Mapping quality requirements for pervasive mobile games
Transcript of Mapping quality requirements for pervasive mobile games
ORIGINAL ARTICLE
Mapping quality requirements for pervasive mobile games
Luis Valente1,2• Bruno Feijo2
• Julio Cesar Sampaio do Prado Leite3
Received: 21 April 2014 / Accepted: 27 August 2015 / Published online: 14 September 2015
� Springer-Verlag London 2015
Abstract Games have not received the full attention of the
requirements engineering community. This scenario is
becoming more critical as we move towards newer forms of
games, such as pervasive games. Pervasiveness (the quality
that distinguishes pervasive games from traditional digital
games) holds several meanings, including being ubiquitous,
permeating something, or spreading something, somewhere,
in a physical space. Pervasiveness can be recognized in by
the boundaries of the game expanding every time it is played,
from the virtual (or fictional) world to the real world. Per-
vasive games are a new form of digital entertainment that has
evolved in different forms, such as alternate reality games,
transmedia games, and crossmedia games. Sensor technolo-
gies, networking capabilities, augmented reality systems,
computer vision technology, the internet, and, especially,
mobile devices have been responsible for the rapid evolution
of this new form of digital product. This paper is focused on
‘‘pervasive mobile games’’, which we define as context-
aware games that use mobile devices. We bear in mind that
mobile devices are currently the main driver for fulfilling the
promises of pervasive game playing. Our investigations and
experiments on this class of games led us to study the quality
requirements for pervasive mobile games. Using different
information sources, we gathered a set of interrelated char-
acteristics that are crucial to the success of these games. In
this paper, we begin to clarify the definition and scope of
pervasive mobile games, which are controversial issues in the
literature. Using these fundamentals, we propose a two-level
conceptual map of non-functional requirements that helps to
realize pervasiveness in pervasive mobile games. These non-
functional requirements are then associated with a set of
questions that help the designers in verifying tasks and
operationalizing the requirements of a game. We also pro-
pose a dependence matrix for pervasive game qualities that
enhances the insight into pervasiveness and reveals important
guidelines for the game designers.
Keywords Pervasive mobile games � Requirements
engineering � Non-functional requirements � Dependencematrix
1 Introduction
The problems of game software development have been
intensifying along with the growing demands of the video
game market—a market that is reported to reach $128
billion by 2017 [1]. Complexity is increasing, the quality
requirements are escalating, and the time to market has
been shortening. Although software engineering processes
have been applied to video game development, there are
few works on requirements engineering for video game
design [2]. Processes of video game requirements engi-
neering have unique characteristics when we compare them
& Luis Valente
Bruno Feijo
Julio Cesar Sampaio do Prado Leite
1 MediaLab/Institute of Computing, UFF, Av. Gal. Milton
Tavares de Souza, s/n8, Sao Domingos, Niteroi,
RJ 24210-346, Brazil
2 VisionLab/Department of Informatics, PUC-Rio, Rua
Marques de Sao Vicente 225, Gavea, Rio de Janeiro,
RJ 22451-900, Brazil
3 Department of Informatics, PUC-Rio, Rua Marques de Sao
Vicente 225, Gavea, Rio de Janeiro, RJ 22451-900, Brazil
123
Requirements Eng (2017) 22:137–165
DOI 10.1007/s00766-015-0238-y
to the processes found in the development of traditional
application software. Some of these characteristics are
discussed in academia and the industry, such as subjective
softgoals (e.g. gameplay, emotions, enjoyment, and fun),
the iterative nature of the requirements management, the
interdisciplinary nature of the development team, and the
fact that games are market-driven products [3–6]. This
whole situation becomes more critical when we move
towards mobile games [4]. In this paper, we focus on an
even less covered area: pervasive mobile games—a sub-
class of pervasive games where the main game device is a
mobile one.
Pervasive games are a recent form of entertainment that
brings the game experience out of the game device and into
the physical world, integrating both virtual and physical
realms. Early examples of pervasive games are Pirates! [7],
Pervasive Clue [8], and Botfighters [9]. The latter is a com-
mercial location-based mobile phone game. More recent
examples are Gbanga Famiglia [10] and Geo Hunters [11],
which are clearly classified as pervasive mobile games. Fig-
ure 1 illustrates Pervasive Word Search (Sect. 4), a pervasive
mobile game developed by the first author, where the goal is
to capture all letters of a word using sources that exist in the
physical world (like colours and Wi-Fi network names).
Historically, game development has been guided by
technological constraints, such as the design of algorithms
with the best possible optimizations regarding run-time and
memory performance. This had led to ad hoc development
that prioritized results over more sophisticated software
engineering methodologies [5].
The game development process is inherently multidis-
ciplinary, comprising activities as diverse as game design,
art production, and software development, among others.
These activities are integrated in two main phases: pre-
production and production. The preproduction phase con-
cerns the work of game designers and artists who produce
creative elements such as the game storyline, game rules,
and multimedia assets. The production phase refers to the
software development.
The integration of these teams that speak ‘‘different
languages’’ is an important (and complex) issue in game
development [3, 4]. The multidisciplinary nature of game
development often raises conflicts among the different
parties, as Callele et al. [3] note:
‘‘Game designers may not understand, for example,
the limitations of artificial intelligence when design-
ing non-player characters while software engineers
may not understand the creative vision or they may
be too willing to compromise that vision in the rush
to ship the product.’’
Callele et al. [3] also mention that many problems in
game development (including failure) arise from the lack
of adequate requirements engineering activities to bind the
preproduction and production phases. For example, a major
source of problems in game development processes relates
to ambitious scope and feature creep [12, 13]. Petrillo et al.
[12] also cite other problems, such as the cutting of features
during the development phase and delays or too optimistic
project schedules.
Requirements engineering in game development needs
to consider some issues that stem from the nature of games.
Games are applications that are designed for entertainment.
Traditional requirements engineering activities focus on
Fig. 1 Pervasive Word Search with the word ‘‘Reynold’’. The player captures a red colour from a red flower to eliminate the letters ‘‘r’’, ‘‘e’’,
and ‘‘d’’ from the target word
138 Requirements Eng (2017) 22:137–165
123
solving problems and supporting work activities, which are
different goals than those found in games.
By having entertainment as the main goal, games have
abstract non-functional requirements, such as fun, flow,
immersion, enjoyment, and delight [3–5]. These require-
ments are related to the emotional aspects that result from
playing a game. For this reason, some researchers refer to
these types of requirements as ‘‘emotional requirements’’
[3, 4], ‘‘psycho-chemical requirements’’ [5], and ‘‘affective
factors’’ [14]. More recently, Callele et al. [15] refine
‘‘emotional requirements’’ into ‘‘experience requirements’’,
which include emotional factors, gameplay factors, and
sensory factors. Apart from these recent advancements,
understanding these requirements in a requirements engi-
neering process is an open issue.
Another problem related to the definition of require-
ments in games stems from the fact that games are mass-
market products. This means that when designing a game,
the real users are unknown. This is similar to what Potts
[16] observed for the construction of commercial off the
shelf software (COTS) requirements; that is, there is no
clear interface between ‘‘customers’’ and ‘‘developers’’.
The game designer ends up acting as the ‘‘customer’’ in the
preproduction phase, which is not a good alternative as the
game designer may produce biased requirements (espe-
cially the ones related to emotional factors such as ‘‘fun’’).
In this regard, the game designer needs support from sev-
eral confirmation techniques, such as focus groups [17].
In this paper, we do not start investigating how sub-
jective softgoals (gameplay, emotions, enjoyment, and fun)
of mass-market products could be modelled and measured.
Instead, we are interested in learning about what makes a
game pervasive—unique qualities and characteristics that
make pervasive games different from traditional digital
games. In this regard, we search for non-functional
requirements (NFRs) that map quality requirements related
to pervasiveness for pervasive mobile games. The perva-
sive mobile games are a subset of pervasive games that use
context-aware mobile devices to enable the interaction
between the environment (physical world) and the virtual
world. This map is based on concrete information sources,
such as books, academic articles, game analyses, and
hands-on experience. Our starting point is an analysis of
the definition and scope of pervasive games, which we
claim contributes to and sheds more light on these currently
controversial issues. Then, we come up with two levels of
non-functional requirements, which are associated with a
set of questions that help the designers in the tasks of
verifying and operationalizing the requirements of a game.
Additionally, we propose a dependence matrix for perva-
sive game qualities that enhances the insight into perva-
siveness and reveals important guidelines for the game
designers. As far as we are aware, there is no other work on
requirements engineering for pervasive mobile games in
the literature. We hope that our proposals can help the
industry to produce better games. Our research work is
pushed by three questions:
R1 What are pervasive games?
R2 What are the differences between pervasive games
and traditional computer games?
R3 How do we make a pervasive mobile game?
We define the results of a requirements construction
process applied to pervasive mobile games as ‘‘a knowl-
edge mapping for pervasive mobile games’’ or ‘‘quality
model for pervasive mobile games’’. Based on the ideas of
Arango&s domain analysis [18], we have performed this
process on the field of pervasive mobile games to answer
questions R1 and R2. Arango [18] defined domain analysis
as ‘‘the identification and acquisition of reusable informa-
tion in a problem domain to be reused in software speci-
fication and construction’’. Domain analysis is part of a
domain engineering process, which Arango [18] defined
originally as a process that concerns producing a reusable
infrastructure to solve problems in a domain.
In the area of software requirements, the term ‘‘non-
functional requirements’’ has been defined differently by
different authors, as surveyed by Glinz [19]. Although
there are discussions about ‘‘non-functional requirements’’
as ‘‘just quality-driven requirements’’ and about ‘‘non-
functional requirements’’ as ‘‘general constraints over
requirements’’, it is a fact that in general quality issues are
treated as non-functional requirements. In this paper, we
use these terms as synonyms: ‘‘non-functional require-
ments’’, ‘‘quality requirements’’, and ‘‘quality’’.
Referring to ‘‘quality attributes’’ is a common practice in
the software architecture area [20]. In software engineer-
ing, Freeman [21] distinguished between basic qualities
(functionality, reliability, ease of use, economy, and safety)
and extra qualities (flexibility, reparability, adaptability,
understandability, documentation, and enhanceability).
Interestingly enough, Freeman [21] defines ‘‘functionality’’
(the cornerstone of software construction) as a ‘‘quality’’.
The remaining qualities have been grouped together as
‘‘non-functional requirements’’, which denotes other qual-
ities in addition to software ‘‘functionality’’.
This paper is organized as follows. Section 2 summa-
rizes the answer to the question R1, while Sect. 3 decom-
poses the qualities of pervasiveness that differentiate
pervasive games from traditional computer games, where
we contribute to partially elucidate questions R2 and R3.
More specifically, Sect. 3 describes a map about pervasive
games knowledge, which includes the process that we
conducted to build this knowledge. This map comprises the
following parts: our vision on pervasive mobile games as a
subset of pervasive mobile games (Sect. 3.1), the
Requirements Eng (2017) 22:137–165 139
123
conceptual map with non-functional requirements
(Sect. 3.2), set of questions to be used to verify and oper-
ationalize the requirements (i.e. the checklists, in
Sect. 3.3), and the dependence matrix of pervasive game
qualities (Sect. 3.4). In Sect. 4, we use the proposed map to
describe how the pervasiveness qualities apply to a work-
ing pervasive mobile game developed by the authors.
Section 5 discusses related works in the literature. Finally,
Sect. 6 concludes the paper and presents some directions
for future works.
2 What pervasive games are: a critical review
While conducting research about pervasive games, we
noticed lack of consensus on definitions and formalisms
about these types of games. We noticed that this field is
highly interdisciplinary and that the approaches to perva-
sive games can be loosely divided into two broad groups:
‘‘cultural’’ and ‘‘technological’’ approaches.
In our interpretation, we use the term ‘‘cultural
approach’’ to refer to approaches originated in the artistic
and social areas, such as design, game studies, and social
studies. The authors in cultural approaches tend to use the
term ‘‘pervasive’’ in its original literal sense (e.g.
‘‘spreading widely throughout an area or a group of peo-
ple’’ [22]) to indicate that some game aspects defy or
expand concepts that are considered central in traditional
game definitions (such as ‘‘the magic circle’’ [23]). Cultural
approaches are more concerned with aspects of gameplay
and the game itself, not emphasizing technology and per-
vasive computing. As an example, for some researchers in
this area [24]) a pervasive game may not use (computing)
technology at all. In these approaches, the concept of
‘‘pervasive games’’ can be very broad, encompassing
simple mobile phone games to ‘‘artistic events’’ with
complex infrastructure.
We use the term ‘‘technological approaches’’ to denote
approaches that place more emphasis on the underlying
technological aspects needed to realize pervasive games.
These approaches often discuss pervasive games through
these ideas: (1) as applications of pervasive/ubiquitous
computing, or context-aware applications; and (2) ‘‘com-
puter-augmented’’ games.
During the literature review, we noticed a set of recur-
ring themes that some authors based upon to discuss per-
vasive games. In this section, we use the term
‘‘perspective’’ to denote a specific theme or set of themes
that underlies a pervasive game discussion. Table 1 sum-
marizes the main perspectives that we identified in the
pervasive game literature.
The work by Nieuwdorp [25] was the first attempt at
clarifying the confusion in the pervasive game discourse.
She used a dichotomy between ‘‘game studies’’ and
‘‘technology’’ to analyse the discussion on pervasive game
research. We agree with Nieuwdorp [25] that the two
perspectives proposed by her (‘‘cultural’’ and ‘‘technol-
ogy’’) are the most appropriate ones to define the basic
approaches to pervasive games. Nieuwdorp [25] explains
that the technological perspective use ‘‘pervasive’’ to
denote ‘‘the technology the games are played with’’, while
the cultural perspective uses ‘‘pervasive’’ ‘‘to describe the
way the game itself (due to this technology) is played in a
real-world setting, and as such blends with it’’.
The ‘‘computer-augmented’’ game perspective considers
the notion of existing real-world (physical) games (e.g.
‘‘capture the flag’’ games) being ‘‘upgraded’’ with some
sort of computing. Magerkurth and co-authors [28] go
further in this regard and consider as ‘‘pervasive games’’
what they categorized as smart toys, affective gaming,
augmented table-top games, and augmented reality games.
This idea implies the existence of a ‘‘mixed-reality’’—the
result of interleaving a virtual layer over a physical world.
