Pers Ubiquit Comput (2004) 8: 135–143

DOI 10.1007/s00779-004-0263-1

O R I GI N A L A R T IC L E

Sakari Tamminen Æ Antti Oulasvirta Æ Kalle Toiskallio

Anu Kankainen

Understanding mobile contexts

Received: 5 December 2003 / Accepted: 10 March 2004 / Published online: 24 April 2004
Ó Springer-Verlag London Limited 2004

Abstract Mobile urban environments present a challenge there are services for arranging ad hoc face-to-face
for context-aware computers because they differ from meetings with friends, finding driving directions, fixing
fixed indoor contexts such as offices, meeting rooms, and blind dates, playing games, and even chatting with un-
lecture halls in many important ways. Internal factors known people.
such as tasks and goals are different—external factors In its complexity, mobile lifestyle presents a chal-
such as social resources are dynamic and unpredictable. lenge for context-aware computing. Context-aware de-
An empirical, user-centred approach is needed to vices are supposed to monitor the changing contexts of
understand mobile contexts. In this paper, we present the user and adapt appropriately through interpreters,
insights from an ethnomethodologically inspired study aggregators, and services [1]. Mobile computers need
of 25 adult urbanites in Helsinki. The results describe ‘‘awareness’’ of several contextual factors including
typical phenomena in mobility: how situational and social, psychological, physical, and the like. What these
planned acts intermesh in navigation, how people con- factors are, and how they should be interpreted and
struct personal and group spaces, and how temporal acted upon, are questions addressed in this study.
tensions develop and dissolve. Furthermore, we provide Having deep social roots and involving a dynamically
examples of social solutions to navigation problems, changing environment, mobile context is a difficult
examine mobile multitasking, and consider design concept to define. It is problematic to decide from an
implications for mobile and context-aware human– armchair which attributes are relevant. In order to be
computer interaction. socially acceptable and useful, context-aware technology
must be based on empirical knowledge of the context
Keywords Mobile context Æ Mobility Æ Ethnography Æ analysed from the perspective of the end-users. In this
Ethnomethodology Æ Context-aware computing Æ paper, we show how different aspects of mobile con-
User-centred design texts are created and maintained by situated actions in
everyday life. Furthermore, we examine various impli-
cations that our findings pose for the design of context-
aware computing.
1 Introduction

Mobile technologies can be seen as new resources for 2 Previous research

accomplishing various everyday activities that are car-
ried out on the move. People have tremendous capa- What a context comprises is a widely debated and con-
bilities for utilizing mobile devices in various innovative troversial issue. A wide philosophically oriented body of
ways for social and cognitive activities. For example, literature exists within HCI about the nature of ‘‘con-
text’’ [2–7]. In addition, attempts to create a standard-
S. Tamminen (&) Æ K. Toiskallio ized definition of use context have been made (e.g. ISO
Information Ergonomics Research Group (SoberIT), 13407, [8]). However, some researchers consider the
Helsinki University of Technology,
P.O. Box 9210, 02015 Helsinki, Finland
present definitions of context too vague and general to
E-mail: Sakari.Tamminen@hut.fi be adapted to any specific design processes. The fol-
Tel.: +358-9-4516042 lowing objection is common [4, 5]: because context is
A. Oulasvirta Æ A. Kankainen
tightly intertwined with users’ internal and social—
Helsinki Institute of Information Technology, continuously changing—interpretations, it seems diffi-
Helsinki, Finland cult to capture context in any general sense that would
136

