Content uploaded by Jarno Ojala
Author content
All content in this area was uploaded by Jarno Ojala on Dec 03, 2020
Content may be subject to copyright.
A Diary Study on Combining Multiple Information Devices
in Everyday Activities and Tasks
Tero Jokela
Nokia Technologies
P.O. Box 1000, FI-33721
Tampere, Finland
tero.jokela@nokia.com
Jarno Ojala
Tampere Univ. of Technology
P.O. Box 589, FI-33101
Tampere, Finland
jarno.ojala@tut.fi
Thomas Olsson
Tampere Univ. of Technology
P.O. Box 589, FI-33101
Tampere, Finland
thomas.olsson@tut.fi
ABSTRACT
As people possess increasing numbers of information
devices, situations where several devices are combined and
used together have become more common. We present a
user study on people’s current practices in combining
multiple information devices in their everyday lives,
ranging from pragmatic tasks to leisure activities. Based on
diaries and interviews of 14 participants, we characterize
the usage practices of the most common devices, including
smartphones, computers, tablets, and home media centers.
We analyze 123 real-life multi-device use cases and
identify the main usage patterns, including Sequential Use,
Resource Lending, Related Parallel Use, and Unrelated
Parallel Use. We discuss the practical challenges of using
several information devices together. Finally, we identify
three levels of decisions that determine which devices are
used in a particular situation, including acquiring, making
available, and selecting the devices for use.
Author Keywords
Information devices; smartphones; tablets; multi-device;
device ecologies; mobile use; user study.
ACM Classification Keywords
H.5.m. Information interfaces and presentation (e.g., HCI):
Miscellaneous.
INTRODUCTION
Visions of ubiquitous computing have long predicted an
evolution from single-device computing towards computing
with multiple devices. Weiser [19] envisioned a future of
multiple computers of different scales per user, all
interconnected by a ubiquitous network. Norman [14]
proposed families of information appliances, specialized to
perform specific activities and capable of sharing
information among themselves, as a solution to the
complexity of personal computers. In line with these
predictions, people today own and use increasing numbers
of interconnected devices with diverse form factors. First
smartphones and then tablets have established themselves
as new device categories alongside personal computers.
Some conventional devices such as televisions and cameras
have become connected to the Internet and gained
capabilities to provide access to many of the same
applications and services as computers. Emerging industry
trends, including wearable devices, connected cars, and the
Internet of Things, suggest continuing increase in the
number and diversity of devices in the future.
While this development towards computing with multiple
devices offers many new opportunities, it has also created a
need for interfaces, applications, and services that better
support multi-device use. Today, many popular applications
and services can be accessed with a range of devices with
different screen sizes and form factors. A session of use can
be saved and closed on one device and re-opened and
continued on another device. Cloud services support
centralized management of device families, including
device settings and installed applications, and file hosting
services allow accessing and synchronizing content
between devices. Recent web browsers provide a common
usage history across all devices and allow moving browser
tabs between devices. A multiplicity of technologies has
been developed for presenting visual or audio information
through other devices, for example, using one device to
show pictures on the screen or to play music through the
speakers of another device. Attempts to harmonize software
and hardware platforms and tools across device categories
aim to make it easier to develop applications that support
multiple devices.
While a wide variety of different technologies have been
proposed to support multi-device use, only a few studies
have addressed how people actually use multiple
information devices together in their everyday lives. Of
these studies, many are relatively old, pre-dating the
emergence of modern smartphones, tablets, and cloud
services. In this paper, we present a recent diary study on
how people today combine multiple information devices in
everyday activities and tasks. Based on diaries and
interviews of 14 participants, we characterize the evolving
usage practices of the most common devices, including
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. Copyrights for
components of
this work owned by others than ACM m ust be honored .
Abstra cting with credit is p ermitted. To copy otherwise, or republish , to
post on servers or to redistribute to lists, requires prior specific permission
and/or a fee. Request permissions from Permission
s@acm.org.
CHI 2015, April 18
- 23 2015, Seoul, Repub lic of Korea
Copyright 2015 ACM 978
-1-4503-3145-6/15/04…$15.00
http://dx.doi.org/10.1145/2702123.2702211
Multi-Device Interaction
CHI 2015, Crossings, Seoul, Korea
3903
smartphones, computers, tablets, and home media centers.
We analyze 123 user-reported cases of multi-device use and
identify the main usage patterns, including Sequential Use,
Resource Lending, Related Parallel Use, and Unrelated
Parallel Use. We discuss the practical challenges of owning
and operating several information devices together. While
most of the earlier studies have been based solely on user
interviews, our study is additionally grounded on user
diaries reporting real-life use cases recorded in real contexts
of use over a time period of one week. Also, while most of
the earlier studies have focused on information work related
use, our study explores a range of different use cases from
pragmatic tasks to leisure activities in various mundane
contexts of use. The objective of our study is to provide
qualitative insights into real-life practices of multi-device
use. The results inform the design of future interfaces,
technologies, and applications that better support multi-
device use.
RELATED WORK
A wide variety of systems, user interfaces, and interaction
methods have been proposed to support computing with
multiple devices. These include device binding methods
that allow connecting several devices to operate together,
ranging from virtual methods (such as scanning for
available devices in the proximity) to physical methods
(such as synchronous user actions, spatial alignment, and
use of auxiliary devices) [18]. Similar techniques have also
been developed to support transferring content objects or
application windows between screens, including transfers
between devices within hand’s reach and between devices
at longer distances from each other [13]. Migratory
interfaces [1] provide techniques for moving application
windows between devices. Solutions have also been
developed for managing and switching tasks in multi-
device computing environments [2]. Beyond single-user
systems, a broad range of multi-user systems that support
collaboration with multiple devices have been developed,
addressing different physical scales and types of social
interaction [17].
