Augotchi: The Augmented Pet

Exploring virtual companion design in Augmented Reality games

Master’s thesis in Interaction Design

KEVIN BJÖRKLUND

SIMON ELIASSON

Department of Computer Science and Engineering
Chalmers University of Technology
University of Gothenburg
Gothenburg, Sweden 2018





Master’s thesis 2018

Augotchi: The Augmented Pet

Exploring virtual companion design in Augmented Reality games

KEVIN BJÖRKLUND

SIMON ELIASSON

Department of Computer Science and Engineering
Chalmers University of Technology

University of Gothenburg
Gothenburg, Sweden 2018



Augotchi: The Augmented Pet
Exploring virtual companion design in Augmented Reality games
KEVIN BJÖRKLUND
SIMON ELIASSON

c© KEVIN BJÖRKLUND, 2018.
c© SIMON ELIASSON, 2018.

Supervisor: Sofia Serholt, Interaction Design, Department of Computer Science and En-
gineering
Examiner: Staffan Björk, Interaction Design, Department of Computer Science and En-
gineering

Master’s Thesis 2018
Department of Computer Science and Engineering
Chalmers University of Technology and University of Gothenburg
SE-412 96 Gothenburg
Telephone +46 31 772 1000

Typeset in LATEX
Gothenburg, Sweden 2018

3



Augotchi: The Augmented Pet
Exploring virtual companion design in Augmented Reality games
KEVIN BJÖRKLUND
SIMON ELIASSON
Department of Computer Science and Engineering
Chalmers University of Technology and University of Gothenburg

Abstract

When designing a digital game, the possibilities in terms of functionality and visual design
are endless, which makes it difficult to find the optimal solution. This is especially true
when exploring the possibilities of new technology or trends, where standards are yet
to be established. One such technology is Augmented Reality (AR), which combines
digital content with information gained from the real world. The focus of the Augotchi
project is the design of a game where the core feature is a virtual companion. The other
essential core feature of Augotchi is the global positioning system (GPS) integration and
the utilization of real world geography, putting an AR aspect into the game. This project
explores how to design a virtual pet being featured in a location-based AR game for
mobile devices.

This thesis covers topics such as provoking emotional engagement in users, involving users
in the design process through co-design workshops and evaluating the results. In this
report, in depth descriptions of the methods used will be explained, the tools that were
utilized will be covered and an extensive description of the final result will be presented.
This report provides an overview of what was found in relation to developing such an
application, and what questions may arise during the development of a similar game. One
of the main topics of discussion revolves around the largest problem that arose during the
project, which was the acquisition of external testers. Other topics of discussion include
the implications of various features and the impact that location-based AR games might
have on society.

Keywords: Augmented reality, virtual companion, Tamagotchi, application development,
co-design, game design, prototyping, iterative development

4





Acknowledgements

Firstly, we would like to thank our supervisor, Sofia Serholt, at the Division of Interaction
Design within the Department of Computer Science and Engineering, Chalmers Univer-
sity of Technology, for supporting us throughout this project. When difficulties arose,
especially regarding the acquisition of testers, Sofia encouraged us to pursue the original
goals of the project, which helped us maintain some of the fundamental themes of the
thesis. She has been thorough in her feedback, both regarding the process and our report
writing, which helped us stay on track, and which we have greatly appreciated.

Another big thanks goes to the eight individuals who took their precious time and par-
ticipated in the co-design workshops, as well as the evaluation session. These people
also provided us with valuable feedback throughout the development. Without these
contributions, this thesis would not have been possible.

We would also like to thank our colleagues at Far North Entertainment for providing us
with an office space, as well as the web hosting service used for Augotchi. Furthermore,
they allowed us to take time off from work, whenever needed, in order to complete this
thesis.

Finally, special thanks goes to our family, relatives and friends who have encouraged us
and supported us through all those years of studies at Chalmers University of Technology.
We would not be able to make it without you.