support design. Consequently, there have been doubts especially in the rapid change of contexts in everyday
about whether the entire concept is of any use [5, 7]. urban navigation. How are mobile contexts actively cre-
The demand for a new, empirical approach has been ated and upheld by people’s interactions with other
noted. For example, Dourish distinguishes between two people, with available technology, and with the outer
strands of empirical context-aware computing research. surroundings of their actions? In addition to the user-
The first is informed by research on physically based centeredness, this study emphasizes two points. First,
interaction and augmented environments. The second our interest is confined specifically to context changes
attempts ‘‘to develop interactive systems around occurring on the move in urban public and semi-public
understandings of the generally operative social pro- places (typically, somewhere between home, leisure
cesses surrounding everyday interaction’’ [4, p. 231]. activities, and work). In contrast to previous work,
The majority of empirical research falling under the which has concentrated on restricted areas such as
first strand has mainly been concerned with fixed indoor museums, offices, or university campuses, we are inter-
contexts (e.g. offices, meeting rooms, and lecture halls). ested in the interplay between dynamic context-changes,
Perhaps because such settings appear to be static, the moving people, and their actions. Second, unlike previ-
researchers have tried to create rigid taxonomies and ous research on mobility, e.g. [10, 11], we are specifically
general ‘‘all-embracing’’ definitions of context—with interested in the majority of people—the elderly, single
negligible success. Technological advances have largely mothers, and youngsters—instead of focussing on busi-
driven this branch of research. As a result, the viewpoint nesspeople or researchers as usual.
of the end-users has been ignored (see [9]). When research
has been carried out in a user-centred way (Dourish’s
second branch), it has focussed on different kinds of work 3.1 Background of this study
contexts and mobile workers. For example, Luff and
Heath’s [10] analysis of different kinds of mobilities and This study was part of a strategic design research project
their relations to collaborative work together with Perry that envisioned new concepts for selected target groups.
et al.’s [11] study of the everyday nature of mobile busi- Understanding their typical ways of moving around the
nesspeople’s work point out several problems and pos- city and providing general design ideas for context-
sibilities related to mobility and mobile contexts. aware mobile technology was one of its goals. Journeys
However, it seems that considerations on how context- in the city of Helsinki are typically done by public
aware mobile devices might make their place in mundane transportation, like in most European cities. In Helsinki,
actions of everyday life outside work settings have been the average annual amount of boardings per capita is
forgotten. The present study falls under the second line of 371, in Western Europe 202 [12]1. Thus, this study
empirical research; we try to understand particular analyses the activities taking place in present-day urban
interactional processes related to mobile contexts outside environments. We do not try to predict the future on the
work settings. As there is a rich variation with situated basis of present-day technology. Instead, we believe that
actions in mobile contexts, they do not lend themselves to an understanding of the present-day activities is neces-
rigid general definitions or static taxonomies. However, sary for gaining an insight into how future devices and
even if mobile contexts are created by particular actions applications could support, or even challenge, the pres-
performed while moving, they also contain regularities ent interactions in mobile contexts2. However, at this
that can be captured by context-aware devices. point, we are not suggesting any particular ideas for
interaction design. The aim is simply to point out some
general aspects relating to the urban navigation and
3 Our approach mobile contexts. These general aspects provide a way of
understanding the challenges and possibilities of the
Our starting point is that contexts are always determined user-centred (interaction) design that apply for different
by their specific use situation loaded with different ac- kinds of new context-aware devices and services.
tion resources: motives, plans, other people, mobile
computers, and the like. By analysing trivial actions, we
focus on the interactional and situated nature of mobile 4 Method
contexts. Thus, rather than trying to give a working
definition for mobile contexts, we believe that, by The data was gathered by ethnographic participant
explicating the subtle actions and resources by which observation and analysed from an ethnomethodological
people manage their mundane doings in particular mo- standpoint [13, 14]. Ethnomethodology is interested in
bile circumstances, we can gain insights on how mobile
contexts ‘‘get done’’ and to what extent these doings
1
could be proactively recognized by mobile and ubiqui- Although public transportation is used frequently in Europe, it is
tous devices. common in other parts of the world too. For example, in New
York, the same annual figure was 152 boardings per capita in the
This paper contributes to user-centred empirical at- years around 2000 [15].
tempts by trying to understand how context-aware 2
Discussions about this kind of ‘‘Technomethodology’’ can be
computing might make its place in mobile activities, found from [16].
 137