We understand that mixed reality is an essential charac-
teristic of pervasive games. Furthermore, this characteristic
is a continuum in the sense presented by Milgram and co-
authors [36], who propose a ‘‘reality-virtuality continuum’’
to represent the varying degrees of complexity that any
mixed-reality situation can exhibit, as Fig. 2 illustrates. We
consider the ‘‘computer-augmented game’’ vision broader
than the ones based on pervasive and ubiquitous computing
concepts (as it includes ‘‘toys’’), although considering
‘‘toys’’ as ‘‘games’’ is up to debate.
In the ‘‘cultural, theatre, technology’’ perspective,
McGonigal [30, 37] propose a broad domain of ‘‘pervasive
play and performance’’ for ‘‘ubiquitous games’’, ‘‘ubicomp
games’’, and ‘‘pervasive games’’. McGonigal [37] defines
‘‘pervasive play’’ as ‘‘consisting of mixed-reality games
that use mobile, ubiquitous and embedded digital tech-
nologies to create virtual playing fields in everyday
spaces’’. She argues that the ‘‘performance’’ aspect enables
the players to maximize their play experience by
Table 1 Perspectives on pervasive games
Perspective Main references
Cultural and technological
dichotomy
Nieuwdorp [25]
Cultural or game studies
approach
Montola et al. [24], Montola et al.
[26], Davies [27]
‘‘Computer-augmented’’ game Schneider and Kortuem [8],
Magerkurth et al. [28], Saarenpaa
[29]
Cultural, theatre, technology McGonigal [30]
Technology, ubiquitous and
pervasive computing,
sensors
Magerkurth et al. [28], Linner et al.
[31], Capra et al. [32], Walther
[33], Hinske [34], Benford [35]
140 Requirements Eng (2017) 22:137–165
123
performing instead of just participating in the game, thus
becoming ‘‘game actors’’. She defines ‘‘pervasive games’’
as ‘‘performance-based interventions that use game ima-
gery to disrupt the normative conventions of public spaces
and private technologies’’ [30], while ‘‘ubicomp games’’
are advanced game prototypes (possibly with custom
hardware) for research into new ubiquitous computing
technologies. Finally, she considers ‘‘ubiquitous games’’ as
commercial entertainment projects that bring computer
games to the real world.
The ‘‘technology, ubiquitous and pervasive computing,
sensors’’ perspective focuses on issues related to context-
aware applications, mainly considering the games as
applications of ubiquitous and pervasive computing. The
perspectives that focus on these kinds of technology often
refer to the concepts of ‘‘pervasive/ubiquitous computing’’
and ‘‘context-awareness’’ with different meanings, which
can be rather confusing. In pervasive games research these
three concepts (pervasive computing, ubiquitous comput-
ing, and context-aware systems) are often correlated.
Nieuwdorp [25] presents a comprehensive analysis on how
researchers use these terms in pervasive game discussion.
Our analysis on these kinds of perspective is as follows.
Ubiquitous computing is a concept that Weiser [38]
proposed in 1991 for the next wave of computing, when
technology would recede into the background of people’s
lives. In this paradigm, computing devices would be
embedded (and hidden) in environments, adapting their
behaviour to human activities. The devices would be small,
inexpensive, networked, and context-aware. The users
would not have to ‘‘think about computers’’ when using
ubiquitous computing applications. Seven years later, IBM
came up with the term ‘‘pervasive computing’’. Their
vision was the possibility to access information and ser-
vices, anytime, anywhere, focusing on ‘‘e-business’’ [39].
Their focus was on e-business and on the corporate world,
and they related ‘‘pervasive computing’’ to mobile com-
puting and wireless technologies. The original pervasive
computing concept relates to the original sense of the word
‘‘pervasive’’ (as found in Oxford English Dictionary [22]).
After considering all issues above mentioned, we have
the following understanding about what a pervasive game
is. Firstly, we understand that ubiquitous computing
applications use context-awareness to implement their
vision in some environment (called the ‘‘smart environ-
ment’’ by ubiquitous computing researchers). This idea
may seem that in ubiquitous computing, the ‘‘smart envi-
ronment’’ is a place of very restricted in size (like a small
room). We consider that a system is ‘‘context-aware’’ if ‘‘it
uses context to provide relevant information and/or ser-
vices to the user, where relevancy depends on the user’s
task’’, as wrote Dey [40]. Dey [40] defines context as fol-
lows: ‘‘any information that can be used to characterize the
situation of an entity. An entity is a person, place, or object
that is considered relevant to the interaction between a user
and an application, including the user and applications
themselves’’. Then, we realize that ‘‘pervasive computing’’
is a way to expand the ubiquitous computing concept to
larger environments and spaces. In this context, pervasive
computing emerges as the key technology to create the
mixed-reality situation found in pervasive games. There-
fore, we are tempted to adopt a technological perspective to
define a pervasive game as a context-aware game system
that has the following characteristics:
• The game is constantly coming back to the real world,
which means that the game is played in physical places
and it is not constrained to stationary computing
devices;
• The physical world (places, objects) is part of the game,
and it is combined with the virtual world;
• Mixed reality is always existent, and it is created
through pervasive computing technologies;
• The spatial mobility occurs in a physical ‘‘open’’
environment; that is, the ‘‘game world boundary’’ is not
‘‘well-defined’’, and sometimes it can be unconstrained;
• The players use mobile devices (e.g. smartphones,
tablets, custom hardware) to interact with the game and
with other players;
• The game may last for several days or weeks (or more),
blending with the daily lives of players. The game may
define a persistent world that progresses without player
intervention. If some important event happens in the
game, the game may notify the player to take action.
These aspects are not mandatory;
Fig. 2 The reality-virtuality
continuum. Extracted from [36]
Requirements Eng (2017) 22:137–165 141
123
• The game may focus on promoting social interaction
among players. Social interaction in a pervasive games
may happen directly (face-to-face interaction) or indi-
rectly (mediated through technology). This aspect is not
mandatory as pervasive games may be single-player
games.
Although this attempt to define pervasive games covers
the characteristics found in the majority of the definitions
analysed in our literature review, it is too broad to allow an
effective support to the game designer. Therefore, in the
next section, we narrow down our scope to pervasive
mobile games.
3 Building our map
In this section, we present the process we have used to
elicit, model, and analyse the knowledge related to quality
requirements for pervasive mobile games. To understand
what pervasive games and pervasive mobile games are,
firstly we had to select suitable information sources [41] to
start the elicitation process. The available information
sources were reading material (books and academic arti-
cles), analyses of game projects, and hands-on develop-
ment experience. The key information source for the
elicitation process was the reading material. The elicitation
process was pushed by these research questions outlined
formerly:
R1 What are pervasive games?
R2 What are the differences between pervasive games
and traditional computer games?
R3 How do we make a pervasive mobile game?
We started the elicitation process by searching for
specific journals and conferences dedicated to pervasive
games. We discovered no specific journals for pervasive
games. Conferences dedicated to pervasive games are
nearly non-existent, with PerGames (‘‘International
Workshop on Pervasive Gaming Applications’’,
2004–2007) being a rare example. In 2014, a new confer-
ence dedicated to pervasive games emerged also as ‘‘Per-
Games’’ (‘‘International Conference on Pervasive
Games’’), although it is not related to the former initiative.
Considering this situation, we explored the following areas:
game development, entertainment computing, and perva-
sive computing. An important information source that we
discovered was the IPerG project [42], which was a large
European project dedicated to studying pervasive games
that spanned from 2004 to 2008. The IPerG project pub-
lished several papers, technical reports, and pervasive
game prototypes. Besides the PerGames conferences and
IPerG project, we searched for relevant works in the main
publishers ACM, IEEE, and Springer, which included
proceedings of conferences dedicated to traditional games
and digital entertainment. Examples of these conferences
include ACE (ACM International Conference on Advances
in Computer Entertainment), ICEC (IFIP International
Conference on Entertainment Computing), and DiGRA
(International Conference on Digital Games Research
Association). We also searched for initiatives in non-aca-
demic domains because game developers usually share and
discuss the state-of-the-art information through informal
channels, such as discussion forums and game develop-
ment portals.
The elicitation process was based on keywords, such as:
pervasive games, ubiquitous computing, mixed reality,
games, context-awareness, context-aware games, mobile
game development, and location-based games. We selected
and rejected papers firstly by evaluating the paper title. If
the title seemed important to our research work, we eval-
uated the abstract. If the paper abstract suggested that the
paper could answer our research questions, we kept the
paper for further reading. Otherwise, we did a quick read to
check whether the paper was really not important to discard
it with more confidence. We were especially interested in
papers that described actual pervasive game projects.
While conducting the process to map pervasive game
knowledge, we discovered that the literature lacks con-
sensus on definitions and formalisms about pervasive
games. The literature also mixes up issues related to
diverse areas, such as game design and computer science.
Moreover, we realized that most of the pervasive games
definitions are excessively broad to support a quality
requirements mapping for pervasive games. With these
initial findings, we understood that we needed to narrow
our scope to find answers to our three research questions.
We were also interested in developing pervasive games
with mobile devices from the very beginning. In this
regard, we established criteria to define ‘‘pervasive mobile
games’’—a set of the possible pervasive games found in
the literature, which roughly correspond to context-aware
games that use mobile devices. Section 3.1 discusses in
detail our vision on ‘‘pervasive mobile games’’, and
Table 2 summarizes the defining criteria to select game
projects according to our vision.
Pervasive games and pervasive mobile games have
unique characteristics that distinguish these kinds of games
from traditional computer and mobile games. We consider
that these characteristics contribute to an overall quality of
‘‘pervasiveness’’. Pervasiveness is a quality that makes
pervasive mobile games unique and very different from
traditional computer-based games. According to our
investigation, pervasiveness seems to be an implied and
taken for granted concept in pervasive game discussion. It
is very difficult to define ‘‘pervasiveness’’ objectively
142 Requirements Eng (2017) 22:137–165
123
because it is an elusive and subtle concept. As we were
interested in developing pervasive mobile games, we
noticed that understanding characteristics that contribute to
in these kinds of games (i.e. ‘‘what makes a game perva-
sive’’) could be more practical in fulfilling our goals of
creating pervasive mobile games than trying to model
pervasiveness itself in a top-down approach. With this idea
in mind, we started investigating the pervasiveness quality
in pervasive mobile games using the following concrete
information sources:
• Twenty-four game projects, summarized in Table 3;
• Books and academic articles related to pervasive
games, which were retrieved from the literature using
these keywords: pervasive games, ubiquitous comput-
ing, mixed reality, games, context-awareness, context-
aware games, mobile game development, and location-
based games;
• The mobile game development experience of the first
two authors, which includes developing games using
embedded sensors.
Using the criteria defined in Table 2, we selected 24
projects to analyse. Table 3 presents these games. We
included games that are not part of the domain and some
borderline cases. Due to space constraints, the reader
should refer to [43] for descriptions and main references
regarding the games in Table 3. Valente and Feijo [43] also
provide information about platforms, sensor, and net-
working technologies that these games use (when
applicable).
The ‘‘borderline cases’’ in Table 2 refer to ‘‘transmedia
pervasive games’’ and games that use other media and
devices as secondary (support) modules. We discuss these
issues in Sects. 3.1.2 and 3.1.3. The game Day of Figurines
is a pervasive game based on mobile devices. However, it
is not part of the domain because it does not satisfy the
mandatory condition C1 in Table 2. Although this game is
out of the domain, it presents some qualities (related to
pervasiveness) that other pervasive mobile games share. In
order to map the knowledge about pervasive mobile games,
we have followed a process composed of 10 steps, which
we detail below:
1. Contextualizing: we started a literature review to
understand what pervasive games were all about;
2. Eliciting from the literature (an information source):
we became overwhelmed and confused about what to
consider as ‘‘pervasive games’’ because we discov-
ered several perspectives from which to approach this
subject. Therefore, we decided to establish criteria to
select suitable games that matched our interests
(develop pervasive games using context-aware
mobile devices). We named the games that matched
these criteria as ‘‘pervasive mobile games’’
(Sect. 3.1);
3. Eliciting from the literature (an information source),
but using feedback from step 2: we started another
literature review, now focused on selecting games
that matched our criteria for pervasive mobile
games.
4. Eliciting from the game experience (an information
source): we started analysing the selected games with
these questions in mind: (1) ‘‘what features does this
game have that makes it a pervasive game?’’; and (2)
‘‘how is this game different from traditional computer
games?’’. Intuitively, this analysis was influenced by
the experience of the first two authors in researching
and developing games;
5. Mapping or modelling the knowledge (conceptual
maps): we gathered all of the material that we have
collected and started to refine it. We were able to
identify an initial list of 16 qualities that pervasive
mobile games have in common. At that time, we had a
single set of first-level qualities for pervasiveness.
Later (step 6), we understood that they were actually
second-level qualities—Fig. 3 illustrates them;
6. Mapping (conceptual maps): while analysing these
qualities, we ‘‘tagged’’ them using labels that meant an
Table 2 Boundary criteria for pervasive mobile games
Condition Description
C1* Games that are context-aware
C2* Games using mobile devices (e.g. tablets, smartphones)
C3 Games that access remote data on the move
C4 Multi-player games
The asterisk denotes mandatory condition
Table 3 Games selected according to the criteria summarized in
Table 2
Condition Games
Out of the
domain
Day of figurines
Borderline
cases
Can you see me now?, Uncle Roy all around you,
Mogi*, Epidemic Menace, Botfighters*, GPS
Mission*, Manhattan Story Mashup
1 and 2 The Journey*, Ere Be Dragons, Pervasive Clue,
REXplorer*
1 and 2 and 3 –
1 and 2 and 4 Pirates!, PAC-LAN, Feeding Yoshi
All conditions Songs of North, Tycoon, Hitchers, Mythical: The
Mobile Awakening, Seek n0 Spell* , Gbanga
Famiglia*, Gigaputt*, GEO Hunters*, Insectopia,
and borderline cases
The asterisk denotes commercial projects
Requirements Eng (2017) 22:137–165 143
123
attempt at classifying these qualities into groups. Later,
we refined these tags and came up with an initial set of
six qualities (‘‘first-level qualities’’) that grouped
second-level qualities. At a later step (step 10), we
have refined this initial set of first-level qualities and
arrived at the results that Fig. 3 illustrates;
7. Analysis (verification): for each second-level quality,
we identified motivations in the literature to support
each one. In other words, for each quality we collected
use cases in game projects, design rationales, and other
examples that justified the existence of the referred
quality;
8. Mapping (checklists): we used the motivations and
examples identified in step 7 to elaborate checklists for
each second-level quality.