Kevin Björklund & Simon Eliasson, Gothenburg, 2018

6





Contents

List of Figures 11

1 Introduction 1
1.1 Aim . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Delimitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Stakeholders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

2 Background 3
2.1 Digital Reality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Mobile Augmented Reality Technology . . . . . . . . . . . . . . . . . . . 4
2.3 Augmented Reality in Mobile Games . . . . . . . . . . . . . . . . . . . . 4
2.4 Formal Game Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2.4.1 Formal Game Analysis Conclusion . . . . . . . . . . . . . . . . . 6

3 Theory 9
3.1 AR - Benefits and Problems . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.2 Emotional Engagement in Entertainment Media . . . . . . . . . . . . . . 10
3.3 Relationships with Virtual Companions . . . . . . . . . . . . . . . . . . . 13
3.4 Virtual Creature Graphical Design . . . . . . . . . . . . . . . . . . . . . 15
3.5 Researching Emotional Engagement . . . . . . . . . . . . . . . . . . . . . 18

4 Methodology 19
4.1 Concept Ideation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

4.1.1 Formal Game Analysis . . . . . . . . . . . . . . . . . . . . . . . . 19
4.1.2 Quick online surveys . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.1.3 Brainstorming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.1.4 Participatory Design and Brainstorming . . . . . . . . . . . . . . 21

4.2 Concept Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
4.3 Concept Creation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.4 MVP Development and Vertical Slicing . . . . . . . . . . . . . . . . . . . 24
4.5 The Agile Development Process . . . . . . . . . . . . . . . . . . . . . . . 24

4.5.1 Result Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.5.2 Forming New Requirements . . . . . . . . . . . . . . . . . . . . . 26
4.5.3 Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.5.4 Iteration Stop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

4.6 Interview Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.7 Ethical Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

5 Planning 30
5.1 Conceptualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.2 MVP-development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.3 Iterative development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.4 Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

6 Process and Execution 32

8



6.1 Preparation and Research . . . . . . . . . . . . . . . . . . . . . . . . . . 32
6.1.1 Establishing Roles . . . . . . . . . . . . . . . . . . . . . . . . . . 32
6.1.2 Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

6.2 Utilized Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
6.2.1 Platform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
6.2.2 Software Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . 34
6.2.3 Media Creation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

6.3 Methodology Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
6.4 Quick Surveys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
6.5 Conceptualization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

6.5.1 Co-Design workshop I . . . . . . . . . . . . . . . . . . . . . . . . 44
6.5.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
6.5.3 Feature Categorization and Election . . . . . . . . . . . . . . . . . 51

6.6 MVP Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
6.6.1 Version 0.1 - First Implementation . . . . . . . . . . . . . . . . . 52
6.6.2 Version 0.2 - Expanded MVP . . . . . . . . . . . . . . . . . . . . 53

6.7 Iterative Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
6.7.1 General methods and Documentation . . . . . . . . . . . . . . . . 55
6.7.2 Version 0.3 - Player Data Collection . . . . . . . . . . . . . . . . . 56
6.7.3 Version 0.4 - Player Convenience Update . . . . . . . . . . . . . . 56
6.7.4 Recruiting Testers: Wave 1 . . . . . . . . . . . . . . . . . . . . . . 57
6.7.5 Co-Design Workshop II . . . . . . . . . . . . . . . . . . . . . . . . 58
6.7.6 Recruiting Testers: Wave 2 . . . . . . . . . . . . . . . . . . . . . . 67
6.7.7 Version 0.5 - Augotchi 2.0 . . . . . . . . . . . . . . . . . . . . . . 68
6.7.8 Recruiting Testers: Wave 3 . . . . . . . . . . . . . . . . . . . . . . 69
6.7.9 Version 0.6 - Dungeoneering & Content . . . . . . . . . . . . . . . 70
6.7.10 Version 0.7 - Garden Decoration . . . . . . . . . . . . . . . . . . . 70

6.8 Evaluating the prototype . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

7 Result 74
7.1 Augotchi Prototype . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

7.1.1 General Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
7.1.2 The World Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
7.1.3 The Pet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
7.1.4 Pet Customization . . . . . . . . . . . . . . . . . . . . . . . . . . 77
7.1.5 The World Markers . . . . . . . . . . . . . . . . . . . . . . . . . . 78
7.1.6 Currency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
7.1.7 Pet Accessories . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
7.1.8 Quests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
7.1.9 The Ranch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
7.1.10 Produce . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
7.1.11 Garden Decorations . . . . . . . . . . . . . . . . . . . . . . . . . . 84
7.1.12 Dungeon Exploration . . . . . . . . . . . . . . . . . . . . . . . . . 85