In addition to the development of methods and solutions to
support multi-device use, a number of studies have
addressed how people actually use multiple devices in real
life. Several reasons and motivations for using multiple
devices have been identified [15, 3]. Different devices suit
different tasks, social situations, and physical environments.
Estimated effort to set up the device is an important
consideration in selecting which device to use, especially
for short tasks. No single device may have all the functions
or data needed, forcing the use of multiple devices.
Additional devices can also be used as data or battery
backups. Sometimes personal preferences and habits may
influence the decision to use different devices for different
tasks.
Practices and workflows in using multiple devices have
been found to vary between different individuals and
professional groups [3, 9, 16]. People tend to divide tasks
between devices, assigning each device a specific role
within the workflow [5, 3, 16]. In addition to serial patterns
of multi-device use [9], also parallel patterns have recently
become more common [16, 4, 12].
Accessing and managing content across devices has been
observed to be one of the key concerns in multi-device
environments [15, 3, 16, 12]. People have been found to
assemble their own personal patchworks of solutions by
combining multiple different tools and approaches, often in
creative ways. Conventional solutions include manual
synchronization and mirroring between devices, dedicating
a certain device for certain kind of information, portable
storage devices such as USB memory sticks, e-mailing
content items to oneself, and network drives. More recently,
various cloud-based storage and synchronization solutions
have increased their popularity. In addition to content, other
kinds of data such as interaction histories should also be
synchronized between devices [15, 3, 8].
In mobile use, managing different device configurations
may require significant physical effort and planning [15].
People have been observed to address this problem by
adopting different strategies, including development of
stable habits, making just-in-case preparations for potential
situations, and doing careful advance planning. A Mobile
Kit refers to a stable set of multiple personal devices kept
together and carried while traveling [11]. Jung, et al. [7]
suggest examining the set of digital devices that a person
uses as an ecology of interactive artifacts, in order to
understand how people experience and strategize the use of
interactive artifacts and the development of their ecologies
over time.
While a number of studies have addressed how people use
multiple devices in real life, they are still relatively few
given the broadness of the topic. Further, several of the
earlier studies have been made before the availability and
widespread adoption of modern smartp hones, tablets, and
cloud services. Therefore, there is room for an updated
view into how the new extended ecologies of devices and
services are managed by users. Most of the earlier studies
have also focused on technologically advanced users and
use cases related to information work. We believe our study
addressing diverse groups of users and exploring beyond
work-related use is useful in broadening the understanding
of the evolving practices and needs in multi-device use.
OUR STUDY
Objectives
In this study, we were interested in how people today
combine multiple information devices in their everyday
activities and tasks. For the purposes of this study, we
defined an information device as any device that can be
used to create or consume digital information, including
personal computers, smartphones, tablets, televisions, game
devices, cameras, music players, navigation devices, and
Multi-Device Interaction
CHI 2015, Crossings, Seoul, Korea
3904
wristwatch computers. In particular, we were interested in
the following three research questions related to multi-
device use: (1) What devices do the participants have and
what are the roles of the different devices in their personal
device ecologies? (2) In which situations do they combine
multiple information devices and what kinds o f practices do
they have for multi-device use? (3) What kind of challenges
and problems they face in multi-device use? Overall, based
on a qualitative analysis, we aimed at creating a rich picture
of the current and emerging practices and challenges in
using multiple information devices together, in order to
support the development of future multi-device interfaces,
technologies, and applications.
Participants
We recruited a total of 14 people living in Southern Finland
to participate in the study by posting an advertisement on
local mailing lists and social media groups. We selected a
diverse group of people who owned multiple devices and
actively used them for a wide range of different
applications and services. Six of the participants were
female and eight male. Seven of the participants were
students (age M=24.4, SD=2.8 years), while the other seven
were professionals working in different fields (age M=35.6,
SD=6.7 years). Regarding education and professional
background, four participants had their primary background
in information technology. The remaining ten participants
represented a wide variety of other professions, including
graphic and industrial designers, a medical doctor, a
goldsmith, a teacher, and a management assistant. Despite
not being pure technologists, the participants were fairly
advanced users of information technology: on a scale
between 1 and 7 (1=novice, 7=expert), the participants rated
their familiarity with information technology above average
(M=5.6, SD=0.8). The participants received small rewards
for their participation.
Method
The study consisted of three phases: an initial briefing
session, a one-week self-reporting period, and a final
interview. All the briefing sessions and interviews were
individual with the participant and one or two researchers
present. Five briefing sessions were held over the phone –
other briefing sessions and all interviews were made face-
to-face, either in a meeting room in our laboratory or in the
participant’s place of study or work. The purpose of the
briefing session was to provide the participant with
necessary information and materials for the self-reporting
period. After introducing the participant to the scope and
the objectives of the study, the researcher gave the
participant a workbook for the self-reporting period and
explained the contents of the book.