9. Analysis (verification) and mapping (matrix): while
analysing the second-level qualities (by checking the
literature and comparing our results), we noticed that
some qualities could influence or interfere with other
qualities. We recorded these dependencies as a
matrix;
10. Refining the results: we submitted all results (the first-
and second-level qualities, checklists, and depen-
dence relationships) to a requirements engineer expert
for feedback about the process that we conducted to
build our knowledge about pervasive mobile games.
Based on the feedback we received, we refined these
results.
Next subsections present our vision on pervasive mobile
games (Sect. 3.1), the conceptual map (Sect. 3.2), the
checklist (Sect. 3.3), and the dependence matrix, which
represents correlations among qualities (Sect. 3.4).
3.1 What are pervasive mobile games?
We realize that mobile devices are currently the main
driver to fulfil the promises of pervasive game playing
because they are naturally networked, full of sensors,
widespread, and easily accessible. In this context, we
define a ‘‘pervasive mobile game’’ as a game that is played
in the physical world, where players use context-aware
mobile devices to enable the interaction between the
environment (physical world) and the virtual world. To
support mixed-reality situations, we understand that per-
vasive mobile games must be ‘‘context-aware’’. We follow
the definitions created by Dey [40] regarding the concepts
of ‘‘context’’ and ‘‘context-awareness’’.
We use the term ‘‘mobile device’’ for a device that has
the following characteristics:
1. They have networking capabilities that are not limited
to a specific physical place. Examples include mobile
phones and tablets that use networking infrastructure
provided by a mobile operator;
2. They are portable devices—battery-powered devices
that users are able to carry around;
A ‘‘context-aware mobile device’’ is a ‘‘mobile device’’
equipped with sensors that enable it to sense ‘‘relevant
context’’ [40], especially environmental properties.
By using context-aware mobile devices as elements of a
mixed-reality situation, pervasive mobile games are able to
Fig. 3 Non-functional requirements view on pervasiveness for pervasive mobile games
144 Requirements Eng (2017) 22:137–165
123
have game activities out in open areas or at least not
confined to the static nature of using desktop computers. A
consequence is that these games may require or foster
player movement in local spaces (not necessarily using
network resources on the move).
We see pervasive mobile games as having three general
types, regarding player participation:
• Single-player. This corresponds to a game played with
mobile devices that does not use network resources;
• Single-player connected. This is a single-player game
that uses network resources;
• Multi-player. This is a context-aware game with
multiple players that are connected through a network.
The players can be co-located or not.
Referring to the defining criteria for pervasive mobile
games (Table 2), ‘‘single-player’’ games satisfy conditions
(C1, C2), ‘‘single-player connected’’ games satisfy condi-
tions (C1, C2, C3), and ‘‘multi-player’’ games satisfy all
conditions. Currently, we consider smartphones and tablets
as the main platform for pervasive mobile games because:
• These devices are easily accessible to the general
public;
• These devices are equipped with embedded sensors that
make it possible to implement (some) context-aware-
ness required by pervasive mobile games;
• These devices make it possible to have mobility in
pervasive mobile games (e.g. gameplay in physical
places using networking or not).
We do not consider ‘‘portable’’ or ‘‘mobile’’ games as
pervasive mobile games. Both types of games are played
with mobile devices. However, a ‘‘portable game’’ is not
context-aware and does not use networking. A ‘‘mobile
game’’ requires networking, but it is not context-aware.
The concepts we have developed so far are not enough
to support a proposal of quality requirements for perva-
sive mobile games. Therefore, we have to develop a
number of other concepts to underlie a conceptual map
for quality requirements related to pervasiveness. The
rest of this section is dedicated to these new concepts,
which are: pervasive experience, transmediality, game
scene complexity, and game autonomy. For simplicity,
henceforth we will use the terms ‘‘mobile device’’ and
‘‘context-aware mobile device’’ as synonyms, mainly
referring to ‘‘smartphones and tables’’, unless otherwise
indicated. We expect that in the near future new classes
of context-aware mobile devices become important for
pervasive mobile games. Examples include wearable
devices such as ‘‘smart glasses’’, ‘‘smart bands’’, and
‘‘smart watches’’.
3.1.1 The pervasive experience
We define ‘‘pervasive experience’’ as the experience that
arises from playing the pervasive mobile game. This defi-
nition is the starting point to understand pervasiveness.
Moreover, pervasive experience can be used as an infor-
mation source in knowledge elicitation.
The pervasive experience occurs in mixed-reality envi-
ronments. As a consequence, physical spaces are always
present in pervasive experiences, as physical spaces are
required to build a mixed-reality environment. Further-
more, the pervasive experience is essentially ‘‘local’’
because this type of experience emerges when the player is
playing the game in some place.
We understand that ‘‘pervasiveness’’ is an inherent
quality of a pervasive experience. However, trying to
define pervasiveness is a difficult task because this is an
abstract concept. Instead of trying to define pervasiveness,
we understand that it is more practical to investigate the
qualities that contribute to create pervasiveness. In
Sect. 3.2, we present a conceptual map of pervasiveness in
terms of non-functional requirements that contribute to this
quality.
3.1.2 Pervasive games and transmediality
There are some pervasive games (e.g. Epidemic Menace
[44]) that offer several modes of participation. In these
games, players are able to use different types of devices
(such as mobile phones and desktop PCs) to perform dif-
ferent types of tasks in a complementary way. Players
usually have distinct roles and experience different envi-
ronments. Some researchers [44, 45] have used the term
‘‘crossmedia games’’ to refer to these types of games,
referring to each mode of participation as a ‘‘gaming
interface’’. In this concept, a game interface (or game role)
is part of a greater whole (the game). However, we con-
sider ‘‘transmedia pervasive game’’ a more adequate term
to refer to those types of games, as we understand as
‘‘transmediality’’ the possibility of a game offering differ-
ent participation modes (game roles) through different
devices, media, or environments. In contrast, we consider
as ‘‘crossmedia games’’ the games that can be played on
different (heterogeneous) devices but offer only one mode
of participation.
Transmedia pervasive games may offer some game
interfaces (or participation modes) based on context-aware
mobile devices, as do Can you see me now? [46] and Mogi
[47]. These game interfaces share the same concerns as the
ones found in general pervasive mobile games. In this
paper, we consider these kinds of games as ‘‘borderline
Requirements Eng (2017) 22:137–165 145
123
pervasive mobile games’’—we refer to the quality of
transmediality for these games, but we are not concerned
about general forms of transmedia pervasive games. In
other words, in this paper we present a conceptual map for
pervasiveness that concerns only the mode of participation
that is carried out through the mobile device. We have
identified some general cases of pervasive mobile games
concerning the concepts presented in this section:
1. The game offers just one game interface, based on
context-aware mobile devices. This includes games
that are available for multiple mobile platforms, such
as Android and iOS. In this case, there is no
transmediality;
2. The game offers several game interfaces, some of them
based on context-aware mobile devices and some of
them based on other kinds of devices (e.g. portable
devices, laptops, desktop computers). Each game
interface is tied to a specific type of device. In other
words, a game interface designed for a mobile device
cannot be played with a laptop computer. These are the
‘‘borderline pervasive mobile games’’;
3. The game offers several game interfaces, all of them
based on context-aware mobile devices. The mobile
device can be used with any of the available game
interfaces. In this case, there is no transmediality;
4. The game offers game interfaces (based on context-
aware mobile devices) and offers secondary interfaces
(based on other kinds of device or not) for support
purposes. For example, in Botfighters [9] players use a
web module (on desktop PCs) to configure their game
avatar, while the actual gameplay is carried out on a
mobile phone. In these cases, we consider that there is
no transmediality.
Referring to the conceptual map of pervasiveness that
Sect. 3.2 presents, transmediality contributes to a quality of
pervasiveness called ‘‘accessibility’’.
3.1.3 Game scene complexity and game autonomy
The integration of the physical and virtual worlds through
context-awareness in pervasive games creates a mixed-re-
ality environment that we refer to as game scene. The game
scene may have various degrees of complexity depending
on the number and types of systems and objects that
compose the mixed-reality environment. We identify three
general cases of increasing game scene complexity:
GSC1 The game scene content is generated dynamically
solely though sensors. Our prototypes Pervasive
Word Search (Sect. 4) and Insectopia [48] are
examples of this case;
GSC2 The game scene uses predetermined context. For
example, the game might require a player to reach
specific physical locations (e.g. a square in a city)
to complete game activities. Examples are Uncle
Roy all around you [49] and PAC-LAN [50]. This
also includes requiring a specific physical place
for the game scene (e.g. a museum, a school
floor);
GSC3 The game scene requires some dedicated and
customized infrastructure to support the mixed-
reality. For example, the infrastructure may take
form as local networking, tracking technologies,
placing sensor technologies in the physical
environment, and placing environment devices in
the game area. This case encompasses several
subcases of different complexities regarding the
required infrastructure and required environment
devices
Environment devices are physical objects that reside in
the physical environment where the game happens, and
they may be able to generate effects or behaviours in the
physical world. Some possibilities for environment devices
are (this is by no means an exhaustive list):
1. Simple output devices, not limited to digital media
(e.g. public displays, audio speakers, smell generators,
weather effects, heat generators, and smoke
generators);
2. Smart objects (input only);
3. Smart objects (input and output);
4. Mechanical or electrical devices equipped with actu-
ators (e.g. sliding doors, elevators, wind generators).
Smart objects are custom computing devices equipped
with sensors and may have networking and output capa-
bilities (not limited to digital media). These devices may be
connected to other smart objects and to a central entity (e.g.
a server) that runs the game. Smart objects may be sta-
tionary or moving.
The literature provides some examples of using objects in
similar ways as the environmental devices that we described
in this section, concerning pervasive games and interactive
installations. The ALICE project [51] creates mixed-reality
environments using public displays and mechanical devices
equipped with actuators. In Manhattan Story Mashup [52],
the game uses a public display to convey information to
players. The Fun-in-Numbers project [53] presents several
interactive installations that use physical objects to com-
municate with players. An example is Magnetize Words
[54], which uses light sources as output devices. Montola
and co-authors [55] present a game using wireless sensor
devices, which could be considered as a type of smart object.
146 Requirements Eng (2017) 22:137–165
123
Game autonomy refers to what degree the game is able
to exist as an independent system. Game autonomy is
directly related to the game scene complexity, in the sense
that a game with high autonomy requires a low degree of
scene complexity. Two general classes of pervasive mobile
games arise from the concept of game autonomy: ‘‘event
games’’ and ‘‘autonomous games’’. Autonomous games are
able to function independent of other systems and require
no specific places and time to happen, presenting high
game autonomy. On the other hand, event games are per-
vasive games that happen at a specific time and physical
place, for a specific duration, presenting low game auton-
omy. These games may require specific infrastructure. We
can say that these two classes of games represent the
opposite sides of a game autonomy spectrum. Event games
usually share some of the following characteristics:
• The game may happen in a predetermined area. For
example, this area could be a room, a museum, or some
specific city area;
• The physical place where the game takes place is
augmented with technology (as sensors) that enables
important game features. This requires the place to be
prepared before game sessions;
• The game may use human actors as live non-player
characters;
• The game may require orchestration, which means
providing real-time in-game support to help players in
difficulty and to prevent issues from impairing the
game session. More information about this concept can
be found in [56];
• The game may require custom hardware;
• The game may require local wireless networking;
• The game may have a specific time to start and has a
specific duration;
• They may use references to local physical context (e.g.
specific addresses, statues, buildings) as part of the
game.
The game scene complexity also impacts the networking
requirements of a pervasive mobile game. For example,
games that fall in cases GSC1 and GSC2 may require no
networking. Games that fall in case GSC3 require net-
working in most cases. Referring to the conceptual map of
pervasiveness that Sect. 3.2 presents, game autonomy
contributes to a quality of pervasiveness called ‘‘context-
awareness’’. In its current version, this conceptual map
concerns games in cases GSC1, GSC2, and some cases of
GSC3.
3.2 The conceptual map
Figure 3 illustrates a two-level conceptual view on the
pervasiveness quality for pervasive mobile games,
according to our investigation. After studying the 24
selected games (Table 3) and exploring the mobile game
development experience of the first two authors (that
includes games using embedded sensors), we were able to
identify sixteen pervasive game qualities (the second-level
qualities in Fig. 3). Later on, we grouped these sixteen
qualities into the seven first-level NFRs (non-functional
requirements) that Fig. 3 illustrates. Each non-functional
requirement contributes to create pervasiveness, but none
of them in isolation is able to describe the entire perva-
siveness quality. In Fig. 3, we have defined acronyms that
appear in the checklist tables (Sect. 3.3) and in the
dependence matrix (Sect. 3.4, which presents correlations
among qualities).
These results are our first attempt at identifying impor-
tant qualities of pervasiveness. We do not claim that these
results are complete—a more definitive pervasive quality
list would require further investigation about design theory
and additional inspections using more games and
designers.
3.3 The checklists
The checklists are a set of questions that designers can use
to evaluate pervasive game qualities (and thus, pervasive-
ness) in existing projects. We have formulated the check-
lists related to each NFR using two information sources:
1. Observing how the pervasive game qualities were
applied in the pervasive mobile games that we have
studied;
2. The experience of the first two authors in developing
context-aware mobile games.
We were able to identify seven first-level non-functional
requirements that contribute to create pervasiveness: spa-
tiality, sociality, permanence, accessibility, context-
awareness, communicability, and resilience. Each non-
functional requirement contributes to create pervasiveness,
but none of them in isolation is able to describe the entire
pervasiveness quality. Figure 3 illustrates these first-level
qualities along with the second-level pervasive game
qualities that contribute to them. This section describes
each of these non-functional requirements.
The checklists for each quality have questions of dif-
ferent natures. There are questions that ask whether an
aspect of the quality is applicable (‘‘do’’ and ‘‘are’’ type of
questions) and that are questions that ask ‘‘how’’ some
aspect of the quality applies or not (i.e. ‘‘how’’, ‘‘what’’,
and ‘‘which’’ questions). The ‘‘how’’ questions implicitly
ask whether the aspect is present or not (i.e. these questions
include an implicit ‘‘do’’ or ‘‘are’’ question), which means
that they only make sense if the implicit ‘‘do’’ or ‘‘are’’
question is true. The ‘‘how’’ questions invite the designer
Requirements Eng (2017) 22:137–165 147
123
or developer to investigate how that aspect is present or to
formulate a solution to have that aspect present in a game.
