Copy-and-Paste Between Overlapping Windows
Olivier Chapuis1,2 Nicolas Roussel1,2
{chapuis,roussel}@lri.fr
1LRI - Univ. Paris-Sud & CNRS 2INRIA
Baˆt. 490, F-91405 Orsay, France Baˆt. 490, F-91405 Orsay, France
ABSTRACT
Copy-and-paste, one of the fundamental operations of mod-
ern user interfaces, can be performed through various means
(e.g. using the keyboard, mouse-based direct manipulation
or menus). When users copy and paste between two differ-
ent windows, the process is complicated by window man-
agement tasks. In this paper, we propose two new window
management techniques to facilitate these tasks in the par-
ticular case of partially overlapping windows. We describe
an experiment comparing four commonly used copy-and-
paste techniques under four window management conditions
– non-overlapping windows, partially overlapping windows,
and partially overlapping ones with one of our two window
management techniques. Results show that our new win-
dow management techniques significantly reduce task com-
pletion time for all copy-and-paste techniques. They also
show that X Window copy-and-paste is faster than the other
three techniques under all four window management condi-
tions.
Author Keywords
Copy-and-paste, Window Management, Overlapping Win-
dows.
ACM Classification Keywords
H.5.2 [Information interfaces and presentation]: Interaction
styles, Windowing systems.
INTRODUCTION
Copy-and-paste (or copy-paste) is the basic mechanism for
replicating part of a document in the same or another doc-
ument. Already available in early systems such as Sketch-
pad [28] or NLS [11], copy-paste is one of the fundamen-
tal services provided by modern graphical user interfaces.
Copy-paste requires the user to specify two things: the ob-
ject(s) to copy and the destination. These can be done in dif-
ferent orders and using various means such as the keyboard,
mouse-based direct manipulation or menus. Over the years,
several “standard” techniques have emerged, such as the use
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies are
not made or distributed for profit or commercial advantage and that copies
bear this notice and the full citation on the first page. To copy otherwise, or
republish, to post on servers or to redistribute to lists, requires prior specific
permission and/or a fee.
CHI 2007, April 28 - May 3, 2007, San Jose, California, USA.
Copyright 2007 ACM 978-1-59593-593-9/07/0004...$5.00.
of Ctrl-C to copy previously-selected objects and Ctrl-V to
paste them. But although these techniques are used by mil-
lions of people several times a day, the interaction is still
poorly understood. The techniques are implemented differ-
ently across operating systems and among applications1 but
most importantly, to our knowledge, they have never been
formally evaluated.
Copy-paste operations between two different windows usu-
ally require users to perform additional window manage-
ment tasks. If the source and destination windows overlap,
for example, the user often has to temporarily change the
stacking order to specify the objects to copy and the des-
tination. Yet again, the interactions and potential interfer-
ences between copy-paste and window management opera-
tions have received very little attention. A notable excep-
tion is Dragicevic’s work on the Fold n’ Drop technique [9]
that could be applied to the particular case of drag-and-drop
copy-paste.
In this paper, we propose two new window management
techniques, restack and roll, to facilitate copy-paste between
partially overlapping windows. We describe an experiment
comparing four commonly used copy-paste techniques (key-
board shortcuts, a context menu, drag-and-drop and a tech-
nique specific to X Window) under four window manage-
ment conditions: non-overlapping windows, partially over-
lapping windows, and partially overlapping ones with one of
our two new window-management techniques. Results from
this experiment show that restack and roll significantly re-
duce the task completion time for all copy-paste techniques.
They also show that the X Window technique is faster than
the three others under the four window management condi-
tions.
The paper is organized as follows. In the next three sections,
we review some of the related work, describe some common
copy-paste techniques and report on a series of interviews
that illustrate how they are used in practice. We then present
our restack and roll techniques, detail the experiment that
was conducted to evaluate them and compare the four copy-
paste techniques. Finally, we discuss some implications of
our results, propose some solutions to the problems identi-
fied in the paper and generalize our key ideas into the con-
cept of fine-grained window management.
1On Microsoft Windows XP, for example, selecting text with the
mouse in an overlapped window works with NotePad but not with
WordPad.
CHI 2007 Proceedings • Navigation & Interaction April 28-May 3, 2007 • San Jose, CA, USA
201

RELATED WORK
Although tiled windows may be more efficient for certain
tasks [6, 14], the overlapping model is the de facto standard
for all modern window systems and plays an essential part in
the realization of the desktop metaphor. The overlapping ap-
proach supports both time-based and space-based multiplex-
ing of the screen space by switching (between windows) and
splitting (the screen) [16]. However, as the number of win-
dows increases, it imposes time-consuming and potentially
complex management tasks on the user. The goal of the work
presented in this paper is to reduce this overhead. Exam-
ples of related work include techniques for leafing through
stacked windows, peeling them back or making them selec-
tively transparent to access windows underneath [3, 9, 13],
and dynamic space management algorithms to reduce over-
lapping [4].