The main part of the workbook was a diary where the
participant was asked to report all situations and tasks
where they used multiple information devices together. We
asked the participant to keep a complete diary for at least
three days (two working days and one non-working day)
during the study period, but we also encouraged them to
report any interesting situations that occurred on the other
days. For each situation, the participant was asked to report
the context of use, overall course of events, devices that
were used, practices for and motivations of using multiple
devices, frequency and importance of the situation, as well
as satisfaction with the course of events and possible
problems. The researcher emphasized the importance of
filling in the book as soon as possible after the event
occurred. In addition to the diary, the workbook included
other tasks, where the participant was asked to provide
information about their devices, the tasks they were used
for, and the contexts they were used in. The participant was
also asked to describe the typical set of devices that they
carried with them in mobile situations, as well as their ideal
multi-device setup. The purpose of these tasks was to help
the participant to prepare for the interview by collecting
information about their device ecology and practices, and
reflecting them in real contexts of use (for example, at
home or at the workplace).
The participant then started the one-week self-reporting
period during which they filled in the workbook. At the end
of the period, the researchers tentatively analyzed the
workbooks to prepare for the semi-structured interviews.
We asked the participant to bring with them the devices
they carried on a typical work day. The interview started by
asking about the participant’s devices and their use,
followed by a detailed discussion of two or three different
situations that the participant had reported. Specific
questions were asked about various topics, including
sharing of content between devices, maintaining multiple
devices, and obtaining new devices. At the end of the
approximately one-hour interview, the participant was
asked to summarize the main benefits and drawbacks of
using multiple devices. The interviews were audio recorded
and photographs of the participant’s devices and their use
were taken when considered relevant.
For each interview, the researcher who made the interview
wrote notes about it based on the audio recordings. Three
researchers then analyzed the data and built an Affinity
Diagram [6] in a series of interpretation sessions. Based on
interpretative content analysis, the notes were grouped
based on similarity. The groups were then further clustered
to broader categories that were identified from the data and
jointly revisited, discussed, and refined. The diary data was
separately analyzed and the reported use cases were
categorized into different patterns of multi-device use based
on a bottom-up data-driven categorization. This was done
to develop an understanding of different patterns of multi-
device use based on the collected data rather than to prove
any a priori hypothesis. The categorization scheme was
primarily based on the roles and the configuration of the
devices as well as the user’s motivations for multi-device
use. Finally, based on the data from the other tasks in the
participants’ workbooks, we constructed matrices
Multi-Device Interaction
CHI 2015, Crossings, Seoul, Korea
3905
describing relationships between devices, tasks, and
contexts of use.
RESULTS
We begin by giving an overview of the participants’ device
ecologies, describing the most common devices and their
practices of use. We then present an analysis of the
collected multi-device use cases and identify the main
usage patterns, including Sequential Use, Resource
Lending, Related Parallel Use, and Unrelated Parallel Use.
Finally, we highlight various practical challenges related to
owning and using several devices together and discuss the
problem of selecting which devices to use.
Device Ecologies
On the average, each participant was actively using 7.9
information devices in their everyday life. Four participants
had mainly Apple devices, while the remaining 10 had a
more balanced mixture of devices from different
manufacturers. In addition to their personal devices, seven
participants also had devices provided by their employers.
Not surprisingly, the most commonly used devices by the
participants were smartphones, computers, tablets, and
home media centers (Fig. 1).
All participants (14/14) had a smartphone; two participants
had two smartphones, a work phone and a personal phone.
The smartphone was the only device that every participant
said they used every day. Several participants commented
that smartphone was their most important device and their
everyday life depended on it. “[P7] I am hooked on my
[smartphone]. There is no life without it. I use it for
everything. … It is my most important device that I use the
most.“ As the participants carried their smartphones almost
always with them, the phones allowed them to be reached
and kept them up to date anywhere they went. The
smartphones were used for a wide variety of different tasks,
including calendar, e-mail, messaging, phone calls, web
browsing, social media, music, and calculator. Smartphones
were also commonly used for taking pictures and for
several participants, the phone was their primary camera.
However, the use of smartphones was focused on retrieving
and entering small amounts of information and the
smartphones were not considered capable of handling large
amounts of information or complex tasks. While many
applications and services had smartphone versions, they
often lacked features or did not work properly.
All participants (14/14) also had laptop computers and more
than half of them (8/14) had several laptops in their use.
Nearly half of the participants (6/14) also had a desktop
computer. As computers were replaced by other devices in
consumption and entertainment oriented tasks, their use was
increasingly characterized by more complex tasks, which
included entry of long texts, creation of complicated
content, detailed work, and handling large amounts of
content. These tasks were often related to work or studies,
and they were described as important but serious and
somewhat boring. “[P13] My laptop is my trusted
companion, which I use for all important things: school and
work.” The participants considered their computers as their
most powerful devices that provided options beyond what
their other devices could offer. Computers were also
regarded as fallback devices that the participants turned to
when their other devices failed. Sometimes computers
formed a link between devices that could not otherwise
communicate with each other. Among modern smartphones
and tablets, the participants felt that their computers were
like legacy devices from the old world but they simply
could not manage without them. “[P6] After I got the
tablet, my laptop feels really ancient, somehow it is awfully
old-fashioned. But certain things you just cannot do without
it.”
Almost all participants (12/14) actively used tablets [12]
and two participants had several tablets. The two
participants who did not use tablets had tested them but
considered that they did not have use for a tablet as other
devices (especially smartphones) served the same purposes.
The tablets were primarily used for searching for and
consuming content and for entertainment, such as reading,
playing games, and watching videos. Some participants
considered tablets as smartphones with larger screens.
