1992 [24].)

---

5 5
H U M A N FA C T O R S A S H C I T H E O R I E S
modeled similarly, and several revised Fitts’s laws (e.g., for desktop 
computer interface, mobile interface) have been derived as well [24].
3.3.2 Motor Control
Perhaps the most prevalent form of input is made by the movements of 
our arms, hands, and fingers for keyboard and mouse input. Berard et 
al. have reported that there was a significant drop in human motor con-
trol performance below a certain spatial-resolution threshold [18]. For 
instance, while the actual performance is dependent on the form factor 
of the device used and the mode of operation, the mouse is operable 
with a spatial resolution on the order of thousands of dpi (dots per inch) 
or ≈0.020 mm, while the resolution for a 3-D stylus in the hundreds.
In addition to discrete-event input methods (e.g., buttons), modern 
user interfaces make heavy use of continuous input methods in the 
two-dimensional (2-D) space (e.g., mouse, touch screen) and increas-
ingly in the 3-D space (e.g., haptic, Wii-mote). While the human 
capabilities will determine the achievable accuracy in such input 
methods, the control-display (C/D) ratio is often adjusted. C/D ratio 
Figure 3.21 Applying Fitts’s law to a computer interface (dragging a file icon into the trash-
can icon). (From MacKenzie, I. S., Movement Time Prediction in Human–Computer Interfaces, in 
Proceedings of the Conference on Graphics Interface ’92, Morgan Kaufman, San Francisco, 1992, 
pp. 140–150 [25].)

---

5 6
H U M A N – C O M P U T E R I N T E R A C T I O N 
refers to the ratio of the movement in the control device (e.g., mouse) 
to that in the display (e.g., cursor). If the C/D ratio is low, the sen-
sitivity of the control is high and, therefore, travel time across the 
display will be fast. If the C/D/ ratio is high, sensitivity is low and, 
therefore, the fine-adjust time will be relatively fast.
Obviously, humans will exhibit different motor-control perfor-
mances with different devices, as already demonstrated with the two 
types of device mentioned previously (e.g., mouse vs. 3-D stylus). The 
mouse and 3-D stylus, for instance, belong to what is called the isomet-
ric devices, where the movement of the device directly translates to the 
movement in the display (or virtual space). Nonisometric devices are 
those that control the movement in the display in principle with some-
thing else such as force, thus possibly with no movement input at all.
Control accuracy for touch interface presents a different problem. 
Despite our fine motor-control capability of submillimeter perfor-
mance—and with recent touch screens offering higher than 4096-dpi 
resolution—it is the size of the fingertip contact (unless using a stylus 
pen), 0.3–0.7 cm, that makes it hard to make selection for relatively 
small objects. Even larger objects, once selected, are not easy to con-
trol if the touch screen is held by another hand or arm (i.e., unstable).