how people make sense of their social world, on how able context. A good example of such a nodal event is
everyday life gets ‘‘done’’ with the help of different the space claiming act (see Sect. 5.2). These nodal
trivial actions and resources that are available to people events, if reoccurring throughout the data and carried
in a given situation. out on the move by our subjects, were then analysed
Twenty-five adults (names changed in the following) further.
were observed while moving from place to place during
their normal days in the Helsinki metropolitan area.
Altogether, 25 people from five different user groups 5 Characteristics of mobile contexts
participated in our user research. They included the el-
derly, young singles, journalists, amateur actors, and We describe five characteristics of mobile contexts. The
middle-aged apartment house neighbours. All partici- characteristics can be seen as the sum of different
pants received a small monetary compensation. resources and actions by which the mobile contexts are
The study took place during the summer of 2001. situationally constructed and upheld. They were se-
Five researchers spent 1–3 days with each participant. lected on the basis of three criteria: (1) diagnosticity
Video cameras, digital cameras, and field notes were (specific to mobile contexts, but rare to fixed contexts),
used to document the observations. The focus of the (2) generality (recurring in many travel episodes), and
observations was on subjects’ everyday activities, espe- (3) concreteness (specific enough to be taken into
cially activities related to their urban everyday journeys account in the design of context-aware computers). We
(see [17]). For the analysis of the mobile context, present the characteristics as closely linked to the
observations were transcribed and represented in a constituting activities and illustrate them with exam-
notepad-like format. Photographs were presented on the ples.
left side of the notepad, and explanations of the pictures
and the storyline were on the other. The storylines were
in the form of a thick description, including all of the 5.1 Situational acts within planned ones
details that could be observed and documented. These
documents were then divided to travel episodes. A travel When moving, people usually have a mental plan on
episode consists of temporally organized (i.e. it has a how to navigate from place A to place B, and what
beginning, middle, and end) action patterns depicting a actions must be performed on the way in order to fulfil
meaningful journey between two places. ‘‘Meaningful- the plan. However, several actions can be performed in a
ness’’ here means that actions were performed in se- situational, ad hoc manner during the journeys. One
quences in order to carry out a culturally recognizable important aspect of mobile navigation seems to be that
doing. A special emphasis was given to finding nodal unplanned context changes lead to unplanned situa-
events (e.g. see [18]), that is, events where an action tional acts. Our participants often popped in somewhere
transformed the present context into another recogniz- or bumped into acquaintances on their way to their
primary destination (see Fig. 1a, b).
Suchman’s [19] notion that plans do not simply
Fig. 1 a Semi-planned sidestepping. Jane was on her way to a café determine action, but provide resources through which
to meet with her friend. She got off the tram in front of another café individuals organize their own actions and interpret the
that provided internet access and dropped in to read emails, and actions of others in certain situations, is useful here.
then carried on to meet her friend. b Popping in to a store. After People keep their main target in mind while simulta-
missing her bus, Anne was walking home from work. She noticed
some nice postcards in the bookstore window and decided to go neously doing something else, which is usually some
into the store to have a closer look at the cards. She realized that minor-scale activity. Thus, the pattern of the modern
she needed to buy one, since she was going to a party the following urban journey can be similar to those of Gladwin’s [20]
day. c Ad hoc. Unplanned sidestepping triggered by a change in native Polynesian navigators, who do not forget their
social context. Jane was riding in a tram on the way to visit her
friend. While talking on the phone, she noticed her friend, Albert, final target, even they are constantly reacting to their
stepping in. Jane finished the phone call and went to Albert. They immediate environment, such as the waves and winds of
started to chat about latest happenings the ocean.
138

Certain contexts enable people to perform actions paper, watching out of the window, talking on a mobile
that are significant only at that specific moment. These phone, or with similar actions. In our example (Fig. 2a),
actions do not necessarily replace the ‘‘main’’ plan; ra- free newspapers were read when travelling alone. With
ther, they are little sidesteps on the way to the goal. the newspaper, the reader claims a personal space [25]
However, the threshold for doing sidestepping has to be and signals off a private sphere in the midst of other
low if they are to be performed while still adhering to the passengers. The number of people in the group is an
main plan—otherwise the sidestep could become the important factor affecting how, and what kinds of spaces,
main plan. are claimed, as well as the actions carried out in the
Furthermore, unplanned acts are often social in created context. An example from our data shows a
nature. Sometimes, it takes time for people to recognize group of friends sitting around a table and sharing a
old acquaintances from a crowd. But after mutual rec- newspaper, turning their backs to other people, and, that
ognition, people usually stop to chat about their latest way, isolating from them. In this situation, a territorial
happenings—it would be impolite to act otherwise (see group space is claimed. The group in another example
Fig. 1c). takes a territorial group space by indicating that it will
not yield to others while roaming ahead (see Fig. 2b).
What is apparent in these examples is the way public
5.2 Claiming personal and group spaces places are transformed to personal and group spaces by
using the resources available. The free newspaper on
People need space for themselves and for the actions trams and the circling act in groups provide physical
they are about to take. For this reason, they create at resources to mark the place socially as a relevant space.
least two kinds of spaces around themselves; personal With these actions, the actor manifests themselves as a
and group spaces. According to several psychosocial participant or as an outsider of the ongoing social
studies, the upholding of personal space is a universal activity.
need—only the dimensions of this space are culturally
dependent [21]. These spaces both constitute and indi-
cate the nature of the social interaction at hand in a 5.3 Social solutions to problems in navigation
given situation. People regulate their involvement in
social interactions [22] by different situational resources Navigation in urban spaces is difficult for a number of
(e.g. by turning away from others in conversation). Even reasons. Maps are complicated and hard to remember,
if the phenomena of space claiming and privacy in social streets and buildings resemble each other, exact ad-
interactions have been recognized in the field of social dresses are difficult to keep in mind, and complicated
psychology decades ago, they have emerged as research bus routes are difficult to envision. These problems are
topics in the HCI area only recently [23, 24]. Spatial usually solved with interaction with other people. When
considerations, and space claiming in particular, are a people on the move come up against obstacles, or are
central issue in the study of mobile contexts, as we simply feeling unable to navigate their routes correctly,
demonstrate below. The space claiming in our everyday they often seek help via their social channels.
journeys was done in various ways. For example, space In our data, problems were solved with the help of a
claiming on a bus was achieved by picking up a news- mobile phone. For example, when people realized that a
bus had already gone, mobile phones were quickly
picked out of pockets. This does not suggest that people
Fig. 2 a Newspaper marking a boundary of personal space. Jane would have been totally lost in navigation, rather, the
was riding alone on a bus. All of the passengers were reading a telephone connection was invoked mainly for two rea-
similar free newspaper. So did Jane. There were no discussions sons: to announce that the schedule has changed, and to
among the passengers. bGathering in a circle as a sign of claiming
group space. Anne and Maija met Jaana at a metro station. They
negotiate what to do next (see Fig. 3). In our example,
got together in a circle to talk about the clothes that Jaana had missing the bus affected not only Kaarina’s and her
bought. They share a territorial group space children’s schedule, but also the schedule of her friend.
 139