Compared to laptops, the tablets were considered to be
more lightweight and insta ntly available for use. Two
common patterns of use were observed: 1) a shared family
tablet at home for multiple purposes; 2) a personal tablet
that was partly replacing a laptop in mobile use. However,
tablets were not considered suitable for entering long texts
or doing complex or detailed work. One participant
described this limitation also as an advantage since it forced
him to focus on the essential. In general, tablets were still
looking for their role and their use involved more
experimentation than the use of the other devices.
For the purposes of this study, we defined Home Media
Center to consist of a TV or other large screen together with
game consoles, Home Theater PCs (HTPCs), and other
media appliances connected to it. All participants had some
kind of a Home Media Center system. The Home Media
Center was the heart of social entertainment at home. It
provided the largest screen in the household supporting
Figure 1. Most commonly used devices.
Smartphone
Home Media Center
Computer
Tablet
Camera
Handheld
Game Console
Music
Player
Health Monitoring
Device
Navigation
Device
Home
Automation
System
Multi-Device Interaction
CHI 2015, Crossings, Seoul, Korea
3906
collective viewing of and sharing of TV programs, movies,
games, music, and personal photos and videos. Video
consumption was dominated by on-demand Internet
services and local content providers. Lack of ready-made
solutions supporting all content sources led the participants
to build custom solutions, which were often PC based.
In addition to these four most commonly used devices, the
participants reported using a variety of other devices. Of the
other devices, the most common were digital cameras,
which were used in pre-planned situations where high
quality photographs and videos were desired. Other devices
included music players, handheld game co nsole s,
navigation devices, health monitoring devices, and home
automation systems.
The smartphones were considered the most perso nal
devices and were rarely shared, with only a few exceptions
such as a child playing games on a parent’s phone or a wife
taking a photo with her husband’s phone. Sharing tablets
and computers was more common, and within families, all
devices were in principle shared. Still, most devices had a
clear primary user and others used the device only
randomly. The primary user had their services and
applications configured and often always running on the
device. “[P5] All of our devices are shared. But I use
primarily some devices and my husband uses other devices,
because on those you always have the web pages, files, and
apps you need already opened.” Some participants also
expressed a desire to keep some of their devices private.
“[P6] I have tried to keep the tablet to myself, but of course
[my daughter] can use it as well if she wants. But some
devices you just want to keep to yourself.”
Patterns of Multi-Device Use
The participants reported a total of 111 situations and tasks
in which they had used multiple information devices
together during the one-week diary period. Initial analysis
of the data indicated that 23 of the situations involved
several separate instances of multi-device use and therefore
these situations were split into multiple cases. On the other
hand, 12 situations were rejected, either because the
description of the situation was too unclear or because the
situation did not involve use of multiple information
devices (for example, there was only one information
device used). Eventually, the preliminary analysis resulted
in a total of 123 cases of multi-device use.
Fig. 2 illustrates the categorization of these multi-device
use cases into the main patterns of use. On the highest level,
multi-device use can be divided into Sequential Use and
Parallel Use. In Sequential Use, the participant changed the
device during the task. In Parallel Use, the participant used
several devices simultaneously. Of the multi-device use
cases analyzed in the study, 37% were Sequential Use while
the remaining 63% were Parallel Use.
The cases of Parallel Use can be further divided into three
subtypes: Resource Lending, Related Parallel Use, and
Unrelated Parallel Use. In Resource Lending, the
participant’s task focused on a single device, but the device
borrowed some resources from other devices. In Related
and Unrelated Parallel Use, the participant used several
devices simultaneously. The difference is that in Related
Parallel Use, the participant was working on a single task
and all devices were involved in this task, while in
Unrelated Parallel Use, the participant was working on
multiple tasks simultaneously and different devices were
involved in different tasks. Of the cases of Parallel Use
analyzed in the study, 43% were classified as Resource
Lending, 44% as Related Parallel Use, and 13% as
Unrelated Parallel Use.
In general, the participants’ motivations for using multiple
devices could be divided in two broad categories. First, in
83% of the cases, the participant voluntarily decided to use
multiple devices, for example, in the hope of improved
performance, efficiency, or convenience of use. Second, in
the remaining 17% of the cases, the participant was forced
to use several devices, for example, because of technical
limitations or errors, or because the original device used to
start the task had become unavailable.
In the following subsections, we discuss each pattern of
multi-device use as well as related motivations and
behaviors in more detail.
Sequential Use
In Sequential Use (Fig. 3.a), the participant changed the
device during the task. The participant could change the
device once or several times in a sequence.
In 68% of the Sequential Use cases analyzed in the study,
the user voluntarily changed the device. We observed
several reasons that triggered the participant to switch
devices. A common reason was a transformation in the
character of the task, which made the participant to consider
another device better suited for the task and to change the
device. “[P7] I googled the phone number of my
physiotherapist with the tablet and called with my phone.”
Another common reason was a change in the physical
environment or the social context, which resulted in another
device deemed more appropriate for the task. “[P1] In the
bus, I was browsing with my phone and found an
interesting page about teacher’s copyrights. At work, I
Figure 2. Patterns of multi-device use.
Multi-Device Use
Unrelated
Parallel Use
Related
Parallel Use
Resource
Lending
Parallel Use
Sequential Use
Multi-Device Interaction
CHI 2015, Crossings, Seoul, Korea
3907
continued reading the page with my laptop.” Often, both
the task and the context changed at the same time. In these
cases, there was typically a workflow consisting of a
sequence of clearly separate steps, each done at a different
location. “[P13] I searched for a recipe with my laptop.