3.3.1 Spatiality (Spa)
Spatiality refers to aspects related to the physical space
usage. For example, it includes how the game defines the
‘‘game scene’’, how the game influences the physical
space, and how the physical space influences the game.
Mobility (Mob, Table 4) This quality relates to two gen-
eral aspects. The first one is mobile computing, which means
‘‘the capacity of moving computing services with us, through
reducing the size of devices, providing network access while
moving, or both’’ [57]. The second aspect refers to the
requirements of movement in a game. This second aspect
may be motivated by design decisions such as promoting
physical health, tourism games (where players would visit
some specific places), and technological constraints (e.g.
limit playing areas to places where location technology
works well). The amount of movement in pervasive games
may vary greatly, ranging from restricted small places (e.g.
Pirates! [7]), dedicated (controlled) parts of a city (e.g. Uncle
Roy all around you [49], PAC-LAN [50]), and unbounded
areas (e.g. Botfighters [9], Insectopia [48], Mogi [47], Gi-
gaputt [58], Gbanga Famiglia [10]). Another important issue
related to Mobility in mixed-reality environments refers to
how the game keeps the consistency between the physical
location of the player and its virtual representation (i.e. how
the rest of the system sees this virtual representation,
including other players). Inconsistency among physical
locations and their virtual representation affects the game
experience negatively, especially in multi-player games.
Local space redefinition (LSR, Table 5) This quality
relates to how the game is able to change the meaning of
the places where the game is played. By ‘‘changing
meaning’’, we denote augmenting value to places, incor-
porating the place (or objects belonging to the place) as
game objects, integrating live (human) non-player charac-
ters, or making the players perceive the place with alter-
native views. This is different from just ‘‘being at a place’’,
or using some property (like location) without referring to
the local context.
3.3.2 Communicability (Com)
Communicability refers to aspects related to how the game
is able to communicate with players and other game
components (as smart objects).
Connectivity (Con, Table 6) Connectivity may exist in
global or local scope. Global scope refers to connecting to
remote peers that are not in the same physical place. Local
scope refers to connecting peers that are co-located. The
connectivity scope directly relates to the space scope of the
game. Global connectivity makes it possible for activities
to happen in very different places, possibly very distant. In
this case, the game space scope can be large. Local con-
nectivity restricts activities to happen in places with more
restricted size. This quality influences all pervasive game
qualities where obtaining remote information may be
required.
3.3.3 Permanence (Perm)
This quality refers to the game as defining a persistent
world. A further degree of permanence would be a game
that possibly evolves by itself over time without player
participation. Other examples of concerns related to this
quality are persistence issues and interleaving game
activities with other non-game activities.
Daily life interleaving (DLI, Table 7) This quality
relates to how the pervasive game is able to become diffuse
through daily life, enabling the player to integrate gaming
sessions with other non-game activities.
Persistency (Per, Table 8) Persistency refers to main-
taining game state to be accessed through different game
sessions over time. This quality is a requirement for games
that aim at simulating a parallel world that evolve by itself.
This includes all games that have social networking or
community aspects. However, not all pervasive games
require persistency.
Table 4 Checklist for MobilityNo. Question
Mob1 Does the game need to use networking while the players are moving?
Mob2 If the game needs networking, how does the game handle networking connectivity issues while
players are on the go?
Mob3 Does the game depend on specific locations to be played? Are players required to move between
specific locations?
Mob4 Does the game require players to walk long distances? What is the order of magnitude of the
potential game area?
Mob5 How does the game keep physical player locations consistent within the game world?
148 Requirements Eng (2017) 22:137–165
123
3.3.4 Accessibility (Acc)
Accessibility refers to how players are able to access the
game. It relates to the game reach. In this sense, we con-
sider ‘‘accessibility’’ as having a wider scope than the usual
definitions that focus on social issues. We consider acces-
sibility as concerning three groups of issues:
• Social issues (e.g. people with disabilities, gender-
biased content);
• Economic issues (e.g. expensive devices, services and
technology);
• Technological issues (e.g. technical constraints, tech-
nology not widespread).
Access usability (Usa, Table 9) This quality corre-
sponds to traditional usability issues from human–com-
puter interaction focused on mobile devices and
accessibility. An example of an important issue related to
this quality is designing activities that do not require
players to constantly look at the mobile device screen.
This issue especially concerns pervasive games, as play-
ers might be moving during game activities. In this case,
looking at the device screen might risk their safety or may
make them miss part of the game experience. Some
authors refer to this approach as ‘‘design for ‘heads-up’
experience’’ or ‘‘head up game’’ [59, 60]. Other examples
of this concern can be found in [61–63]. Dixon et al. [64]
recommend simple interaction mechanics in game activ-
ities. We interpret this recommendation as a measure to
avoid having the player focus extensively on the game
device interface and thus missing part of the game
experience. Also, this issue can be related to using mul-
tiple modalities for interacting with the player. For
example, games may explore audio and haptic feedback,
as we did in [65, 66]. Another example comes from [67],
which use speech synthesis to keep players informed
about game events, reducing the need to look at the device
screen to continue playing the game.
Research on general usability issues in pervasive games
is still under-explored. The work by Jegers [68] is among
the few ones, where the results indicated that there are
general usability issues that are unique to pervasive games.
A related work is [69], which explored game usability
evaluation in pervasive games. Another example is the
work by Martins and co-authors [70], which explored
usability issues for a wearable interface designed for
‘‘ubiquitous games’’.
Game autonomy (GA, Table 10) This quality refers to
the degree that the game is able to work as an ‘‘independent
system’’. For example, a pervasive mobile game with
Table 5 Checklist for Local
space redefinitionNo. Question
LSR1 Does the game integrate physical places (or their elements) in the gameplay?
LSR2 Does the game give roles to physical places (or their elements) in the gameplay?
LSR3 Does the game help players to have alternative views of the physical environment?
LSR4 What physical resources does the game use as game elements?
LSR5 Does the game use ‘‘live non-player characters’’ (human actors) to reinforce
the mixed-reality overlay?
LSR6 How do game activities affect the physical world?
Table 6 Checklist for
ConnectivityNo. Question
Con1 What are the connectivity requirements for the game? Global? Local? None?
Con2 How does the game handle uncertainties related to connectivity?
Con3 What is the desired space scope for game activities?
Table 7 Checklist for Daily
life interleavingNo. Question
DLI1 Does the game provide a persistent game world?
DLI2 Are game activities designed to be interruptible?
DLI3 Do game activities require long time commitment from players?
DLI4 Which approaches or techniques does the game apply to blend with daily life?
DLI5 Does the game provide equal opportunities for playing in any time?
DLI6 How does the game communicate game-related events to players?
DLI7 Does the game world evolve by itself, while the player is not playing?
Requirements Eng (2017) 22:137–165 149
123
‘‘high autonomy’’ does not require supervision and prepa-
ration for game sessions. On the other hand, games with
‘‘low autonomy’’ can be considered as ‘‘event games’’
(Sect. 3.1.3).
Transmediality (TM, Table 11) Transmediality mainly
refers to providing different modes of participation (as
different game interfaces) in the game to deliver a broader
transmedia experience (Sect. 3.1.2). In this paper, we are
concerned with issues such as the number of different game
interfaces and the impact of using these different interfaces
in the game—in other words, if the game experience is
balanced among the various interfaces. However, the
qualities related to pervasiveness that we present in this
paper relate to game interfaces based only on context-
aware mobile devices. The pervasive mobile game may
also use smart objects to enrich the game experience.
Device independence (DI, Table 12) This quality relates
to the possibility of using the same (primary) game inter-
face (Sect. 3.1.2) in different mobile platforms.
3.3.5 Sociality (Soc)
Sociality refers to the social aspects and social implications
of the game. For example, it includes aspects such as pri-
vacy, ethical issues, how people relate with each other, and
how activities disturb non-players in public places.
Conformance to physical and social settings (CPS,
Table 13) This quality relates to ethical and privacy con-
cerns, conforming to social conventions, safety concerns
(e.g. protecting players from harm while moving around),
and adequacy to physical settings (e.g. relying on audio
feedback in noisy places).
Social communication (SC, Table 14) This quality
relates to how the game is able to foster or promote social
communication, acting as a medium for people to com-
municate. Social communication in pervasive games may
happen directly (face to face) or indirectly (using tech-
nology as a mediator). Also, the scope of communication
encompasses co-located and distributed people. Social
communication may generate emergent gameplay—com-
plex interaction patterns or behaviours that arise from
people playing the game. A game designer usually cannot
anticipate the effects that might emerge in the gameplay
dynamics, although the designer may foster emergent
gameplay through design decisions. For example, discus-
sion on emergent gameplay issues on general pervasive
games can be found in [71, 72]. A pervasive mobile game
may encourage social interaction by providing different
participation modes that require players to collaborate in
order to complete game activities.
Involving non-players (INP, Table 15) This quality
relates to integrating non-players into the game. Some
games might use this to create ambiguities about who is
playing and who is not, creating what Montola et al. [24]
have defined as ‘‘social expansion’’. This idea is also
related to players having to discover who the other players
are and meet them (physically). By doing so, players might
have to approach strangers on public places and then pro-
ceed to complete game activities. It is important to notice
that involving non-players in the game might raise ethical
issues.
Non-player participation can be passive or active. An
example of passive participation is our prototype Pervasive
Word Search (Sect. 4), which uses other people’s Blue-
tooth devices as a source of content. An example of active
Table 8 Checklist for
PersistencyNo. Question
Per1 Do game sessions depend on previous sessions or stored data?
Per2 Does the game support internal social networks or communities?
Per3 Does the game world evolve by itself, while the player is not playing?
Table 9 Checklist for Access
usabilityNo. Question
Usa1 Do game activities require players to focus on the device screen all the time?
Usa2 Do game activities use various modalities to interact with the player?
Table 10 Checklist for Game
autonomyNo. Question
GA1 Is the game bound to specific places or local context?
GA2 Does the game require configuring the physical space for a game session?
GA3 Does the game require any kind of supervision (orchestration) when players are playing it?
GA4 Does the game require custom hardware, human actors, or other kind of related resources?
150 Requirements Eng (2017) 22:137–165
123
participation is Uncle Roy all around you [49], which uses
actors to represent non-player characters.
3.3.6 Context-awareness (CA)
Context-awareness refers to aspects related to how the
game creates and maintains the mixed-reality environ-
ment—the integration between virtual and physical worlds
through exploring context.
Game content adaptability (GCA, Table 16) This quality
relates to context acquisition (through sensors) and
dynamic content generation. Game content adaptability
refers to how a game is able to keep the same (or expected)
gameplay functionality despite favourable or unfavourable
environment conditions. For example, a game that uses Wi-
Fi access point data to generate content should be designed
to work correctly in areas with lots of Wi-Fi access points
(as dense urban areas) as well as in areas with poor Wi-Fi
coverage. Otherwise, the experience that the game aims at
providing may be wanting in some situations.
The accessibility concept referred in question CGA3 is
the same as defined by the ‘‘accessibility’’ non-functional
requirement (Sect. 3.3.4).
Game object tangibility (GOT, Table 17) This quality
relates to how the game uses the mobile device (and other
environment elements) as tangible objects [74]. Ishii and
Ulmer [74] define ‘‘tangible object’’ as a metaphor for
‘‘human–machine interfaces that couples digital informa-
tion to physical objects’’. The users manipulate the physical
object to interact with the application. For example, on a
general context, tangible objects could be pieces of aug-
mented table-top games [28]. Existing research on perva-
sive games suggest that this feature contributes to
increasing player immersion in the game [45, 52, 75].
When considering mobile devices, this quality refers to
aspects such as assigning a purpose for the device (e.g. a
specific role) in the game and having players manipulate
the device according to this purpose to interact with the
game. An example of tangible interaction comes from the
mobile game The Audio Flashlight [65, 66] (although not a
Table 11 Checklist for
TransmedialityNo. Question
TM1 Does the game offer different participation modes (game roles)?
TM2 If the game offers different modes of participation, are these roles played through
different media (devices)?
TM3 Does the game balance the game experience for the various game roles?
TM4 Does the game need to use other media (e.g. desktop PCs) as support modules?
TM5 Is this game part of a broader transmedia experience?
Table 12 Checklist for Device
independenceNo. Question
DI1 Is it possible to use the game interface across multiple mobile device platforms?
Table 13 Checklist for
Conformance to physical and
social settings
No. Question
CPS1 How does the game handle player privacy?
CPS2 Do game activities possibly disturb non-players?
CPS3 Do game activities expose players (or non-players) to embarrassing situations?
CPS4 Do game activities conform to local social conventions/etiquette?
CPS5 Are game activities adequate to the physical setting of the game?
Table 14 Checklist for Social
communicationNo. Question
SC1 How does the game use technology as means to improve communication among people?
SC2 How does the game transform the relationships among players?
SC3 Does the game stimulate players to approach/start interactions with other people?
SC4 Does the game use technology to foster communities or social networks?
SC5 Does the game present emergent gameplay?
Requirements Eng (2017) 22:137–165 151
123
pervasive game) transforms the mobile phone into an
‘‘audio radar’’. The player interacts with the game using
gestures (i.e. tilting the device) according to the volume
and tempo of music. The game uses the device
accelerometer and vibration motors to create this interac-
tion interface. Another example comes from the pervasive
mobile game REXPlorer [76], where the mobile phone is
disguised as a game object through plastic shells to create
the ‘‘The REXplorer ‘detector’’’. The physical object
designed for a specific role in the game is called ‘‘a game
prop’’, which is a term inspired by theatre terminology.
3.3.7 Resilience (Res)
Resilience refers to how the game is able to prevent the
pervasive experience from breaking apart despite the
existence of technological issues that might disrupt the
pervasive experience.
Uncertainty handling policy (UHP, Table 18) This
quality relates to how the game handles noticeable
boundaries, breaks, or gaps among technology compo-
nents, which are known as ‘‘seams’’. This also includes
handling breaks in the smoothness of user experience, due
to inherent technology limitations in precision, accuracy,
availability, and other uncertainties. Some researchers [46,
61] have identified five general strategies to handle these
issues:
• Remove: consists of designing activities so that limi-
tations never appear in the game. This includes using
improved technologies (which is not always possible)
or designing activities that fit the technology limitations
into them;
• Hide: anticipates issues and ‘‘corrects’’ them before the
player has a chance to see them. Contrary to the remove
strategy, in this case the limitations appear in the game,
but are ‘‘corrected’’ before the player notices them;
• Manage: consists of including fallback modes to use
when the primary mode of operation fails. In other
words, the game adapts to the circumstances by having
several modes of operation;
• Reveal: presents the limitations to users and letting
them decide how to act. For example, mobile phones
display the operator signal strength in the user
interface;
• Exploit: means acknowledging the existence of issues
and integrating them into the game as a game design
feature.