Fig. 4 a Hurrying. Kaarina needed to run with her children to

catch a bus. Before running, they were in a fast-food restaurant and
Fig. 3 Missing a bus triggers social navigation. Kaarina was on her the children didn’t eat fast enough. b Waiting. Maikku was
way to meet a friend with her children. They missed the bus they swimming with her grandson. She rushed to the bus stop in order to
had planned to take and she called her friend in order to ask when go home in time, but after checking the timetable, she noticed that
the next bus leaving. She preferred to call even though the timetable she would have to wait for the next bus. After a while, she decided
was in front of her to call her son (the father of the grandson) to let him know that
everything went well in the swimming hall
This episode underlines the complex relationship be-
tween navigational procedures and possible disruptions
in the socially organized and co-ordinated schedules. their current doings to fit in a time frame—the planned
Disruptions in these actions require re-negotiation of action is not unfolding as anticipated, and more ‘‘do-
both the next navigational action as well as shared ings’’ must be squeezed into a same amount of time
schedules. Generally, Kaarina’s action in the situation than in a normal situation to keep up with the original
shows orientation towards social awareness: she man- plan. When moving in a hurry, physical and social
ages her relationship to the other person indirectly in- surroundings change rapidly, but attention is directed
volved in her journey. In future, this mundane, but still mainly to space (e.g. shortest route) and time (e.g.
rather complex, re-negotiation procedure could proba- monitoring time). Amidst human and inhuman moving
bly be recognized and assisted by proactive devices and actors, finding the fastest route becomes important
services. (Fig. 4) and co-ordinating the hastening requires all of
the attention of the person. Activities that need con-
stant monitoring involve route selection, checking
5.4 Temporal tensions timetables, informing related persons, and anticipating
changes in the surrounding environment.
Time plays a crucial role when moving through urban Sometimes, people simply hurry too much, which
areas. It has been argued that mobile devices free results in waiting. When waiting, the relationship be-
people from the limitations of time and place. Our tween time and action is stretched. Temporal aspects of
data, in agreement with previous work, e.g. [26], firmly the situation are again a main orientation in the situa-
disagrees with this. In fact, time and place are over- tion as an anticipated action is about to happen (Fig. 5).
emphasized in mobile contexts. Fluctuations in the The waiting time can then be made use of by various
importance of time and space as contextual factors are time-killing activities (e.g. by calling somebody, see
called temporal tensions here. Temporal tensions refer Figs. 3 and 4b).
to situations where time becomes problematic in rela-
tion to the action at hand and where, at the same time,
the temporal aspect of a situation is actively used to 5.5 Multitasking
orient action (making temporal tensions in actions
visible). Analytically, temporal tensions could be clas- The fact that navigating through an urban environment
sified into four stages: acceleration, normal (antici- requires paying constant attention to surroundings
pated) proceeding, slowing down, and stopping. Some means that attentional resources available for interact-
situations get accelerated so that multiple tasks should ing with a device are limited. The navigator must ac-
be performed more or less simultaneously, and some tively position and reposition him/herself in the social
prepared tasks may become impossible at that very context of other people and physical objects. Social
moment. They can be carried out only if events unfold conventions reduce some of the cognitive load in navi-
as anticipated. Sometimes, everyday life gets ‘‘slower’’, gating through a mass of other people (e.g. in Finland,
or even ‘‘stops’’—for example, because of a suddenly people use the right side of the pavement), but not all
cancelled appointment or a misunderstood timetable of actions can be as streamlined and may require socially
a public transport vehicle. The two important tensions coordinated sequencing—turntaking—with other people
emerging from our data are hurrying and waiting. to achieve the goal (e.g. turntaking is required among
When hurrying, people are actively orienting to the people whose lines of walk are crossing). Furthermore,
temporal aspects of their actions. They are managing finding the way from the starting point to the goal
140