Then I moved to the kitchen, opened the recipe with the
tablet, and started cooking.” There were also a few cases in
which the task and the physical and social context remained
unchanged, but the participant still decided to change the
device. In these cases, the participant’s initial evaluation of
the task proved incorrect as more information was learned
about the task, ultimately resulting in a device change.
“[P13] When meeting the other members of the project
group, I tried to browse the course materials with my
phone, but finally switched to my laptop because the phone
was too slow and the display too small.”
The dilemma of whether to change the device or not can be
illustrated with Fig. 4.a. The horizontal axis represents the
progression of a task, while the vertical axis represents the
user effort. Each device requires an initial setup effort,
represented by the y-intercept of the line, before it can used
to work on the task. Each device also has certain efficiency,
represented by the slope of the line. As illustrated in Fig
4.b, the efficiency of the device may change as the
characteristics of the task change. In order to switch the
device (Fig. 4.c), the user has to do additional effort. While
the person may gain improved performance after the
switch, it may take a long time to outdo the switching
effort. The problem may be further complicated by the
difficulty of predicting the future evolution of the task.
In 32% of the Sequential Use cases, the user was forced to
change the device. The most common reason was technical
problems that prevented continuing with the original device
[9]. “[P6] I opened the electronic messaging system of my
daughter’s school with my tablet and tried to reply to a
message, but a scrolling text field did not work with the
tablet. I mailed the text I had entered to myself and
continued with my laptop.” Overall, technical problems
were the reason for using multiple devices in 25% of the
Sequential Use cases. Other reasons that forced the user to
change the device included the original device running out
of battery or the need to access content that was stored on
another device.
We observed a variety of different methods of moving a
task from one device to another. It was common that the
devices and applications did not support any way of moving
the task from one device to another, or the user was
unaware of or unwilling to use them. In these cases, the
user just started the task from the beginning with the new
device and manually copied information from the old
device to the new one as necessary. However, in many
cases it was possible to save the task on a network resource,
for example, on a cloud service or a network disk, and to
reopen it with another device. Other less common
approaches were to send a message, typically an e-mail, to
oneself, or to use a physical medium such as a USB stick or
memory card to transfer the task.
Parallel Use
In Parallel Use, the participant used two or more devices
simultaneously. The cases of Parallel Use can be further
divided into three subtypes: Resource Lending, Related
Parallel Use, and Unrelated Parallel Use.
Resource Lending
In Resource Lending (Fig. 3.b) use cases, the participant’s
activity primarily focused on a single device, but this device
borrowed some resources from other devices.
Common examples of Resource Lending included
borrowing the input and output capabilities of another
device. For example, the screen of another device could be
used to display visual information or speakers to play audio.
“[P6] I connected my laptop to my TV to watch an episode
of a TV series.” Alternatively, the input devices of one
device could be borrowed to provide input for another
device. “[P8] I controlled my home theatre system with an
app installed on my phone.” In some cases, both input and
output capabilities of another device were borrowed and all
interaction took place through another device. “[P2] I used
my laptop to organize messages and music on my phone.”
In addition to input and output capabilities, another
common resource shared between devices was the network
connection. “[P2] In the bus, I used the hotspot feature on
my phone to connect my laptop to the Internet.” Resource
Lending was almost exclusively done using direct wireless
connections between devices. Other approaches included
traditional cables and docking stations as well as lending
resources through a cloud service.
Device 1
Task A
d) Unrelated Parallel Use:
Devices involved in
different tasks.
c) Related Parallel Use:
All devices involved in a
single task.
b) Resource Lending:
Borrowing resources from
other devices.
a) Sequential Use:
Changing the device during
the task.
Device 2
Task B
Device 3
Task C
Device 1
Task A
Device 2
Task A
Device 3
Task A
Device 2
Resource A
Device 3
Resource B
Device 1
Task
Device 1
Task A
Device 2
Task A
Figure 3. Different patterns of multi-device use.
Multi-Device Interaction
CHI 2015, Crossings, Seoul, Korea
3908
Related Parallel Use
In Related Parallel Use (Fig. 3.c), the participant was
working on a single task using two or more devices. All
devices were involved in the same task.
In most cases, the motivation for using multiple devices in
parallel was to have multiple views to the task, dedicating
different devices to different content or applications. A
common situation was that a participant was using another
device to view additional information when watching video
content with a home media center. “[P5] While watching a
movie on TV, I opened the IMDB page with my phone.” In
other cases, a participant was viewing instructions on one
device while working on a task on another device. “[P1] I
installed software on my phone and looked at instructions
on my computer.” The participant could also work on a new
document on one device, while having the source
documents open on the other devices. ”[P10] While making
a project presentation with my laptop, I had my notes open
on my tablet.” Sometimes, there simply were several
documents that needed to be viewed in parallel. “[P10] I
displayed CAD drawings on my laptop and images on my
tablet.” A second device could also be used for simple
auxiliary tasks, for example, to access the dictionary. “[P8]
While watching TV, I checked the translation of a word
with my phone.”
Another common category of cases of Related Parallel Use
was situations where the participant was collaborating with
remote persons over a real-time communication link. One
device, typically a phone, was then assigned for handling
the communication, while other devices were used for
accessing related content and applications. “[P8] My friend
called me and asked me to the movies. I checked my
calendar and information about the movie with my laptop
while talking on the phone.” Other less common
motiva tions for Related Parallel Use included speeding up a
computationally complex task like 3D graphics rendering
by using multiple devices as well as various technical
reasons which forced the use of several devices.