The case study in Sect. 4 provides examples on using
the hide strategy.
Game pacing (GP, Table 19) This quality relates to how
technology influences or limits the pacing of the game (or
vice versa). For example, if game activities have high pac-
ing, some sensor technologies might not work well [61, 73].
Some technologies may not provide responses in a real-time
fashion (e.g. Bluetooth device scanning) that might be
required by some games. In those cases, it is necessary to
design game activities that are compatible with the inherent
characteristics of the technology that the game uses. These
issues mainly concern networking and sensor technologies.
3.4 The dependence matrix
During our investigation on pervasive game qualities, we
found several dependencies among them, which clearly
reveal an embedded structure. In this paper, we propose a
matrix, called the dependence matrix, to describe this
structure. Firstly, we present the dependence matrix as a
general instrument (Sect. 3.4.1), although we give exam-
ples with quality requirements for pervasive mobile games.
Afterwards we go into details about the use of this matrix
for the specific case of pervasive mobile games
(Sect. 3.4.2). In this matrix, we kept the terms ‘‘quality’’
Table 15 Checklist for
Involving non-playersNo. Question
INP1 Does the game involve non-players? How does it do it?
INP2 Does the game have activities where players need to find out who the other players are?
INP3 Does the game generate/use content that is based on other (non-player) people?
INP4 Does the game use human actors for non-player characters?
Table 16 Checklist for Game
content adaptabilityNo. Question
GCA1 How does technology availability affect the game content and the player experience?
GCA2 Does context usage require players to move more in the physical world?
GCA3 Does context usage affect accessibility?
GCA4 Are players able to play the game in various places without manual intervention/re-installation/re-
adaptation/orchestration?
152 Requirements Eng (2017) 22:137–165
123
and ‘‘quality requirements’’ to facilitate the reading of the
rest of the paper (although general terms would be more
appropriate, such as NFR or softgoals). In this section, we
use the acronyms defined in Fig. 3 (Sect. 3.2) to refer to the
pervasive game qualities.
3.4.1 The general case
We propose that dependencies among non-functional
quality requirements can be understood from the perspec-
tive of two complementary views:
• Operationalization view: this type of dependence arises
when the presence of a quality i is required by the
techniques used to operationalize the quality j. Depend-
ing on to the specific software being developed (e.g. a
specific game), this relationship can be a necessary or
an optional condition for operationalization. For exam-
ple, game content adaptability (GCA) is always
required by Game Autonomy (GA), regardless the
game being developed. On the other hand, depending
on the game being developed, Mobility (Mob) may be
required or not by Local Space Redefinition (LSR);
• Interdependence view: this type of dependence arises
when the presence of a quality helps or hinders another
one, regardless of the software being developed. For
example, game activities with intense game pacing
(GP) may hinder Connectivity (Con) and gathering
sensor information (GCA—game content adaptability)
due to limitations of the involved technologies. An
example of intense game pacing (GP) occurs when
players are moving fast while playing the game (e.g.
playing while sitting on a train or on a car).
In our proposal, the operationalization view has the
following basic relationship:
1. i R j: quality i is required by quality j;
The interdependence view has two types of basic
relationships:
1. i ? j: quality i helps quality j;
2. i - j: quality i hinders quality j.
Also we need some symbols of modality to express
necessity, possibility, and conditionality, which affect a
basic relationship:
• a: ‘‘always holds’’;
• m: ‘‘may hold, in some specific cases’’;
• d: ‘‘will hold if the intensity of the first parameter is
above some threshold’’. This modality applies only to
the help and hinder relationships.
For example, in the domain of pervasive mobile games,
we have: UHP aR GCA (i.e. UHP is always required by
GCA for the operationalization of GCA) and Mob mR LSR
(i.e. Mob may be required, in some cases, by LSR for the
operationalization of LSR), GOT ? LSR (i.e. GOT helps
LSR), Usa - Mob (i.e. Usa hinders Mob), GP m - TM
(i.e. GP may hinder TM in some cases), and GP d - GCA
(i.e. GP will hinder GCA if the intensity of GP is above
some threshold). We present other examples in Sect. 3.4.2.
Table 17 Checklist for Game object tangibility
No. Question
GOT1 Does the game specify a role for mobile devices in the game, or are they mere access terminals?
GOT2 How does the game transform mobile devices into game objects?
GOT3 Does the game use other physical objects (other than mobile devices) equipped with sensors or actuators as game objects?
GOT4 Does the game use mobile devices to create various game props (i.e. various game interfaces)?
GOT5 How does using tangible objects improve the socialization of players in a game?
Table 18 Checklist for
Uncertainty handling policyNo. Question
UHP1 Which strategy does the game use to handle technology limitations?
UHP2 How does the game handle technology limitations?
UHP3 Does technology correctly acknowledge physical world limits?
Table 19 Checklist for Game pacing
No. Question
GP1 Is the technology the games uses compatible with the game
pacing?
GP2 Do game activities adapt their pacing to accommodate
technology limitations?
GP3 Does the pacing of game activities require the player to focus
exclusively on the game? For how much time?
Requirements Eng (2017) 22:137–165 153
123
The operationalization and interdependence views in
this section concern implicit relationships—a kind of
relationship that is different than the one that Fig. 3 rep-
resents. Figure 3 represents explicit relationships among
qualities as level refinements—there are first-level NFRs
and second-level NFRs. Figure 3 does not present any
implicit relationship, such as the one that exists between
Game object tangibility (GOT) and Local space redefini-
tion (LSR) (e.g. GOT ? LSR). Figure 4 organizes these
implicit relationships (the operationalization and interde-
pendence views) in a dependence matrix. The dependence
matrix also represents the explicit relationships (the two-
level conceptual map defined in Fig. 3) as horizontal and
vertical headers (e.g. the first-level NFR Spa and its sec-
ond-level NFRs LSR and Mob). The first-level NFR names
are represented in bold italic font and the second-level
NFR names are represented in bold font. We analysed
several games (Table 3) and use our design experience to
define the matrix of Fig. 4. It is important to understand
that Fig. 4 represents the best of our current knowledge,
which is evolving considering the volume of information
we are dealing with.
In Fig. 4, we represent any dependence with a shaded
cell containing the type of relationship. The reader may
read the matrix in row-wise or column-wise direction.
Considering a matrix cell (i,j), where i is a row and j is a
column, reading in row-wise direction reveals straight in-
fluences, such as ‘‘i helps j’’, ‘‘i is required by j’’, and
‘‘i hinders j’’, while reading in column-wise direction
reveals dependencies, such as ‘‘j is helped by i’’, ‘‘j re-
quires i’’, and ‘‘j is hindered by i’’.
Our experience in building the dependence matrix
reveals that we should start filling the matrix by trying first
to find what each quality requires (i.e. column-wise
direction). The other two basic relationships (? and -) are
more naturally defined in row-wise direction. For the sake
of readability, we do not represent self-relationships (e.g.
LSR 9 LSR, Mob 9 Mob, GOT 9 GOT).
The proposed matrix allows a quick verification of
special interdependence relationships, such as the presence
of conflicting qualities revealed by symmetrical cells
containing the symbol -. This is the case of Access
usability (Usa) and Mobility (Mob) in Fig. 4, i.e. the matrix
reveals a challenge to be solved by designers (something
found in many other domains, where conflicting qualities
should be harmonized; such as the case of satisfying both
security and accessibility in a network system).
A visual inspection of the dependence matrix (Fig. 4)
reveals that some qualities are more influential or depen-
dent than others. For instance, UHP has a great influence
on others qualities, because the UHP row has a large
number of non-empty cells. Mob depends on several other
qualities, since the column Mob has a great number of non-
empty cells. We may be tempted to sum the number of
non-empty cells to set a measure of influence or depen-
dency, but this would not capture interrelationships. We
propose to calculate a relationship strength value s for each
cell of the matrix based on the combination of basic
strength values associated with the basic relationship types
(R, 1, -) and their modalities (a, m, d). For example, a
cell with aR would have a strength value greater than a cell
with mR, since ‘‘always required’’ is stronger than ‘‘may be
required in some cases’’. A cell with just R (i.e. not
knowing if always required or not) would be between those
two cases (aR and mR). In this approach, the sum of the
cell values s in a row/column would be a better measure of
influence/dependence. Furthermore, we could sort the
quality requirements according to their strengths. In fact,
any formula that permits a relative numerical comparison
between quality requirements is valid for the above-men-
tioned calculations, provided that it could incorporate the
essence of the interrelationships. In this paper, we propose
the following empirical expression (based on the experi-
ence of the present authors as software engineers) to cal-
culate the relationship strength value s, in which we give
more importance to the R relationship and consider the
negative effect of the hinder relationship as a subtractive
term:
s ¼ f ðsRÞ þ f ðsþÞ � f ðs�Þ ð1Þ
where sR, s?, and s- are the basic strength values of the
relationships R, ?, and -, respectively. In this model, we
consider sR, s?, s- [ [0, 1.0], and sR,[ s?, sR[ s- to
increase the importance of the relationship R. We proposed
the modality function f, which intensifies or reduces the
basic strength values (i.e. sR, s?, s-) according to the
modality symbol (i.e. m, d, a, or none) present in the
relationship, as follows:
f ðxÞ ¼0:5x if m; d
ffiffiffi
xp
if a
x otherwise
8
<
:
ð2Þ
Considering that in Eq. 2, x [ [0, 1.0] (i.e. x can be sR,
s? or s-), we used the square root to increase the impor-
tance of the a modality (‘‘always required’’) over other
possibilities. As a hypothetical example, we may instanti-
ate this model using the arbitrary basic values sR = 0.6,
s? = 0.2, and s- = 0.2. If a cell (i,j) contains ‘‘mR’’ (i.e.
i may be required by j), the relationship strength value
s will be 0.3 (this is the case of this is the case of ‘‘Mob
mR LSR’’ represented in Fig. 4). As another example, if a
cell (i,j) contains ‘‘aR’’ (i.e. i is always required by j), the
relationship strength value s will be 0.775 (e.g. ‘‘UHP
aR LSR’’ represented in Fig. 4).
154 Requirements Eng (2017) 22:137–165
123
In pervasive mobile games, the cells in the dependence
matrix have only one relationship (such as aR, mR, and
m-). However, on a general case a single dependence
matrix cell (i,j) might contain multiple relationships, such
as ‘‘aR, -’’. (i.e. i is always required by j, but i also
hinders j). As a final remark, we do not recommend using
cells with multiple ? or - (e.g. ?? or ???) because this
type of difference is difficult to be noticed at a higher level
of abstraction. We think this would be more appropriate
when the designer is using specific techniques to opera-
tionalize specific qualities. The reader should refer to [77]
for a discussion of such issues.
The sum of the relationship strengths of a row represents
the degree of influence of the corresponding quality
requirement. The sum of the strengths of a column repre-
sents the degree of dependence of a particular quality
requirement. The matrix is a primary source for several
other analysis and views. The degrees of influence/depen-
dence can be sorted in a descending/ascending order and
help us to identify the most important or critical require-
ments. These identifications provide some rationale for the
software development decisions. This is something similar
to the ‘‘claim softgoals’’ proposed by Chung and co-authors
[78], although here we are at a much higher abstraction
level.
3.4.2 Dependence matrix using pervasive mobile game
examples
The dependence matrix can generate dependence diagrams,
where the more influential qualities (based on their degrees
of influence) are drawn at the higher levels of the graph and
those qualities with the highest degrees of influence are
marked by shaded rectangles. We use the methodology of
Sect. 3.4.1 to analyse the dependencies among the quality
requirements for pervasive mobile games.
For example, we are able to find the most influential
qualities and most dependent qualities by applying Eq. 1
with custom values for sR, s?, and s- to calculate degrees
of influences and degrees of dependence. For example, we
can arbitrarily use sR = 0.6, s? = 0.2, and s- = 0.2 to
boost the R relationship, because we consider this rela-
tionship stronger than the other basic relationships. Fig-
ure 5 illustrates a partial example of influence and
dependence order (the first three elements only) calculated
by applying Eq. 1 with the above-mentioned values for sR,
Fig. 4 The dependence matrix for pervasive mobile game qualities
Requirements Eng (2017) 22:137–165 155
123
s?, and s-. Figure 5 reveals that among second-level
NFRs, UHP is the most influential quality and LSR is the
most dependent one. This analysis reveals the importance
of UHP, Con, and GCA as second-level NFRs. These
calculations can be carried out by hand or with the help of a
computer program (which is currently under development
by the present authors). For example, we can calculate the
degree of influence of UHP by looking up the row that
contains UHP (in Fig. 4) and applying Eq. 1 in each shaded
cell in that row. Using the above-mentioned values for sR,
s? and s-, this calculation yields: degree of influence =
(0.7746 ? 0.3 ? 0.7446 ? 0.7446 - 0.1 ? 0.7746 ?
0.7746 ? 0.7746) = 4.8476. Similarly, we can calculate
the degree of dependence of LSR by looking up the column
that contains LSR and applying Eq. 1 in each shaded cell in
that column. This calculation yields: degree of depen-
dence = 0.3 ? 0.2 ? 0.3 ? 0.7446 ? 0.2 ? 0.3 ? 0.3 =
2.3746.
Using the dependence matrix in Fig. 4 as reference, we
can draw dependence diagrams. For example, Fig. 6
illustrates a dependence diagram considering only aR re-
lationships (‘‘always required’’) and second-level NFRs.
We chose the UHP quality as the root of the example
diagram because UHP is the most influential quality among
second-level NFRs. To create this diagram, we firstly look
up the row that contains UHP in Fig. 4. Next, we browse
the UHP row searching for aR relationships, which reveals
these qualities: LSR, GOT, GCA, GA, Per, and Con. For
each of these qualities, we repeat this process until there
are no more aR relationships to consider.
As another possible view, we can partition the interde-
pendence matrix into submatrices by rearranging rows and
columns in such a way that regions with great incidences of
strong relationships are grouped. In this later case, the
rearranged matrix may reveal more independent sets of
quality requirements. Also this reordered matrix may
reveal good approaches to the design of the system archi-
tecture. The dependence matrix is a type of Design
Structure Matrix—DSM [79, 80] and, consequently, enjoys
the benefits, algorithms, and tools related to DSMs.