Fig. 5 Temporal tensions:

acceleration, deceleration,
hurrying, normal (or balanced),
and waiting. The horizontal axis
denotes objective time in the
task and the vertical bars show
task-related actions

requires monitoring performance and reasoning whether capturing this nodal event (since it requires eye gaze and
the goal is getting closer or not. hands).
Navigating while moving obviously restricts multi-
tasking. As discussed in Sect. 5.1, people seem to have a
longer-term plan of navigation (e.g. ‘‘to go and visit 6 Design implications
parents’’) divided into sub-goals (e.g. ‘‘turn left at the
next corner’’). Therefore, monitoring the environment to The purpose of the following suggestions is not to
notice whether a sub-goal has been reached requires advocate a certain specific technological platform or
attentional resources from other tasks. The requirement infrastructure, but to direct designers’ attention to
for attention is higher in situations involving more higher-level issues in mobility and mobile contexts. The
uncertainty, such as when getting closer to a signal that suggestions below fall into four broad categories: navi-
indicates the fulfilment of a sub-goal (e.g. ‘‘the street gation, mobile communication, mobile games, and
corner’’). Another attention-demanding task is manip- general user interface issues.
ulating or creating sub-goals ‘‘on the move’’ (e.g. infer-
ring the shortest route when coming to a street-crossing).
These are likely reasons why our participants tended to 6.1 Navigation
have less multitasking while moving than whilst waiting.
It is also known that navigation in an urban environ- Monitoring sidestepping The phenomenon of occasional
ment places heavy demands on working memory, the sidestepping, for example, reading one’s emails in an In-
visuo-spatial resources in particular [27]. However, when ternet café on the way to meeting friends (Fig. 1a), creates
the route is familiar and navigation is more automatic, a challenge for route-guidance systems. These systems
and working memory is not as taxed, more multitasking could monitor and learn information about recurring
can be carried out. An extreme example is given in sidestep destinations (and activities performed therein).
Fig. 6. This information could be utilized, for example, in a busy-
While waiting, people tend to engage only in the kind alerting service that would watch a person’s timetable and
of multitasking that does not hinder them from noticing warn if the ongoing process might make him or her busy in
the signal in the environment that indicates the end of the next phase. Thus, understanding temporal tensions
waiting. This signal constitutes a nodal event that must caused by sidestepping would be important in creating
be noticed. For example, making a call with a mobile this kind of service. This is also relevant from the view-
phone while waiting for a bus does not interfere with the point of affective computing [28], as temporal tensions are
demand of seeing the bus as it comes closer. In contrast, usually accompanied by changes in the emotional moods
writing an SMS message or email would interfere with of the users (e.g. frustration caused by missing the bus).

Fig. 6a–e Mobile multitasking.

a After her work shift, Jane
took a tram to visit to her
friend. On the tram, she
received several phone calls.
b She continued talking while
stepping out from the tram,
crossing the street, walking to a
kiosk, c grabbing some money
out from her bag, d actually
buying the tobacco, and
e trying to manipulate
malfunctioning door bell
buttons. We noticed that
moving and waiting pose
somewhat different demands on
how the environment on the
one hand, and the task itself on
the other, are attended
 141