Users with large displays tend to leave more applications
running and associated windows open [21]. Like Hutchings
and Stasko [12], we believe that overlapping windows will
not disappear with the advent of larger displays. First, a vari-
ety of devices will keep using small or medium-size screens.
More fundamentally, although large displays make it easier
to develop tiling strategies, interactions across large screen
distances may become more complex and time-consuming
than keeping windows together on a smaller space. Large
displays are probably better used to differentiate primary and
peripheral activities, i.e. for tiling tasks, not windows [21].
We anticipate that larger displays will lead to fewer maxi-
mized (full-screen) windows that completely hide others. In
some cases, the previously-obscured windows may become
tiled on a larger display, but in many others, they will par-
tially overlap. Therefore, copying and pasting between par-
tially overlapping windows will remain important.
Modifying an existing document or combining pieces from
several is always easier than creating a new one. Design-
ers of the Xerox Star said it elevated the concept of copying
to the higher level of “a paradigm for creating” [25]. We
indeed believe that the combined use of overlapping win-
dows and copy-paste supports innovation and creativity [24].
Copy-paste has been studied in specific domains such as pro-
gramming environments [31, 15], graphical editors [8] or
ubiquitous computing environments [19, 20]. Much previ-
ous research has tried to make it “smarter” by analyzing the
selected data. Citrine [27], for example, can recognize that
structured text has been copied, and paste it in multiple form
fields in a single operation. Other systems have been pro-
posed to support fast copy-paste of multiple selections or
text entities like phone numbers [18, 5]. In this work, we are
not interested in the objects being copied, or in optimizing
copy-paste for a particular domain. Rather we are interested
in the low-level interactions between copy-paste and win-
dow management operations.
COPY-PASTE TECHNIQUES
In this section, we describe what we believe are currently
the four most common copy-paste techniques. Note that al-
though we focus on copy-paste, most of these techniques
can also be used for cut-and-paste operations, the main dif-
ference being that the selection is deleted after the copy has
been made. Other differences between copy and cut will be
further explained, as needed.
Copy-paste usually starts by using the mouse to select one
or more object using one or more click(s) or a press-drag-
release gesture2. This selection might be assisted, for ex-
ample by automatically snapping to the edges of objects for
example. The user must then (1) activate the copy command,
(2) specify the destination – in the same window or another
one – and (3) activate the paste command. We will now de-
scribe several ways of accomplishing these three operations.
Using the Keyboard
Sketchpad and the Xerox Star had specific Delete, Copy and
Move keys that could be used in conjunction with the point-
ing device. Pressing the Copy key on a Star, for example,
attached the selection to the cursor, and then a mouse click
specified the destination. Modern systems do not have spe-
cific keys for these functions but support keyboard-based
copy-paste in a less modal way: (1) a first shortcut, e.g. Ctrl-
C, causes the selection to be copied; (2) the user navigates
to the destination using the mouse and/or the keyboard; (3)
a second shortcut, e.g. Ctrl-V, performs the paste.
We refer to the use of keyboard shortcuts to activate the
copy-paste commands as KEY copy-paste.
Using Menus
In addition to being accessible through keyboard shortcuts,
copy-paste commands are usually found in the standard
menu bar of applications, e.g., under the Edit item, as icons
in palettes and toolbars, and in context menus accessible
from the selected objects, e.g., using a right click.
Menu bars are very similar to context menus but impose ad-
ditional mouse travel to reach them after selecting objects
and after indicating the insertion point. Copy and paste icons
in toolbars or palettes have the same problem, so we decided
to focus on the use of context menus to activate the copy and
paste commands.
We refer to the use of context menus to copy-paste as MENU
copy-paste.
Using Drag-and-Drop
Drag-and-drop offers a more direct way of performing a
copy-paste operation. The user simply has to press a mouse
button on one of the selected objects, drag the mouse pointer
to the destination and release the button. However, this tech-
nique has several problems. First, its semantics are not al-
ways easy to determine: although one can reasonably as-
sume that dropping something on a trash icon deletes it,
dropping it somewhere else might copy it or move it. As a
consequence, application designers often disagree with users
on what the drag-and-drop operation should do [10].
2As a notable exception, users of the Xerox MESA programming
environment had to first specify the destination and then the text to
be copied [29]. Note that objects might also be selected using the
keyboard, but this does not affect the descriptions that follow.
CHI 2007 Proceedings • Navigation & Interaction April 28-May 3, 2007 • San Jose, CA, USA
202