Window Management Strategies
Introduction
The output of early computers was printed by Teletype on an ever growing
scroll of paper. As designers switched to high-speed displays, the need
to go back was sometimes supported by electronic scrolling of the session.
This technique is useful, but designers became aware of similar situations
in which users have to jump around to related text or graphics. Programmers
have to jump from procedural code to data declarations, or from procedure
invocations to procedure definitions. Authors of scientific papers jump
from writing the text to adding a bibliographic reference to reviewing empirical
data to creating figures to reading previous papers. Airline reservationists
jump from working on a client itinerary to reviewing schedules to choosing
seat assignments.
The general problem for many computer users is the need to consult multiple
sources rapidly, while minimally disrupting their concentration on their
task. With large desk- or wall-sized displays, many related documents can
be displayed simultaneously, but visibility and eyep;head movement might
be a problem. With small displays, windows are usually too small to provide
adequate information or context. In the middle ground, with 9- to 27-inch
displays (approximately 640 x 480 to 2048 x 2048 pixels), it becomes a design
challenge to offer users sufficient information and flexibility to accomplish
their tasks while reducing window housekeeping actions, distracting clutter,
and eyep;head movement. The animation characteristics, three-dimensional
appearance, and graphic design play key roles in efficacy and acceptance
(Gait, 1985; Kobara, 1991; Marcus, 1992).
If users' tasks are well understood and regular, then there is a good chance
that an effective multiple-window display strategy can be developed.
The airline reservationist might start a client-itinerary window and, review
flight segments from a schedule window, and drag selected flight segments
to the itinerary window. Windows labeled seat selection or food preferences
might appear as needed, and then the charge-card information window would
appear to complete the transaction. When the sequence is varied and unpredictable,
users will need to have more control of the layout and will need more training.
Window housekeeping is an activity related to the interface domain
of the OAI model and not directly related to the user's task. If window-housekeeping
actions can be reduced, then users can complete their tasks more rapidly.
In an empirical test with eight experienced users, the windowed version
of a system produced longer task-completion times than did the nonwindowed
(full-screen) environment (Bury et al., 1985). Multiple smaller windows
led to more time arranging information on the display and more scrolling
activity to bring necessary information into view. However, after the time
to arrange the display was eliminated, the task-solution times were shorter
for the windowed environment. Fewer errors were made in the windowed environment.
These results suggest that there are advantages to using windows, but these
advantages may be compromised unless automatic window arrangement is provided.
On small displays with poor resolution, opportunities for using multiple
windows are limited unless the user can tolerate frequent and annoying horizontal
and vertical scrolling. With medium-resolution displays and careful design,
multiple windows can be practical and esthetically pleasing. Window-border
decorations can be made to be informative, useful, and attractive. On larger,
high-resolution displays, windows become still more attractive, but the
manipulation of windows can remain as a distraction from the user's task.
Opening windows, moving them around, changing their size, or closing them
are the most common operations supported (Card et al., 1984; Myers, 1988).
The visual nature of window use has led many designers to apply a direct-manipulation
strategy (see Chapter 6) to window actions. Instead of typing a command
to stretch, move, and scroll a window, users can point at appropriate icons
on the window border and simply click on the mouse button (Billingsley,
1988; Kobara, 1991; Marcus 1992). Since the dynamics of windows have a strong
effect on user perceptions, the animations for transitions (zooming boxes,
sequencing of repainting when a window is opened or closed, blinking outlines,
or highlighting during dragging) must be designed carefully.
It is hard to trace the first explicit description of windows (Hopgood et
al., 1985), although several sources credit Doug Engelbart with the invention
of the mouse, windows, outlining, collaborative work, and hypertext as part
of his pioneering NLS system at the Stanford Research Institute during the
mid-1960s (Engelbart, 1988). Movable, overlapping windows appeared in the
Smalltalk graphical environment (Figure 13.1) as it evolved in the 1970s
at Xerox PARC, with contributions from Alan Kay, Larry Tesler (1981), Daniel
Ingalls, and many others. In 1981, the highly graphical Xerox Star (Figure
13.2) (Smith et al., 1982, Johnson et al., 1990) allowed up to six nonoverlapping
windows (with limited size control, movement, but no dragging of windows
or icons) to cover the desktop plus multiple property sheets to overlay
temporarily parts of the windows or desktop. Soon after, the Apple Lisa
and, in 1984, the Apple Macintosh (Figure 13.3) made popular their style
of graphical user interface with overlapping windows (Apple, 1987). Microsoft
followed with the graphical MS Windows 1.0, 2.03, 3.0, and Windows 95 (Figure
13.4ap;d) for IBM PCs, while IBM offered OS/2.
The notion of collections of windows assembled into rooms is an important
step forward in matching window strategies to users' tasks (Henderson and
Card, 1986; Card and Henderson, 1987). Users can open and leave a set of
windows in one room for reading electronic mail, while another room might
have a set of windows for composing an article or a program. Rooms can be
seen as a form of window macro that enables users to specify actions on
several windows at a time. Hewlett-Packard's HP-VUE implements the rooms
idea as a set of six workspaces users can visit (Color Plate 7). Much progress
has been made, but there is still an opportunity to reduce dramatically
the housekeeping chores with individual windows and to provide task-related
multiple-window coordination. Innovative features, inventive borders or
color combinations, individual tailoring, programmable actions, and cultural
variations should be expected. An effective overview of windowing strategies
is available in videotape form (Myers, 1990).
 |
Please send
comments and suggestions to the Booksite
Director
Last Updated:
11 December 2002
|