Adapting to sidestepping The notion of situational acts consider not only one person’s doings, but also the do-
within planned acts implies that it is important to allow ings of other relevant people. Simple implementation of
users to choose or create a route that is not necessarily this idea would mean, for example, crosschecking of a
the shortest or fastest, but otherwise rewarding. Side- user’s schedule or conversation history with one of their
stepping is relevant because travelling should not be friends and to compare it to their past or current phys-
understood only as moving from place A to place B. ical location. On the basis of these crosschecks, the
Collected information about sidestepping could be used systems could provide information to the user about
to support more flexible, customized routes that are not their social contexts and networks of communication
based only on minimizing distance. Ideally, routes (social awareness), and also use this knowledge to pro-
would represent possibilities for beneficial sidestepping actively infer the user’s next action. Another imple-
such as popping into a store (Fig. 1b). On the other mentation could consider the fact that, at the beginning
hand, sidestepping can be explicitly supported by visu- of mobile phone conversations, it is very common to
alizing the resources available and close to the route (e.g. both ask about and describe one’s own contextual situ-
gas stations for drivers). Sidestepping calls for adapta- ation [29, 30]. Context-aware technology could provide
tion and flexibility on the part of the system that aims to additional representational tools for the presentation of
support navigation. The system may need to re-adjust its user environments (e.g. what kind of resources are
navigation suggestions after sidestepping from the pro- present in the present or remote context, how easily
posed optimal route has occurred, as is done in many of available other social contacts are, or how likely they are
the current in-car route-guidance systems. to know an answer to the problem). During the tem-
poral tension of hurrying (see Fig. 4a), contextual
Predicting navigation problems Agents could benefit information would be useful in preventing an unneces-
from observations of common problems in navigation. sary waste of time caused by calling a busy or inacces-
When facing an obstacle, such as not finding the way to sible person. Prototypes of contextual information of
a meeting place or missing a bus, people tend to find a others communicated to and represented in a mobile
social solution. In our data, problems were solved, for device do exist (e.g. [31]), but field tests have not fo-
example, by calling to the friend (Fig. 3). Navigation cussed on the aspect of what contextual information
could be supported by a digital or live agent as in route- would be useful for a mobile person to know about
guide services (direct social navigation), or by a repre- another mobile person. Wizard of Oz studies in the
sentation of a history of solutions other people have mobile context could be carried out in the future, similar
made in the same situation (indirect social navigation). to the study performed by Hudson et al. [32] in the office
Proactive agents are already in use in some in-car nav- environment to address this question.
igation devices that create a direct connection to a live
person, who advises the driver about the route, or even Recognition of personal and group spaces The obser-
gives advice in the case of a car accident. Another vation that people actively create personal and group
example illustrating this kind of service could be a mo- spaces (see Fig. 2a, b) points out that issues of space
bile agent, triggered by a bus station, that would be claiming and privacy in social interaction are important
available to people who have missed their bus. in mobile contexts. Many of the markers of changes in
personal and group space could be used as starting
Enhanced awareness of navigation-related changes in re- points in the recognition of new mobile contexts. Space
mote contexts To address the problems of waiting for claiming indicates the nature of the interaction at hand
environmental signals (Fig. 4b) that indicate an impor- in relation to other people situated in the vicinity. Thus,
tant forthcoming context change, we suggest devices context-aware devices could recognize acts of space
that ‘‘boost’’ these signals. For example, the vibration of claiming because they are important indicators of the
a mobile device could indicate that one’s bus is user’s activity and willingness to use different types of
approaching the bus stop. This service would free re- services (e.g. single/multi-user). It would also be fruitful
sources for other, potentially more interesting, activities. to consider services that could extend the claiming of
Our analysis of multitasking implies that it is important personal or group space to the (shared) virtual space. At
that interacting with the device does not interfere with the present moment, there are prototype services [33]
the noticing of the signal in the environment that indi- that can be used to overlap physical and virtual worlds,
cates the transformation from one context (e.g. waiting) and to turn a virtual space into a location-dependent
to another (e.g. getting on the bus). social place. In addition, virtually shared group spaces
pose challenges for the design of shared displays and of
privacy in interaction.
6.2 Social awareness in mobile communications
Representation of nearby associates Ad hoc meetings
Communication of context information As humans are among acquaintances (Fig. 1c) could be arranged easily
inherently social beings, our actions are always directly (also in virtual space) if people on the same route, or on
or indirectly linked to other people (e.g. Fig. 3). Con- routes close to another, were aware of each other’s
text-aware technology should, thus, be designed to movements (e.g. [31]).
142