7.2 Evaluation Session Results . . . . . . . . . . . . . . . . . . . . . . . . . . 86
7.2.1 TLX form results . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
7.2.2 Interview results . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

9



7.3 Quantitative data analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 91

8 Discussion 93
8.1 Process Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

8.1.1 Prototype development . . . . . . . . . . . . . . . . . . . . . . . . 93
8.1.2 Co-Design & Evaluation . . . . . . . . . . . . . . . . . . . . . . . 94
8.1.3 Tester Acquisition . . . . . . . . . . . . . . . . . . . . . . . . . . 96

8.2 Result Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
8.2.1 Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
8.2.2 Prototype Features . . . . . . . . . . . . . . . . . . . . . . . . . . 99
8.2.3 Evaluation Data . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

8.3 Project Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
8.4 Generalizability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
8.5 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104

8.5.1 User Testing in AR . . . . . . . . . . . . . . . . . . . . . . . . . . 104
8.5.2 Concept Considerations . . . . . . . . . . . . . . . . . . . . . . . 105
8.5.3 Future Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

9 Conclusion 107

References 109

Appendices 118

A Formal Game Analysis Design Patterns 118
A.1 Design Patterns of Ingress . . . . . . . . . . . . . . . . . . . . . . . . . . 118
A.2 Design Patterns of Pokémon GO . . . . . . . . . . . . . . . . . . . . . . 121
A.3 Shared Principles of Design . . . . . . . . . . . . . . . . . . . . . . . . . 124
A.4 Unique Principles of Design . . . . . . . . . . . . . . . . . . . . . . . . . 125

B Augotchi Content Tables 128

10



List of Figures

1 Mori’s uncanny valley graph [1] . . . . . . . . . . . . . . . . . . . . . . . 16
2 Player warnings in Pokémon GO . . . . . . . . . . . . . . . . . . . . . . 28
3 Summary of the Pokémon GO quick-survey responses. . . . . . . . . . . . 41
4 Summary of the Ingress quick-survey responses. . . . . . . . . . . . . . . 42
5 Result of the grid brainstorming round. . . . . . . . . . . . . . . . . . . . 46
6 The setup for participants investing ”currency” in the various ideas . . . 49
7 The participants evaluating pros and cons of various themes . . . . . . . 59
8 Refinement of the various themes . . . . . . . . . . . . . . . . . . . . . . 67
9 A representation of the pet coloring scheme . . . . . . . . . . . . . . . . 78
10 Augotchi coin icon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
11 Building material icon . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
12 Depiction of the in-game Ranch and garden of Augotchi . . . . . . . . . . 83

11



1 Introduction

The company Far North Entertainment, which both thesis students are part of, is a
startup company active in the video game industry. This being the case, the company
is looking for new concepts and prototypes that might turn into its next big project; it
has so far specialized in virtual reality (VR) games, but is also interested in exploring
new areas within digital entertainment. Augmented reality (AR) is one such area that
has shown a lot of promise lately, and it is also related to VR, which makes it a natural
subject of interest for Far North Entertainment. One of the biggest hits in mobile game
history is Pokémon GO, an AR mobile game in which players collect magical creatures
found in the real world, which became a worldwide phenomenon upon its release [2, 3, 4].
Not only did it reach over half a billion downloads in just two months [2], it also became a
worldwide news story [4], bringing AR games the attention of the masses. In short, there
seems to be huge potential in the mobile AR market, which means it is also interesting
from a commercial standpoint.

The thesis will therefore consist of developing a prototype of a mobile AR game. The
name of this prototype is Augotchi, which is a combination of the words augmented, as
in augmented reality, and Tamagotchi [5] , which is the game that has provided the main
inspiration for Augotchi’s gameplay. Mival et al. [6] describes the Tamagotchi as an
egg-shaped electronic device that features a virtual egg, which hatches and becomes a
creature that needs to be taken care of by its owner. A virtual pet, so to speak. The
name Augotchi thus reflects the two components which are studied in this thesis.