In most cases, the participants used two devices together.
However, in a few cases the participants had built more
sophisticated device configurations which involved Parallel
Use of three or more devices. “[P4] While watching sports
on TV, I have live statistics open on a laptop. At the same
time, I can do real-time betting on my tablet or phone.” In
practically all cases, the devices provided no technical
support of any kind for Related Parallel Use. The
participants had to manually connect and transfer
information between the devices.
Unrelated Parallel Use
In Unrelated Parallel Use (Fig. 3.d), the participant was
working on several tasks simultaneously using several
devices. Different devices were involved in different tasks.
In cases of Unrelated Parallel Use, there was typically a
primary foreground task and a secondary background task.
Typical examples of background tasks included watching
videos or listening to music. “[P10] I listen to music with
my phone while doing homework with my laptop or tablet.”
In a few cases, the participant was working on two equal
tasks in parallel, for example, when the participant received
two simultaneous phone calls. “[P1] I was talking with my
mother on my personal phone, when my girlfriend called to
my work phone.” There were also a few cases where
technical restrictions required the participant to use
different devices for different tasks. While different devices
were involved in different tasks, in some situations
coordination between the devices would have been
beneficial. ”[P6] I was listening to a net radio with my
laptop, when I received a phone call. I would have wanted
to turn down the laptop volume when the phone rang.”
The participants reported relatively few cases of Unrelated
Parallel Use. It is possible that the participants did not
recognize many common situations as Parallel Use of
multiple devices, as the unrelated parallel tasks may not
always be conscious. For example, carrying a phone in
order to be reachable by voice calls or messages when
working on other tasks with other devices (undoubtedly a
very common case) could be classified as Unrelated Parallel
Use.
Deciding Which Devices to Use
Owning multiple devices creates the problem of choosing
which devices to use in a specific situation. In our analysis
of multi-device use cases, we identified three levels of
decisions which determined which devices to use: 1)
deciding which devices to acquire, 2) deciding which of
your devices to make ava ilable in a specific context, a nd 3)
Figure 4. a) Different devices have different setup efforts and task efficiencies. b) Changing task characteristics may influence
device efficiency. c) Switching devices requires additional effort that may exceed the benefits.
a) Task Progress
Effort
b) Task Progress
c) Task Progress
Effort
Phone
Laptop
Read
e
-mail
View
attachment
Compose
reply
Phone
Switch to laptop
Effort
Multi-Device Interaction
CHI 2015, Crossings, Seoul, Korea
3909
deciding which of the available devices to actually use.
Each level of decisions narrowed the available options until
finally the devices to use were left. The participants’ device
selections determined the pattern of multi-device use: the
decision to change device resulted in a Sequential Use
pattern, while the decision to use two or more devices
simultaneously resulted in a Parallel Use pattern.
Regarding acquisition of new devices, most participants
said that interoperability with their current devices was a
major factor when considering which new devices to
obtain. On the other hand, other factors such as price or
curiosity to try and opportunity to learn new devices and
systems often overrode it. The level of influence the
participants had over the acquisition decisions varied: the
participants had more control over devices they purchased
for their personal use than over devices that were provided
to them by their employers. Overall, the participants could
not fully control their device environments as they were
constrained by the choices of other people and
organizations, such as their employers, educational
institutions, clients, and friends.
For fixed devices, the decisions on which devices to make
available in specific contexts primarily related to the
physical locations of the devices, for example, in which
room to place the device at home. For mobile devices, the
decisions related to which devices to take with you in
mobile situations. Regarding the mobile devices, we
observed similar practices as Oulasvirta, et al. [15]. Most
participants had a fixed Mobile Kit [11] of devices they
took with them every day. Heavier devices such as laptops
and special devices such as USB drives were taken based
on expected need. In addition to computing devices, the
participants carried with them a wide variety of auxiliary
devices such as chargers, headsets, cables, and memory
sticks.
Decisions on which of the available devices to use in
specific situations were based on several criteria. The
device characteristics that were considered included both
user interface capabilities (such as display size, pointing
device, and the level of multi-tasking support), technical
capabilities (such as processing power, storage capacity,
network connection, and camera quality), and physical
characteristics (such as the size and weight of the device).
For tasks requiring entry of long texts, the text entry
capability, especially availability of a physical QWERTY
keyboard, strongly influenced the device selection. “[P4] I
won’t write any long texts with a virtual keyboard. I want
the good old physical QWERTY.” As already discussed in
section Sequential Use, another major factor influencing the
decision on which device to use was the easiness of starting
to use the device. If the required content was available only
on certain devices, the participant was forced to use those
devices. People had also developed different habits of using
certain devices for certain tasks or in certain contexts of
use. Finally, the participants also considered the context of
use, including both social aspects (such as acceptability of
using a certain device in a certain social situation) as well
as physical aspects (such as the available space). “[P13] As
I met new persons, I did not want to hide behind the
[laptop] screen, as we were supposed to work together and
be social.”
Accessing Content Between Devices
Being able to access any content on any device was the
most commonly requested feature to support multi-device
use by the participants. For transferring individual content
items between devices, sending e-mail to oneself was still
common, but also direct network transfers between devices
over WiFi or Bluetooth were used. Traditional physical
methods of memory sticks, cables, and DVDs had largely
been replaced by wireless or cloud-based solutions, but
were still used as fallbacks when more advanced solutions
failed to work, especially between systems of different
owners.