Our research on the dependence matrix is not com-
pleted, but we are confident enough to claim that it can be a
promising technique for non-functional requirements elic-
itation in several other domains. As an example of how we
defined the matrix of Fig. 4, the reader can check the LSR
quality in some games as follows:
1. Mob may be required by LSR for operationalization
(Mob mR LSR): In Feeding Yoshi [61], the players
need to feed a special character by searching for food
in ‘‘plantations’’. The plantations correspond to secure
Wi-Fi access points that are spread on the physical
environment, which requires players to wander around
to find these ‘‘food sources’’;
2. GOT helps LSR (GOT ? LSR): The Gigaputt game
[58] redefines physical places as mini-golf courses. To
help in improving LSR, in these game players manip-
ulate the mobile devices as if they were golf clubs,
using a tangible object metaphor.
3. UHP is always required by LSR for operationalization
(UHP aR LSR): The game needs to use UHP to handle
uncertainty issues related to location data—the game
obtains these data through location systems such as
GPS;
4. INP helps LSR (INP ? LSR): Our prototype Pervasive
Word Search (Sect. 4) uses Bluetooth devices of non-
players to create special game zones that have specific
purposes in the game;
5. Per and Con may be required by LSR for operational-
ization (Per mR LSR, Con mR LSR): this is the case
when the game needs to associate some specific data to
some physical place. For example, our prototype
Location-based quiz [81, appx. C] defines some
questions that the players need to answer when they
arrive at a physical room. Each room has a different set
of questions.
As far as the conditional modality, d, is concerned, we
can produce the following example:
• GP hinders Con if the intensity of GP is above some
threshold (GP d2 Con): If the pacing of activities in the
game is high, connectivity may not work well. For
example, this is an issue that Bell and co-authors [61]
reported for Feeding Yoshi. In that game, when players
were moving fast (on cars, trains), they had trouble
Fig. 5 Partial example of influence/dependence order
Fig. 6 Dependence diagram (aR only) created by visually inspecting
the dependence matrix in Fig. 4, starting with the UHP quality
156 Requirements Eng (2017) 22:137–165
123
connecting to access points. This caused players on cars
to slow down to be able to catch the virtual content.
4 Case example
This case example demonstrates how the pervasiveness
quality applies to a single-player pervasive mobile game
developed by the authors, Pervasive Word Search. In this
example, we outline which checklist questions are relevant
to the non-functional requirements under discussion. This
section presents the most important non-functional
requirements used to develop this pervasive mobile game.
We had these goals when developing this game (1) a game
that can be played in current smartphones; (2) a context-
aware game that retrieves as much content from the envi-
ronment as possible using current smartphones; (3) a game
that requires no special set-up (regarding infrastructure).
The objective of Pervasive Word Search is to find the
letters from a word that the game draws. The player must
explore the environment surrounding him to find the letters.
While exploring the physical world, the player may interact
with some game zones—the dark, open, and wireless
zones. The ‘‘dark zone’’ is a place with ‘‘low’’ ambient
light. An ‘‘open zone’’ corresponds to an outdoor area. A
wireless zone corresponds to a physical area with a certain
number of Wi-Fi access points and Bluetooth devices.
Interacting with these zones is an important part in this
game.
The player is able to get letters by capturing colours
with the device camera and interacting with the wireless
and dark zones. For example, 1 illustrates a player in
Pervasive Word Search focusing the device camera on a
red flower to capture the letters ‘‘r’’, ‘‘e’’, and ‘‘d’’—thus
getting the letters ‘‘r’’, ‘‘e’’, and ‘‘d’’, and eliminating ‘‘r’’,
‘‘e’’, and ‘‘d’’ from the target word. The game identifies a
finite set of colours: red, yellow, orange, green, purple,
blue, pink, black, white, and grey. The player goes to
wireless zones to get letters that do not exist in the basic
colour names (like ‘‘f’’, considering names in English). The
player has a finite time period to find all letters. Interacting
with the game zones changes the game clock behaviour.
For example, if a wireless zone has lots of devices, the
game clock runs slower. If the player is inside an open
zone, the game clock runs faster. For more specific details
on this game, the reader should refer to [81].
4.1 Analysing pervasiveness
In Pervasive Word Search, the first-level non-functional
requirements that most contribute to pervasiveness are
Spatiality, Resilience, and Context-awareness. This section
uses the checklists (Sect. 3.3) to evaluate the pervasive
game qualities (and thus, pervasiveness) in Pervasive Word
Search. In other words, we describe how the game applies
these requirements and how the other requirements do not
apply to the game. In this section, we refer to checklist
questions by their acronyms.
4.1.1 Spatiality
The game implements the Spatiality quality through Local
space redefinition and Mobility. The game uses Local
space redefinition to define three game zones: the open
zone, the wireless zone, and the dark zone. In this regard,
players are able to perceive physical places as ‘‘game
zones’’ (LSR1, LSR2, LSR3) when they go to physical
places that match the requirements of a specific game zone.
The game does not use live non-player characters (LSR5),
and the game activities do not affect the physical world
(LSR6).
The game applies Mobility, as it requires players to
wander around the physical world to find sources for letters
(Mob4). The game area is unconstrained. This game does
not use networking, so Mob1 and Mob2 do not apply. This
game does not depend on specific locations to be played
(e.g. a specific room, city square, or building), so Mob3
does not apply. The ‘‘game zones’’ are not actual physical
places—they are virtual elements applied to a physical
environment, being created dynamically through context
information. In other words, these ‘‘game zones’’ may
disappear or move over time depending on the current
context. Pervasive Word Search does not use location
information, so Mob5 does not apply.
4.1.2 Resilience
The game implements Resilience by defining Uncertainty
handling policies and adjusting the Game pacing to be
compatible with these policies. The relevant questions for
Uncertainty handling policy are UHP1 and UHP2. The
game applies the general strategy hide to handle several
uncertainties. For example:
• The colours are represented with a limited set of
options. Similar ‘‘colours’’ are grouped with the same
name (e.g. all variations of ‘‘red’’ are considered as
‘‘red’’);
• Bluetooth and Wi-Fi queries are slow operations. It is
also possible that some queries miss some devices or
return false positives. Thus, the game does not require
real-time response to use these data. The interaction
with those sensors is indirect—the game queries the
sensors in the background and announces the results
when they are ready. In other words, the game does not
Requirements Eng (2017) 22:137–165 157
123
use Bluetooth and Wi-Fi information in real-time
fashion.
• Game zone representation is ambiguous—the game
does not display zone maps nor tracks physical world
limits (UHP3). Instead, the game just informs the
player if he/she has entered or left a game zone (e.g. the
wireless, dark, and open zones).
Regarding the Game pacing, the activities were
designed to not require fast responses from players. Also,
the players are not required to move fast while playing
(GP1). Additionally, the game pacing was adjusted to be
compatible with the nature of some operations that the
game performs (GP2). For example, some activities rely on
Bluetooth and Wi-Fi network queries, which are very slow
operations. The pace of game activities enables players to
focus the physical environment most of the time (GP3).
4.1.3 Context-awareness
Pervasive Word Search explores Context-awareness
mainly through Game content adaptability. The game
generates all content through sensing the environment,
except for the target word that comes from an internal
database (GCA4). In this regard, players may have trouble
playing in areas with no Wi-Fi access points or Bluetooth
devices around (GCA1, GCA3). In this case, players would
have to search for places where these devices could be
found. We do not regard this issue as a problem—instead,
it is part of the requirement to have players explore the
physical world (GCA2).
Pervasive Word Search does not use tangible object
metaphors, so Game object tangibility does not apply to
this game. In other words, the mobile phone works as mere
terminal to play the game, with no specific role in the game
design.
4.1.4 Permanence
The Permanence quality in Pervasive Word Search is
limited. For example, the game barely supports Persis-
tency. Game sessions do not depend on previous session
data. However, the game draws the target word from an
internal database to start a new game session (Per1). The
other persistency questions (Per2 and Per3) do not apply to
the game.
Considering Daily life interleaving, the game does not
require long-time commitment from players (DLI3), as
Pervasive Word Search implements the concept of ‘‘short
play sessions’’ [48] that do not depend on previous ses-
sions. In this regard, the game can be easily interrupted
without further loss to the players (DLI2) and can be
restarted anytime (DLI5). Pervasive Word Search does not
provide a persistent world, so DLI1, DLI4, DLI6, and DLI7
do not apply.
4.1.5 Sociality
Considering Social communication and Pervasive Word
Search as being a single-player game, questions SC1, SC2,
and SC4 do not apply. The game activities do not require
players to approach strangers (SC3). However, the game
encourages players to go to areas with lots of people, as the
game could become easier if there are many Bluetooth
devices in these places (the game clock runs slower in this
situation). Finally, the game is not designed to foster
emergent gameplay (SC5).
Regarding Conformance to physical and social settings,
relevant questions are CPS2, CPS3, CPS4, and CPS5. As
the game requires players to point the device camera to
capture colours, this behaviour might create issues related
to these questions. For example, a player might get repri-
manded while trying to capture a colour while playing in
some places. Also, a player might embarrass people while
trying to capture a colour. In our preliminary tests with this
game, we had no issues regarding these questions, although
evaluating this quality thoroughly would require a more
specific study. Pervasive Word Search does not require
user data (CPS1).
Considering Involving non-players, Pervasive Word
Search generates content using Bluetooth devices (i.e. their
name identifiers) that other people carry (INP3). We can
say that the game involves these non-players indirectly
(INP1). However, the game does not contact these devices
nor the player is required to approach these people (INP1,
CPS2, SC3). Non-players are unable to take part in the
game directly (INP1). The INP2 question does not apply to
this game as it is a single-player game, and neither does
Pervasive Word Search use human actors as non-player
characters (INP4).
4.1.6 Communicability
Pervasive Word Search does not support Communicability,
as it is a single-player game that does not use Connectivity.
Although the game queries Bluetooth devices and Wi-Fi
networks, it does not establish remote connections—the
game uses these devices only as content sources.
4.1.7 Accessibility
Pervasive Word Search is not a transmedia game (TM5)—
the game has only one game interface (TM1). In this
regard, questions TM2, TM3, and TM4 do not apply. This
game was developed originally for Symbian/Qt devices.
158 Requirements Eng (2017) 22:137–165
123
This means that Device independence is limited to Nokia
smartphones (DI1).
Considering Access usability, the game does not require
players to focus on the device screen most of the time
(Usa1). Instead, the game encourages players to focus on
the physical world, searching for colours. Concerning
Usa2, the game provides two modes of interaction: touch
input (to capture colours) and context input (information
coming from ‘‘the game zones’’).
Considering Game autonomy, Pervasive Word Search
was designed to generate all game content (excluding the
target word) from any physical environment that has
coloured objects, Wi-Fi network access points, and Blue-
tooth devices (GA2). Also, the game may generate extra
(redundant) content when the player enters a ‘‘dark zone’’
(a place with ‘‘low’’ ambient light) or an ‘‘open zone’’ (an
outdoor area). In other words, the game is not bound to a
specific physical place (GA1). The game requires no
orchestration (GA3) and no custom hardware (GA4).
5 Related work
Considering quality attributes before defining the software
architecture has been the main reason for the literature on
non-functional requirements. In this regard, the central
issue is to define in advance (along with the stakeholders)
the desired quality attributes that a given software product
should address. Freeman [21] defined a set of basic quali-
ties (functionality, reliability, ease of use, economy, and
safety) and extra qualities (flexibility, reparability, adapt-
ability, understandability, documentation, and enhance-
ability) that apply to general software systems. However, a
software product may require specific qualities that are not
included in these previous sets of qualities. Also, not all
stakeholders involved in the software definition may be
aware of which qualities to consider. In such cases, the
literature has been producing lists or taxonomies of quali-
ties attributes for software. Some examples are [78, 82, 83].
In this context, the NFR (non-functional requirements)
framework [78] has been a central reference in require-
ments engineering research. This framework has been
proposed by the Toronto group [78]. We present below an
overview of this framework because it is essential to
understand the present paper.
The overall idea in the NFR framework is to treat NFRs
as first class citizens. The framework proposes a language
designed to represent NFRs and their relationships with
other NFRs and with possible functions that will carry on
the NFR implementation. In that perspective, the main task
of requirements engineers is to elicit which NFRs will be
considered, their relationships and their operationalizations
(i.e. how they will be handled in terms of software
functions). A key point in the NFR framework is treating
NFRs as softgoals. The term ‘‘softgoal’’ is in opposition to
‘‘hard goal’’. Softgoals relate to the idea that the notion of
‘‘success’’ may vary among stakeholders, as they may
perceive and evaluate qualities in different ways due to
diverse world-views and backgrounds. This reasoning is
similar to Simon’s idea of ‘‘bounded rationality’’ [84],
which says that rationality is limited by three factors: the
information available at the time of a decision, the cogni-
tive limitations of the person, and the time available to
make a decision. This idea implies that objective, clear-cut
reasoning might be hard to achieve in practice due to the
three factors mentioned previously. In this context, the
NFR framework uses the word ‘‘satisfice’’ when relating to
softgoals, which means that a softgoal is not satisfied, but
‘‘satisficed’’ (i.e. partially satisfied based on qualitative
judgements such as ‘‘is it good enough?’’). The term
‘‘satisfice’’ denotes that the conditions in which a decision
will be made depend on the current context and involved
viewpoints (without precise determination or clear-cut
criteria), which leads to assessing qualities using qualita-
tive judgements. In this sense, the NFR framework deals
with qualities using a qualitative approach, not relying on
hard metrics to evaluate these qualities.
The NFR framework defines a graphical language that
supports modelling and verification. The model produced is
called SIG (Softgoal Interdependence Graph), and the
verification is conducted by qualitative reasoning based on
label propagation [85]. The NFR framework language
defines four major symbols: one to represent a softgoal (a
cloud), one to represent the softgoal operationalization (a
bold cloud), one to represent contribution, and one to
represent correlation. A softgoal has to be named by two
descriptors: a type descriptor and a topic descriptor. The
relationships (lines) are labelled according to label cata-
logues, together with rules used for verification (propaga-
tion). Contributions are relationships on a branch of the
graph, whereas correlations are relations from different
branches. An important side effect of the NFR framework
is that the rationale of decisions is explicit.