6.3 Mobile games mobile situation, the best predictors may differ. Tem-
poral tensions, such as acceleration, make certain kinds
Mixing reality with the virtual world It is possible that, of information important (e.g. estimations of speed),
in the future, mobile users can be connected to a mul- and other information irrelevant. Moreover, as side-
tiplayer online game whilst moving around in the real stepping introduces breaks and pauses into the use of the
world. Such games could utilize the real world contexts system, its user interface must incorporate effective
presented in this paper in at least three ways. First, a reminders of the stage of the interrupted task.
group of people sharing a group space in the real world
could be given more ‘‘power’’ in the game than those not
in the space, or the people in the same group space could
be given a bonus task to be solved together. Second, 7 Conclusions
people playing a game featuring challenges or problem
could be given a possibility to call and ask for help from If context-aware devices and services are to find their
their friends in the real world to solve the problem. And way into everyday mobile life, careful reconsideration of
third, temporal tensions in the real world could be part the analytical level and role of ‘‘context’’ in the studies of
of the game. For example, when the player stops, the technology and human action should be taken into ac-
game world could become more demanding, or it could count. Social acceptability of new context-aware tech-
create tensions in the real world by providing time-lim- nologies is dependent on how well they fit into the
ited tasks that are connected to certain routes in the real routinely carried out mundane processes of everyday
world (e.g. catching a certain person or bus). life. The results of our study describe interactional social
and psychological restrictions and resources in mobility:
how situational and planned acts intermesh in naviga-
tion, how people construct personal and group spaces,
6.4 General user interface issues and how temporal tensions develop and dissolve.
The results of our data are geo-culturally bound to
Modality selection Our participants rarely just Helsinki, Finland. The everyday journeys made by the
‘‘walked’’ or ‘‘waited’’ but instead, engaged in multiple selected target groups could look different outside Scan-
activities simultaneously (Fig. 5). Multitasking naturally dinavia, or Western Europe in general. Thus, the study
gives rise to multi-modal interfaces. Traditional PDAs, could be replicated in other cultural settings to assess the
for example, require both hands and visual attention to generalizability of our findings. However, even with these
operate, which is clearly inappropriate for mobile con- reservations, the viewpoint and analysis of everyday
texts in which some modalities are preserved for other journeys presented in this paper demonstrate one way of
tasks. On the other hand, nomadic user interfaces dealing fruitfully with the concept of mobile context
(designed for interaction while walking) might be too within the HCI tradition, and of incorporating the find-
clumsy and awkward for situations where all modalities ings into the design process as design implications. The
are available, such as during a longer waiting period. studies of mobile settings based on an empirical, human-
Context-sensitive selection of modalities can be achieved centred approach offer interesting possibilities that allow
by recognizing prototypical ‘‘modes’’ of mobility, such us to see the future of context-awareness in a new and
as walking, waiting, hurrying, navigating, or sidestep- innovative light, and offer a rich source of innovation for
ping, and choosing the interaction channels according to new context-aware services and for the design of under-
the resources that are typically free in that mode. Recent lying context-recognition mechanisms.
developments in context management and recognition in
mobile devices point out that this may not be a distant Acknowledgments We thank Salla Hari, Sauli Tiitta, Tomi Kan-
dream (e.g. [34]). kainen, and Esko Kurvinen, and our industrial partners Alma
Media, Elisa Communications, Nokia, Sonera, and SWelcom. The
second author acknowledges a grant from the Academy of Finland.
Interruption management Mobile computing devices
that filter the information flow (e.g. incoming emails,
SMSs, phone calls, etc.) need to select and intensify task- References
relevant information, and inhibit task-irrelevant infor-
mation. From the point of view of the user, irrelevant 1. Dey AK, Abowd GD, Salber D (2001) A conceptual frame-
information proves to be an interruption, no matter how work and a toolkit for supporting the rapid prototyping of
context-aware applications. Hum-Comput Interact 16:97–166
elegantly it is conveyed. We propose that markers of 2. Bellotti V, Edwards K (2001) Intelligibility and accountability:
context-change also provide indirect indications of the human considerations in context-aware systems. Hum-Comput
interruptability of the person. Space claiming acts, for Interact 16:193–212
example, signify a transformation of context to where 3. Dix A, Rodden T, Davies N, Trevor J, Friday A, Palfreyman K
certain information is attended to (e.g. newspaper), and (2000) Exploiting space and location as a design framework for
interactive mobile systems. ACM Trans Comput-Hum Interact
other information is ignored. In the office environment, 7:285–321
talking to another person in the last 15 s is a very good 4. Dourish P (2001) Seeking a foundation for context-aware
predictor of the interruptability of a person [32], but in a computing. Hum-Comput Interact 16:229–241
 143