The participants utilized cloud storage services (for
example, Dropbox, Google Drive, and iCloud) for a wide
variety of purposes, including storing and accessing
personal content from any device, and synchronizing and
making backups between devices. Cloud storage services
were also used for other purposes, such as sharing and
collaborating between people and for temporary storage.
Using several services in parallel was common as this
allowed the participants to grab free storage space from
every service and reduced the risk of being tied to a single
service provider. It was also common to dedicate different
services to different purposes or content types, for example,
using one service for personal content, another for work,
and a third for collaboration with others.
While the participants utilized cloud storage services in
many ways, they also raised many concerns about cloud
storage and did not trust the cloud as the only storage
solution [16], especially for important content. Accessing
content in the cloud was considered slow and sometimes
unreliable, particularly over wireless networks. “[P5]
Accessing [cellular] networks sometimes causes problems
in certain areas. It is really infuriating.” To address this
problem, one participant used different wireless network
operators on different devices to be able to always select the
best working network. A few participants were concerned
about privacy and security risks, for example, of storing
personal photographs or work-related content in the cloud.
Another source of concern was the persistence and possible
discontinuation of cloud services. Finally, many current
cloud services were considered complicated and poorly
integrated with native applications.
Device Maintenance and Energy Management
Participants with large collections of information devices
had to do significant amounts of maintenance work to keep
their devices up to date and running. However, half of the
participants (7/14) said that this was not a major problem as
Multi-Device Interaction
CHI 2015, Crossings, Seoul, Korea
3910
the software updates were so easy to do. “[P8] All
[devices] have automatic updates and I just have to accept
them.” Other participants (6/14) considered the need to
maintain a large amount of devices as troublesome and
would have preferred more a utomated solutions.
Maintenance responsibility was often concentrated on a
particular person in a social network, for example, the
husband in a family. These people typically enjoyed the
maintenance task and considered it as a positive challenge.
“[P4] Device maintenance in our family is my
responsibility. … I like to do it. I take it as a challenge.
Updating the devices of my parents and my mother-in-law
are my responsibility, too.”
Another challenge related to the use of multiple devices
was managing the energy and charging the batteries of
portable devices. The battery concerns were mainly related
to smartphones and most participants (9/14) said they were
frustrated with the battery lives of the current smartphones.
Other devices were considered to have reasonably good
battery lives – tablets, in particular, were praised for their
long battery lives.
DISCUSSION
In general, the participants wanted all their devices to
seamlessly work with each other. However, in practice,
they continuously encountered problems in multi-device
use, especially between devices from different ecosystems.
Common problems included connecting and transferring
information between devices, incompatible content formats,
web pages that did not work on all devices, and applications
and services that were not available for all devices. The
participants often had to resort to common core functions,
such as e-mail, to work around interoperability problems.
Plenty of work still remains to be done to realize the visions
of smooth and effortless multi-device computing.
Today, most of the commercial support for multi-device use
is aimed towards supporting patterns of Sequential Use, for
example, moving sessions between devices. Also,
traditional forms of Resource Lending, such as displaying
pictures on the screen of another device, are well supported.
However, the results of our study indicate that also patterns
of Related Parallel Use were common among the
participants. This finding is supported by other recent
studies [16, 4, 12]. Still, the devices and systems used by
the participants provided practically no support for Related
Parallel Use.
Being able to access any content with any device was the
most commonly requested feature to support multi-device
use by the study participants. While the participants utilized
cloud storage services in many ways, it was obvious that
the current services were not adequate as the only storage
solutions. A wide range of user concerns should be
addressed in future storage solutions, including capacity,
cost, privacy, security, performance, reliability, persistence,
and complexity. As suggested by one of our participants,
one potential direction to explore might be more direct
sharing of content between devices, in order to provide
improved reliability and performance and lower cost:
“[P10] Information between devices currently flows
through the cloud, but when the devices are close to each
other, why it couldn’t pass directly between the devices?”
The participants described that in the modern world of
smartphones and tablets, the PC felt like a legacy device
from the old world. Some participants had tried to
completely replace their computers with smartphones and
tablets. However, it was clear that while smartphones and
tablets had taken over many simple tasks traditionally done
with computers, in their current form they could not fully
replace computers in more demanding tasks. The main
problems were related to limited text-entry capability,
inaccurate pointing devices, and restricted multi-tasking
capability. This calls for “reinventing the PC” in the age of
multi-device computing, that is, developing multi-device
solutions that are capable of handling the large, complex,
and detailed tasks that currently require the use of a
personal computer. Interestingly, the limited multi-tasking
capabilities of current tablets and smartphones already seem
to encourage the use of multiple devices to overcome them.
As the participants purchased new devices, they sometimes
recycled their old devices for others to use. However, often
the old devices were gradually left unused, partly because
the limitations of current systems to support multi-device
use promote the use of a single device. Over time, this
resulted in large collections of unused devices. Multi-device
ecologies might provide opportunities to extend the life of
old devices. The old devices could adopt more specialized
or supporting roles in the device ecology, for example, an
old tablet could be permanently attached to the kitchen wall
to support cooking activities, or an old computer could be
used as a home server.
Generalizability of Results and Future Work
As we wanted to gain insights into the current practices and
behaviors in combining multiple information devices, we
decided to approach the topic though a detailed analysis of
a limited number of subjects and cases, emphasizing
qualitative research methods. While this approach provided
us with a rich picture of the current practices and the
underlying motivations and needs, it is not possible to
statistically estimate the frequency of the observed
behaviors in the overall population. In order to validate the
results, a quantitative study, such as a survey or a logging
study, with a larger sample would be useful.