Below we provide a simple example of how the NFR
framework deals with quality concerns, using the Pervasive
Word Search game (Sect. 4) and some pervasive qualities
presented in Sect. 3.3. This example refers to an important
feature in Pervasive Word Search—getting letters from the
physical environment. Figure 7 illustrates a possible (par-
tial) quality model concerning the ‘‘get letters’’ feature as a
SIG. The SIG in Fig. 7 defines the overall softgoal Quality
(type) for the PervasiveWordSearch (topic), which is
related to Mobility, Conformance to physical and social
settings, and Game content adaptability using the AND
label. The SIG also relates Mobility to both Local space
redefinition and Game pacing using HELP labels. The SIG
Requirements Eng (2017) 22:137–165 159
123
represents possible implementations of ‘‘get letters’’ as two
operationalizing softgoals: ‘‘GetLetters by Wi-Fi’’ (getting
letters using Wi-Fi access point names as the source of
information) and ‘‘GetLetters by camera’’ (getting letters
using the smartphone camera to capture coloured objects,
using the captured colour name as the source of
information).
Operationalization through Wi-Fi access point names
contributes in a strong manner to Local space redefinition
(represented with a HELP label), as this operationaliza-
tion helps to create the ‘‘wireless zone’’ in this game, but
it contributes negatively to Game pacing because pacing
needs to be low as Wi-Fi access point scanning is a slow
operation. We may notice that the operationalization
through Wi-Fi access point names correlates to Confor-
mance to physical and social settings with a HELP label
because this operationalization helps to avoid issues
related to checklist questions CPS2, CPS3, CPS4, and
CPS5 (Sect. 3.3.5). This operationalization may con-
tribute negatively to Game content adaptability (repre-
sented with a HURT label) as it will not work in places
that have no Wi-Fi access points around. In these situa-
tions, the player will have to move to another place to find
Wi-Fi access points.
Operationalization though the camera may hinder Game
pacing (a relation represented with a HURT label) because
pacing needs to be low so that the player is able to properly
focus an object to get its colour. This operationalization
may hinder Conformance to physical and social settings
(represented with a HURT label) as the player might get
into situations related to issues mentioned in checklist
questions CPS2, CPS3, CPS4, and CPS5 (Sects. 3.3.5 and
4.1.5). Finally, this operationalization contributes posi-
tively to Game content adaptability (represented with a
HELP label), because coloured objects are widespread and
ubiquitous—the player will have plenty of information
sources to draw letters from the environment.
Using Fig. 7 as a model, it is possible to apply rules for
verification. For instance, if ‘‘get letters’’ is implemented
through access point names, then label propagation will
register Local space redefinition as HELP (?), Game
pacing as HURT (-), Conformance to physical and social
Fig. 7 Softgoal Interdependence Graph example
160 Requirements Eng (2017) 22:137–165
123
settings (?) as HELP, and Game content adaptability as
HURT (-), resulting in neutral (0) for the overall quality.
On the other hand, if ‘‘get letters’’ is implemented through
the camera, Game pacing registers as HURT (-), Con-
formance to physical and social settings (-) as HURT, and
Game content adaptability as HELP (-), resulting in an
overall negative (-) quality. When verifying the model
using a reasoning (propagation) strategy, the consequences
of a given design choice become more clear. The depen-
dence matrix in Sect. 3.4 is a similar way of dealing with
the interaction of qualities.
In this paper, Fig. 3 illustrates (on the first level) seven
non-functional requirements (qualities) that contribute to
pervasiveness. On the second level, there are sixteen non-
functional requirements that contribute to the first-level
qualities. Using the language of the NFR framework [78],
the first-level non-functional requirements and the perva-
sive game qualities can be considered as NFR softgoals.
Figure 3 corresponds conceptually to a NFR SIG (softgoal
interdependency graph), and all relationships in Fig. 3 are
equivalent to ‘‘HELP’’ contributions in the NFR
framework.
The relationships in Fig. 3 mean that the non-functional
requirement contributes positively to the parent quality. For
example, Uncertainty handling policy and Game pacing
contribute positively to the Resilience quality, which in
turn contributes positively to Pervasiveness, along with
Spatiality, Sociality, Permanence, Accessibility, Context-
awareness, and Communicability. However, not all quali-
ties must be present in all levels to characterize perva-
siveness. We understand that pervasive mobile games may
have different levels of pervasiveness depending on how
these non-functional requirements (qualities) are applied.
For each pervasive game quality in Fig. 3, we have
created a checklist. The checklists are a set of questions
that designers can use to evaluate pervasive game qualities
(and thus, pervasiveness) in existing projects. Another
application of these checklists is to help designers in
introducing pervasive game qualities in new projects, by
answering the questions. This idea of having questions to
help in verifying and operationalizing requirements is
similar to the idea that the GQO (goal-question-opera-
tionalization) [86] proposes.
The dependence matrix in Fig. 4a presents the possible
impact of a quality onto other ones. We defined some
symbols (such as aR, mR, ?) to represent the relationships
(and impacts) that we had identified so far. This idea shares
similarities with the ‘‘label’’ concept that the NFR frame-
work defines for its ‘‘evaluation procedure’’ [78]. We can
also find other similarities with the work by Chung and co-
authors [78], but in our case we are dealing with NFR at a
higher level of abstraction (with no reference to specific
methods for operationalization). The dependence matrix
can be compared to the Design Structure Matrices (DSMs)
[79, 80], which are used to help in analysing dependencies
between design elements in products. More examples of
works related to DSMs are [87–90].
The qualities that we discovered for pervasiveness can
be used to design pervasive mobile games considering non-
functional requirements from the very beginning of the
project development. As Chung and Leite [91] note, non-
functional requirements place restrictions on how the
desired functionality can be implemented, not being
something that could be considered at a later stage of
project development. Hence, integrating non-functional
requirements and the desired functionality is crucial. For
example, Uncertainty handling policy (Sect. 3.3.7) deter-
mines policies to handle technology uncertainties (espe-
cially related to sensors and networking), which might
involve requirements for accuracy, response time, and
other quality factors.
This domain knowledge (the non-functional require-
ments related to pervasiveness and checklists) might be
used to inspire game ideas that may lead to functional
requirements. For example, a game might apply Local
space redefinition (Sect. 3.3.1) to create special ‘‘game
zones’’ in the physical space. For example, our prototype
Pervasive Word Search creates three types of game zones
(wireless, open, dark) in the physical environment. In
Sect. 4, we use Pervasive Word Search to describe how the
non-functional requirements related to pervasiveness are
applied in this game.
The process that we conducted to build knowledge about
pervasive mobile games shares similarities with the process
that Leite and Cappelli [92] conducted to build their
Transparency SIG. In their process, Leite and Cappelli [92]
had the assistance of domain experts to gather knowledge
about the domain and to validate the requirements. The
process that we conducted also had input from domain
experts, and we evaluated the requirements that we dis-
covered using game prototypes that we developed. For
example, Sect. 4 has described an evaluation we had con-
ducted with one of our prototypes.
6 Conclusions
Digital games are creative projects that require the input of
professionals with very diverse backgrounds, including
game designers, artists, and software developers. The game
development process has a preproduction phase where
game designers work on the central vision of the game. In
the preproduction phase, the game designers and artists
create game elements, such as the storyline, game design,
and art assets. Later, the results of this phase are carried on
to the production phase where software developers
Requirements Eng (2017) 22:137–165 161
123
translate the game vision and integrate multimedia assets
into the software to realize the digital game.
Managing the expectations of these different stake-
holders is a difficult task in game development. Many
projects fail due to problems related to the transition
between the preproduction and production phases.
Improving game development processes requires improv-
ing existing requirements engineering processes to consider
the particularities of game projects. Some of these partic-
ularities follows:
• As game projects are a creative endeavour, the process
must take the preproduction phase into consideration.
Non-game projects do not have a preproduction phase;
• A software requirement may originate in the prepro-
duction phase. For example, a game designer may
propose a functional requirement that turns out to be a
non-functional requirement in software implementa-
tion. In this regard, there are other stakeholders (other
than software developers) that may propose require-
ments; and
• Games are projects that focus on entertainment rather
than productivity. As a result, games have requirements
that are not present in non-game applications, such as
‘‘fun’’, ‘‘gameplay’’, ‘‘enjoyment’’, and ‘‘flow’’.
The advent of pervasive games complicates this scenario
by presenting various issues:
• Research on pervasive games has attracted people with
different backgrounds, such as those that have experi-
ence in computer science, game studies, design, and
theatre, which produces a rich discussion about this
topic but also creates a confusing situation. The end
result is that the discussion on pervasive games is too
broad, especially regarding support for pervasive game
development.
• The literature on pervasive games is confusing and often
mixes up preproduction issues and production ones.
Some of the literature (e.g. game studies and game
design) discusses pervasive games using only prepro-
duction concepts, while other parts of the literature
emphasize technological aspects (e.g. pervasive com-
puting technology). Table 1 summarizes these views.
• A pervasive game integrates the virtual and physical
worlds into a mixed-reality. As a consequence, a game
designer may propose a requirement that can be
operationalized outside of the software realm. For
example, the pervasive game Uncle Roy all around
you [49] uses human actors to play non-player
characters that interact with the player. This is an
example of operationalizing Involving non-players
(Sect. 3.3.5) through means other than software,
which reflects the interdisciplinary nature of pervasive
mobile games.
Considering this confusing scenario, in this paper we
described a process for building knowledge on pervasive
mobile games—a subset of pervasive games that uses con-
text-aware mobile devices to enable the interaction between
the environment (physical world) and the virtual world.
The mapping process that we conducted revolves around
trying to understand what makes a game pervasive—a
question that relates to ‘‘pervasiveness’’, which is a quality
that makes those games unique and it is very hard to
objectively define. Our map regarding pervasiveness in
pervasive mobile games is composed of a conceptual map,
a set of checklists, and a dependence matrix.
The conceptual map describes pervasiveness according
to a set of first-level requirements that we have identified:
spatiality, permanence, accessibility, sociality, context-
awareness, resilience, and communicability. Each first-
level requirement has associated second-level requirements
that contribute to it. For each second-level requirement, we
defined checklists to help to verify and operationalizing
them. Finally, we present the first version of a dependence
matrix that records various relationships among these
requirements and that helps with design decisions.
We see these results, in their current form, as a prag-
matic roadmap to navigate the sea of pervasive mobile
games. These results can help game developers by pro-
viding: (1) a resource to spark novel game ideas; (2) a
language to discuss pervasive mobile games and their
characteristics; (3) a starting point to discover functional
and non-functional requirements for pervasive mobile
games; and (4) a starting point to find other relevant fea-
tures in pervasive mobile games.
We believe that there is a rich research agenda ahead for
pervasive mobile games. In future works, we aim to
explore the following topics:
• Refining and validating the qualities that we identified
using more games and designers;
• Refining the dependence relationships among the
identified qualities;
• Identifying more qualities related to pervasiveness;
• Exploring alternatives to operationalize the qualities
that we reported in this paper. We can begin this
exploration by answering the ‘‘how’’ questions in the
checklists.
Acknowledgments The authors thank CNPq, FAPERJ, FINEP, and
CAPES for the financial support to this paper.
References
1. Gartner (2013) Gartner says worldwide video game market to
total $93 Billion in 2013. http://www.gartner.com/newsroom/id/
2614915
162 Requirements Eng (2017) 22:137–165
123
2. Callele D, Neufeld E, Schneider K (2011) A report on select
research opportunities in requirements engineering for videogame
development. In: 2011 Fourth international workshop on multi-
media and enjoyable requirements engineering—beyond mere
descriptions and with more fun and games (MERE), pp 26–33
3. Callele D, Neufeld E, Schneider K (2005) Requirements engi-
neering and the creative process in the video game industry. In:
Proceedings on 13th IEEE international conference on require-
ments engineering, 2005, pp 240–250
4. Alves C, Ramalho G, Damasceno A (2007) Challenges in
requirements engineering for mobile games development: the
meantime case study. In: 15th IEEE international requirements
engineering conference, 2007. RE’07. pp 275–280
5. Furtado AWB, Santos ALM, Ramalho GL (2010) Streamlining
domain analysis for digital games product lines. In: Bosch J, Lee
J (eds) Software product lines: going beyond. Springer, Berlin,
pp 316–330
6. Norneby J, Olsson T A new attitude to game engineering:
embrace change, re-use, fun. http://www.gamasutra.com/view/
feature/132491/a_new_attitude_to_game_.php
7. Bjork S, Falk J, Hansson R, Ljungstrand P (2001) Pirates! Using
the physical world as a game board. In: Proceedings of interact
2001, IOS Press, pp 9–13
8. Schneider J, Kortuem G (2001) How to host a pervasive game—
supporting face-to-face interactions in live-action roleplaying. In:
Interactions in live-action roleplaying. UbiComp workshop on
Designing Ubiquitous Computing Games
9. Sotamaa O (2002) All the world’s a Botfighter stage: notes on
location-based multi-user gaming. In: Frans M (ed) Computer
games and digital cultures conference proceedings, Tampere
University Press, Tampere, p 10
10. Gbanga: Gbanga Famiglia, http://gbanga.com/portfolio/gbanga-
famiglia/
11. YD Online: GEO Hunters for iPhone, iPod touch, and iPad on the
iTunes App Store. http://itunes.apple.com/us/app/geo-hunters/
id417926906?mt=8&ign-mpt=uo%3D6
12. Petrillo F, Pimenta M, Trindade F, Dietrich C (2009) What went
wrong? A survey of problems in game development. ACM
Comput Entertain 7:13:1–13:22
13. Kanode CM, Haddad HM (2009) Software engineering chal-
lenges in game development. In: Sixth international conference
on information technology: new generations, 2009. ITNG’09,
pp 260–265
14. Bentley T, Johnston L, von Baggo K (2002) Putting some emo-
tion into requirements engineering. In: Proceedings of the 7th
Australian workshop on requirements engineering
15. Callele D, Neufeld E, Schneider K (2010) An introduction to
experience requirements. In: 18th IEEE international require-
ments engineering conference (RE), 2010, pp 395–396
16. Potts C (1995) Invented requirements and imagined customers:
requirements engineering for off-the-shelf software. In: Pro-
ceedings of the second IEEE international symposium on
requirements engineering, pp 128–130
17. Poels K, de Kort Y, Ijsselsteijn W (2007) ‘‘It is always a lot of
fun!’’: exploring dimensions of digital game experience using
focus group methodology. In: Proceedings of the 2007 conference
on future play. ACM, New York, NY, pp 83–89
18. Arango GF (1988) Domain engineering for software reuse. PhD
Thesis, University of California, Irvine
19. Glinz M (2007) On non-functional requirements. In: 15th IEEE
international requirements engineering conference, 2007. RE’07,
pp 21–26
20. Kazman R, Abowd G, Bass L, Clements P (1996) Scenario-based
analysis of software architecture. IEEE Softw 13:47–55
21. Freeman P (1987) Software perspectives: the system is the
message. Addison-Wesley Longman Publishing Co. Inc, Boston
22. Oxford English Dictionary (2015) Pervasive, adj. http://www.
oxforddictionaries.com/definition/english/pervasive?q=pervasive
23. Salen K, Zimmerman E (2004) Rules of play: game design fun-
damentals. MIT Press, Cambridge
24. Montola M, Stenros J, Wærn A (2009) Pervasive games: theory
and design. Morgan Kaufmann, Burlington
25. Nieuwdorp E (2007) The pervasive discourse: an analysis.
Comput Entertain CIE 5(2):13. doi:10.1145/1279540.1279553
26. Montola M, Waern A, Nieuwdorp E (2006) Domain of pervasive
gaming. IPerG
27. Davies H (2007) Place as media in pervasive games. In:
Porceedings of fourth Australasian conference on interactive
entertainment, pp 7:1–7:4
28. Magerkurth C, Cheok AD, Mandryk RL, Nilsen T (2005) Per-
vasive games: bringing computer entertainment back to the real
world. Comput Entertain CIE 3:4
29. Saarenpaa H, Korhonen H, Paavilainen J (2009) Asynchronous
gameplay in pervasive multiplayer mobile games. In: Proceed-
ings of 27th international conference on extended abstracts on
human factors in computing systems, pp 4213–4218
30. McGonigal J (2006) This might be a game: ubiquitous play and
performance at the turn of the twenty-first century. PhD Thesis,
University of California, Berkeley
31. Linner D, Kirsch F, Radusch I, Steglich S (2005) Context-aware
multimedia provisioning for pervasive games. In: International
symposium on multimedia, IEEE computer society, Los Alami-
tos, CA, pp 60–68
32. Capra M, Radenkovic M, Benford S, Oppermann L, Drozd A,
Flintham M (2005) The multimedia challenges raised by perva-
sive games. In: Proceedings of 13th annual ACM international
conference on multimedia, pp 89–95
33. Walther B (2005) Notes on the methodology of pervasive gam-
ing. In: Kishino F, Kitamura Y, Kato H, Nagata N (eds) Enter-
tainment computing—ICEC 2005. Springer, Berlin, pp 488–495
34. Hinske S, Lampe M, Magerkurth C, Rocker C (2007) Classifying
pervasive games: on pervasive computing and mixed reality.
Concepts Technol Pervasive Games Read Pervasive Gaming Res
1:11–38
35. Benford S, Magerkurth C, Ljungstrand P (2005) Bridging the
physical and digital in pervasive gaming. Communications of the
ACM, vol 48, pp 54–57
36. Milgram P, Takemura H, Utsumi A, Kishino F (1994) Aug-
mented reality: a class of displays on the reality-virtuality con-
tinuum. Syst Res 2351:282–292
37. McGonigal J (2003) A real little game: the pinocchio effect inpervasive play. In: Marinka C, Joost R (eds) Level up conference
proceedings: proceedings of the 2003 digital games research
association conference. CD ROM. University of Utrecht, Utrecht
38. Weiser M (1991) The computer for the 21st century. Sci Am
265:66–75
39. IBM Research: The convenience of small devices: how pervasive
computing will personalize e-business. https://web.archive.org/
web/20000420092553/http://www.research.ibm.com/resources/
magazine/1998/issue_3/bregman398.html
40. Dey AK (2001) Understanding and using context. Pers Ubiqui-
tous Comput 5:4–7
41. Leite JCSP, de Moraes EA, de Castro CEPS (2007) A strategy for
information source identification. In: Anais do WER07—work-
shop em Engenharia de Requisitos, Toronto, pp 25–34
42. IPerG: Integrated Project of Pervasive Games. http://www.perva
sive-gaming.org/index.php
43. Valente L, Feijo B (2013) A survey on pervasive mobile games.
Departamento de Informatica, PUC-Rio, Rio de Janeiro
44. Lindt I, Ohlenburg J, Pankoke-Babatz U, Ghellal S (2007) A
report on the crossmedia game epidemic menace. Comput
Entertain CIE 5(1):8. doi:10.1145/1236224.1236237
Requirements Eng (2017) 22:137–165 163
123
45. Lindt I, Ohlenburg J, Pankoke-Babatz U, Ghellal S, Oppermann
L, Adams M (2005) Designing cross media games. In: Proceed-
ings of 2nd international workshop on pervasive gaming appli-
cations, Munich, pp 8–13
46. Benford S, Crabtree A, Flintham M, Drozd A, Anastasi R, Paxton
M, Tandavanitj N, Adams M, Row-Farr J (2006) Can you see me
now? ACM Trans Comput Hum Interact TOCHI 13:100–133
47. Joffe B (2007) Mogi. In: Borries F, Walz SP, Bottger M (eds)
Space time play. Birkhauser Basel, pp 224–225
48. Peitz J, Saarenpaa H, Bjork S (2007) Insectopia: exploring per-
vasive games through technology already pervasively available.
In: Proceedings of the international conference on advances in
computer entertainment technology, ACM, New York, NY,
pp 107–114
49. Benford S, Flintham M, Drozd A, Anastasi R, Adams M, Row-
Farr J, Oldroyd A, Sutton J, Park A (2004) Uncle roy all around
you: implicating the city in a location-based performance. In:
ACE04. ACM Press
50. Rashid O, Bamford W, Coulton P, Edwards R, Scheible J (2006)
PAC-LAN: mixed-reality gaming with RFID-enabled mobile
phones. Comput Entertain CIE 4(4):4. doi:10.1145/1178418.
1178425
51. Bartneck C, Hu J, Salem B, Cristescu R, Rauterberg M (2008)
Applying virtual and augmented reality in cultural computing.
IJVR 7:11–18
52. Tuulos V, Scheible J, Nyholm H (2007) Combining web, mobile
phones and public displays in large-scale: manhattan story
mashup. In: LaMarca A, Langheinrich M, Truong K (eds) Per-
vasive Computing. Springer, Berlin, pp 37–54
53. Chatzigiannakis I, Mylonas G, Kokkinos P, Akribopoulos O,
Logaras M, Mavrommati I (2011) Implementing multiplayer
pervasive installations based on mobile sensing devices: field
experience and user evaluation from a public showcase. J Syst
Softw 84:1989–2004
54. FinN: Magnetize words. http://funinnumbers.eu/index.php?op
tion=com_content&view=article&id=15
55. Mottola L, Murphy AL, Picco GP (2006) Pervasive games in a
mote-enabled virtual world using tuple space middleware. In:
Proceedings of 5th ACM SIGCOMM workshop on network and
system support for games. ACM, New York, NY
56. Flintham M (2003) Uncle Roy all around you: mixing games and
theatre on the city streets. In: Copier M, Raessens J (eds) Pro-
ceedings level up the first international conference of the digital
games research association DIGRA, pp 168–177. University of
Utrecht and Digital Games Research Association
57. Lyytinen K, Yoo Y (2002) Issues and challenges in ubiquitous
computing. Commun ACM 45:62–65
58. Gigantic Mechanic: Gigaputt:: Make the world your golf course!
http://giganticmechanic.com/games_gigaputt.html
59. Ballagas R, Walz SP (2007) REXplorer: using player-centered
iterative design techniques for pervasive game development. In:
Magerkurth C, Roecker C (eds) Pervasive gaming applications—
a reader for pervasive gaming research, vol 2. Shaker, Aachen,
pp 255–284
60. Soute I, Markopoulos P, Magielse R (2009) Head up games:
combining the best of both worlds by merging traditional and
digital play. Pers Ubiquitous Comput 14:435–444
61. Bell M, Chalmers M, Barkhuus L, Hall M, Sherwood S, Tennent
P, Brown B, Rowland D, Benford S, Capra M, Hampshire A
(2006) Interweaving mobile games with everyday life. In: Pro-
ceedings of SIGCHI conference on human factors in computing
systems, pp 417–426
62. Ekman I, Ermi L, Lahti J, Nummela J, Lankoski P, Mayra F
(2005) Designing sound for a pervasive mobile game. In: Pro-
ceedings of the 2005 ACM SIGCHI international conference on
advances in computer entertainment technology, pp 110–116
63. Coenen T, Mostmans L, Naessens K (2013) MuseUs: case study
of a pervasive cultural heritage serious game. J Comput Cult
Herit 6:8:1–8:19
64. Dixon D, Kiani SL, Ikram A (2013) Experiences with AR plots:
design issues and recommendations for augmented reality based
mobile games. Commun Mob Comput 2:1–6
65. Valente L, de Souza CS, Feijo B (2008) An exploratory study on
non-visual mobile phone interfaces for games. In: Proceedings of
the VIII Brazilian symposium on human factors in computing
systems. Sociedade Brasileira de Computacao, Porto Alegre,
Brazil, pp 31–39
66. Valente L, Souza CSD, Feijo B (2009) Turn off the graphics:
designing non-visual interfaces for mobile phone games. J Braz
Comput Soc 15:45–58
67. Schmitz M, Moniri MM (2009) Burgomaster and Pedro—a per-
vasive multi-player game for rural tourism. In: Games and virtual
worlds for serious applications, 2009. Conference in VS-
GAMES’09, pp 205–208
68. Jegers K (2008) Investigating the applicability of usability and
playability heuristics for evaluation of pervasive games. In: Third
international conference on internet and web applications and
services, 2008. ICIW’08, pp 656–661
69. Koivisto E, Eladhari M (2006) User evaluation of a pervasive
mmorpg concept. In: DIME conference, Bangkok, Thailand
70. Martins T, Romao T, Sommerer C, Mignonneau L, Correia N
(2008) Towards an interface for untethered ubiquitous gaming.
In: Proceedings of the 2008 international conference on advances
in computer entertainment technology. ACM, New York, NY,
pp 26–33
71. Vogiazou Y, Reid J, Raijmakers B, Eisenstadt M (2006) A
research process for designing ubiquitous social experiences. In:
Proceedings of the 4th Nordic conference on human-computer
interaction: changing roles. ACM, New York, NY, pp 86–95
72. Reid J (2008) Design for coincidence: incorporating real world
artifacts in location based games. In: Proceedings of the 3rd
international conference on digital interactive media in enter-
tainment and arts, ACM, New York, NY, pp 18–25
73. Ishii H, Ullmer B (1997) Tangible bits: towards seamless inter-
faces between people, bits and atoms. In: Proceedings on SIGCHI
conference on human factors in computing systems, pp 234–241
74. Waern A, Montola M, Stenros J (2009) The three-sixty illusion:
designing for immersion in pervasive games. In: Proceedings
27th international conference on human factors in computing
systems, pp 1549–1558 (2009)
75. Ballagas R, Kratz SG, Borchers J, Yu E, Walz SP, Fuhr CO,
Hovestadt L, Tann M (2007) REXplorer: a mobile, pervasive
spell-casting game for tourists. CHI 07 Ext Abstr Hum Factors
Comput Syst pp 1929–1934
76. Drozd A, Benford S, Tandavanitj N, Wright M, Chamberlain A
(2006) Hitchers: designing for cellular positioning. In: Dourish P,
Friday A (eds) UbiComp 2006: ubiquitous computing. Springer,
Berlin, pp 279–296
77. Cunha H, Sampaio do Prado Leite JC, Duboc L, Werneck V
(2013) The challenges of representing transparency as patterns.
In: 2013 IEEE third international workshop on requirements
patterns (RePa), pp 25–30
78. Chung L, Nixon BA, Yu E, Mylopoulos J (1999) Non-functional
requirements in software engineering. Springer, Berlin
79. Baldwin CY, Clark KB (2000) Design rules vol 1: the power of
modularity. MIT Press, Cambridge
80. Lindemann U DSMweb.org: design structure matrix (DSM).
http://dsmweb.org/
81. Valente L (2011) A methodology for conceptual design of per-
vasive games. PhD Thesis, PUC-Rio
82. Boehm BW, Brown JR, Kaspar H (1978) Characteristics of
software quality. North-Holland, Amsterdam
164 Requirements Eng (2017) 22:137–165
123
83. Barbacci M, Klein MH, Longstaff TA, Weinstock CB (1995)
Quality attributes. Carnegie-Mellon University, Software Engi-
neering Institute, Pittsburgh
84. Simon HA (1996) The sciences of the artificial, 3rd edn. The MIT
Press, Cambridge
85. Giorgini P, Mylopoulos J, Nicchiarelli E, Sebastiani R (2002)
Reasoning with goal models. In: Spaccapietra S, March ST,
Kambayashi Y (eds) Conceptual modeling—ER 2002. Springer,
Berlin Heidelberg, pp 167–181
86. Serrano M, Leite JCSP (2011) Capturing transparency-related
requirements patterns through argumentation. In: 2011 first
international workshop on requirements patterns (RePa),
pp 32–41
87. Browning TR (2001) Applying the design structure matrix to
system decomposition and integration problems: a review and
new directions. IEEE Trans Eng Manag 48:292–306
88. MacCormack A, Rusnak J, Baldwin CY (2006) Exploring the
structure of complex software designs: an empirical study of open
source and proprietary code. Manag Sci 52:1015–1030
89. Lopes CV, Bajracharya SK (2006) Assessing Aspect Modular-
izations Using Design Structure Matrix and Net Option Value. In:
Rashid A, Aksit M (eds) Transactions on Aspect-Oriented Soft-
ware Development I. Springer, Berlin Heidelberg, pp 1–35
90. Matos PJ, Duarte R, Cardim I, Borba P (2007) Using design
structure matrices to assess modularity in aspect-oriented soft-
ware product lines. In: Proceedings of the first international
workshop on assessment of contemporary modularization tech-
niques. IEEE Computer Society, Washington, DC, USA, p 4
91. Chung L, Leite JCSDP (2009) On non-functional requirements in
software engineering. In: Borgida AT, Chaudhri VK, Giorgini P,
Yu ES (eds) Conceptual modeling: foundations and applications.
Springer, Berlin, pp 363–379
92. Leite JCSDP, Cappelli C (2010) Software transparency. Bus Inf
Syst Eng 2:127–139
Requirements Eng (2017) 22:137–165 165
123