5. Greenberg S (2001) Context as a dynamic construct. Hum- 21. Bonnes M, Secchiaroli G (1995) Environmental psychology. A
Comput Interact 16:257–268 psycho-social introduction. Sage, London
6. Schmidt A, Beigl M, Gellersen HW (1999) There is more to 22. Altman I (1975) The environment and social behaviour: pri-
context than location. Comput Graph 23:893–901 vacy, personal space, territoriality and crowding. Brooks/Cole,
7. Svaneas D (2001) Context-aware technology: a phenomeno- Monterey
logical perspective. Hum-Comput Interact 16:379–400 23. Agre PE (1998) Introduction. In: Agre PE, Rotenberg M (eds)
8. ISO 13407 (1999) Human-centred design processes for inter- Technology and privacy: the New Landscape. MIT press,
active systems. International Standard, the International Cambridge, pp 1–28
Organization for Standardization 24. Palen L, Dourish P (2003) Unpacking ‘‘Privacy’’ for a net-
9. Chen G, Kotz K (2000) A survey of context-aware mobile worked world. In: Proceedings of the conference on human
computing research. Technical report, department of Computer factors in computing systems (CHI 2003), Fort Lauderdale,
Science, Dartmouth College, Dartmouth Florida, April 2003, pp 129–136
10. Luff P, Heath C (1998) Mobility in collaboration. In: Pro- 25. Czarnowski TV (1978) The street as a communications artifact.
ceedings of the CSCW ’98 conference on computer-supported In: Anderson S (ed) On streets. MIT, Cambridge, pp 207–211
cooperative work. ACM Press, New York, pp 305–314 26. Kopomaa T (2000) The city in your pocket: birth of the mobile
11. Perry M, O’Hara K, Sellen A, Brown B, Harper R (2001) information society. Gaudeamus Kirja, Helsinki
Dealing with mobility: understanding access anytime, any- 27. Garden S, Cornoldi C, Logie RH (2002) Visuo-spatial working
where. ACM Trans Comput-Hum Interact 8:323–347 memory in navigation. Appl Cognitive Psych 16:35–50
12. City of Helsinki urban facts (link checked 20.2.2004): http:// 28. Picard J, Rosalind W (1998) Toward agents that recognize
www.hel.fi/tietokeskus/tilastoja/taskutilasto/TaskutilastoEng- emotion. In: Actes proceedings IMAGINA, pp 153–165
lanti2003_250603.pdf 29. Laurier E (2001) Why people say where they are during phone
13. Garfinkel H (1967) Studies in ethnomethodology. Prentice- calls? Environ Plann D 19:485–504
Hall, New Jersey 30. Weilenmann A (2003) ‘‘I can’t talk now, I’m in a fitting room’’:
14. Wolcott H (1995) The art of fieldwork. AltaMira Press, Walnut availability and location in mobile phone conversations. J
Creek Environ Plann 9:1589–1605
15. Urban Transport Fact Book: http://www.publicpurpose.com/ 31. Nakanishi Y, Kumazawa S, Tsuji T, Hakozaki K (2003) iC-
ut-us96pcr.htm AMS2: developing a mobile communication tool using location
16. Dourish P, Button G (1998) On ‘‘Technomethodology’’: information and schedule information with J2ME. In: Pro-
foundational relationships between ethnomethodology and ceedings of mobile HCI 2003, Udine, Italy, pp 400–404
system design. Hum-Comput Interact 13:395–432 32. Hudson SE, Fogarty J, Atkeson CG, Avrahami D, Forlizzi J,
17. Kankainen A, Oulasvirta A (2002) Design ideas for everyday Kiesler S, Lee JC, Yang J (2003) Predicting human interrupt-
mobile and ubiquitous computing based on qualitative user ibility with sensors: a Wizard of Oz feasibility study. In: Pro-
data. In: Proceedings of the 7th ERCIM workshop for user ceedings of human factors in computing systems (CHI 2003),
interfaces for all. Lecture notes in computer science, Springer, pp 257–264
Berlin Heidelberg New York, pp 458–464 33. Espinoza F, Persson P, Sandin A, Nyström H, Cacciatore E,
18. Schilit BN, Adams N, Want R (1994) Context-aware comput- Bylund M (2001) GeoNotes: social and navigational aspects of
ing applications. In: Proceedings of the IEEE workshop on location-based information systems. In: Proceedings of Ubi-
mobile computing systems and applications, Santa Cruz. IEEE comp 2001: ubiquitous computing, international conference,
Press, Piscataway, pp 85–90 Atlanta, Georgia. Springer, Berlin Heidelberg New York, pp 2–
19. Suchman LA (1987) Plans and situated actions. The problem of 17
human–machine communication. Cambridge University Press, 34. Korpipää P, Mäntyjärvi J, Kela J, Keränen H, Malm E-J
Cambridge (2003) Managing context information in mobile devices. IEEE
20. Gladwin T (1964) Culture and logical process. In: Goodenough Pervas Comput 2:42–51
WH (ed) Explorations in cultural anthropology: essays in
honor of George Peter Murdock. McGraw-Hill, New York