It should also be noted that the study results reflect current
practices – future devices and technologies may enable new
practices, as can also be seen when comparing these results
with the earlier similar studies. Also, while the participants
represented a rather diverse sample in terms of occupations,
they were still relatively advanced users of technology with
generally high level of education living in a western
Multi-Device Interaction
CHI 2015, Crossings, Seoul, Korea
3911
country. Different user groups in different parts of the
world might display different behavior. Another possible
future direction could be more focused studies addressing
specific user groups, applications, or contexts of use.
CONCLUSION
We have presented a user study on people’s practices in
combining multiple information devices in their everyday
lives, ranging from pragmatic tasks to leisure and
entertainment activities. Based on diaries and interviews of
14 participants, we have characterized the usage practices
of the most common devices, including smartphones,
computers, tablets, and home media centers. We have
analyzed 123 real-life cases of multi-device use and
identified the main usage patterns, including Sequential
Use, Resource Lending, Related Parallel Use, and
Unrelated Parallel Use. Additionally, we have observed
three levels of decisions that determine which devices are
used in a particular situation, including acquiring, making
available, and selecting the devices for use. We have also
discussed the practical challenges related to owning and
operating several information devices together, including
content access, maintenance, and energy management.
While the participants wanted all their devices to
seamlessly work with each other, in practice they
continuously encountered problems in multi-device use. Of
the multi-device use patterns, Sequential Use and Resource
Lending were relative well supported by current devices
and systems, but there was little technical support for
Related Parallel Use even though it was found to be
common among the study participants. Current cloud-based
storage solutions were found to have several weaknesses in
supporting multi-device use. Finally, improved support for
multi-device use might also provide opportunities to extend
the life of old devices by allowing the old devices to take
more specialized or supporting roles in the device ecology.
ACKNOWLEDGMENTS
We thank Prof. Kaisa Väänänen-Vainio-Mattila, Guido
Grassel, and Petri Piippo for contributing to the planning of
the study and for their valuable comments.
REFERENCES
1. Bandelloni, R., and Paternò, F. Migratory user interfaces
able to adapt to various interaction platforms. Int.
Journal of Human-Computer Studies 60, 5-6 (2004),
621-639.
2. Bardram, J. Activity-based computing for medical work
in hospitals. ACM Transactions on Computer-Human
Interaction 16, 2 (2009).
3. Dearman, D. and Pierce, J. It's on my other computer!:
computing with multiple devices. In Proc. CHI '08, 767-
776.
4. Google. The New Multi-Screen World Study.
https://www.thinkwithgoogle.com/research-studies/the-
new-multi-screen-world-study.html
5. Grudin, J. Partitioning digital worlds: focal and
peripheral awareness in multiple monitor use. In Proc.
CHI '01, 458-465.
6. Holtzblatt, K., Wendell, J. B. and Wood, S. Rapid
Contextual Design. Morgan Kaufmann, 2004.
7. Jung, H., Stolterman, E., Ryan, W., Thompson, T. and
Siegel, M. Toward a framework for ecologies of
artifacts: how are digital artifacts interconnected within
a personal life? In Proc. NordiCHI '08, 201-210.
8. Kane, S., Karlson, A., Meyers, B., Johns, P., Jacobs, A.
and Smith, G. Exploring cross-device web use on PCs
and mobile devices. In Proc. INTERACT ’09, 722-735.
9. Karlson, A., Iqbal, S., Meyers, B., Ramos, G., Lee, K.
and Tang, J. Mobile taskflow in context: a screenshot
study of smartphone usage. In Proc. CHI '10, 2009-
2018.
10. Karlson, A., Meyers, B., Jacobs, A., Johns, P. and Kane,
S. Working overtime: patterns of smartphone and PC
usage in the day of an information worker. In Proc.
Pervasive ’09, 398-405.
11. Mainwaring, S., Anderson, K. and Chang, M. Living for
the global city: mobile kits, urban interfaces, and
Ubicomp. In Proc. UbiComp ’05, 269-286.
12. Müller, H., Gove, J. and Webb, J. Understanding tablet
use: a multi-me thod exploration. In Proc. MobileHCI
'12, 1-10.
13. Nacenta, M. Cross-display object movement in multi-
display environments. PhD thesis, University of
Saskatchewan, 2009.
14. Norman, D. The Invisible Computer. MIT Press, 1999.
15. Oulasvirta, A. and Sumari, L. Mobile kits and laptop
trays: managing multiple devices in mobile information
work. In Proc. CHI '07, 1127-1136.
16. Santosa, S., and Wigdor, D. A field study of multi-
device workflows in distributed workspaces. In Proc.
UbiComp '13, 63-72.
17. Terrenghi, L., Quigley, A. and Dix, A. A taxonomy for
and analysis of multi-person-display ecosystems.
Personal and Ubiquitous Computing 13, 8 (2009), 583-
598.
18. Uzun, E., Saxena, N. and Kumar, A. Pairing devices for
social interactions: a comparative usability evaluation.
In Proc. CHI ‘11, 2315-2324.
19. Weiser, M. The computer for the 21st century. Scientific
American 265 (1991), 94-104.
Multi-Device Interaction
CHI 2015, Crossings, Seoul, Korea
3912