1.1 Aim

The thesis will revolve around exploring AR game design in order to create a better
understanding of how such an application should be developed. The main aspect of the
application, and hence the thesis, will be to create a game in which the player’s relation
to a virtual pet is the core feature. For example, in Pokémon GO, thousands of creatures
are collected, filtered by power level and then discarded by the player. There seems to
be little emotional connection between the player and the creatures he/she has caught.
What if all the player’s energy was to be funneled into one single creature instead, similar
to a Tamagotchi? [7] Does this have a significant impact on how the player views the
game and what it is about? These are just a couple of the questions that have been
fundamental to the initial idea of the thesis, which research question is as follows:

What game components are important when designing a virtual pet for mobile
AR games?

The main goal of the thesis has been to formulate a set of guidelines and/or pitfalls that
could help developers in designing an AR application built to emotionally engage people.
While this is the means of achieving the main goal, the aim is to explore ways to develop

1



a successful entertainment product. Note that it is not obvious that, just because an
application is emotionally engaging, it will thus lead to an improved user experience.
However, this is the underlying assumption which resulted in the overarching theme of
Augotchi.

In order to find answers to the research question, a mobile game prototype was de-
veloped. Throughout its development, both qualitative and quantitative data from user
testing, co-design workshops and evaluation sessions were gathered, in order to later draw
conclusions regarding what is good or bad practice when developing an application that,
to some degree, relies on emotionally engaging the user. In this report, the study of
the theory surrounding relevant subjects is presented, as well as the methods for how
to perform the development and user research. Furthermore, the technical foundation of
the game is described. This is followed by the three main chapters which describe the
process, the results of the thesis and a discussion surrounding the project.

1.2 Delimitations

The result of the thesis is a mobile game prototype which features a virtual pet, AR
being the fundamental technology used. The content of the game, and the technology
utilized, are of course highly intertwined. Hence, one goal of the thesis will be to find
out in what ways the two may affect each other, and determine how to capitalize on it.
Virtual reality (VR) and mixed reality (MR) are other technologies that were considered
early on, but due to time constraints, VR and MR are not part of the prototype. VR and
MR are further described in Section 2.1. AR features that revolve around the projection
of the virtual pet in the camera was also considered, but due to technical constraints,
which would complicate the recruitment of testers, this feature had to be neglected.

1.3 Stakeholders

The stakeholders for the project are the following:

• Far North Entertainment - As described in the introduction.

• The students - As part of Far North Entertainment, and as students who have
acquired new knowledge and experience through their work on the thesis.

• The Users - The people who have tried out the Augotchi application, and who will
potentially use it in the future.

2



2 Background

In this section, the background of AR mobile games will be briefly presented. The
technology will be more thoroughly described and compared with similar concepts, and
some of the mobile AR games that already exist on the market are presented.

2.1 Digital Reality

Before describing in detail what AR is and how it can be used, the different forms of
digital realities will be briefly presented. These are Augmented reality (AR), Mixed
reality (MR) and Virtual reality (VR). In this thesis, the focus has been AR technology,
with some influences of MR. VR has not been touched upon at all, but the comparison
of the different concepts might clarify what AR is, and what it is not. AR can be defined
as:

“Computer-generated content overlaid on a real world environment” [8]

This is done by either utilizing data from the real world to dictate what happens in an
application (for example GPS data or camera data), or by projecting digital or virtual
images/objects on top of the real world through the use of a camera or similar. In contrast
to AR, is Mixed reality. MR can be defined as:

“Virtual spaces where real world objects or people are dynamically integrated
into virtual worlds to produce new environments and visualizations where
physical and digital objects co-exist and interact in real time” [9]

This may be achieved by, for example, mapping the position of an object and placing a
similar looking object in the virtual environment, with the same relative position to the
player as the real world object. The benefit of this is that the user can physically interact
with the virtual and real object (the “MR” object) in the virtual world without seeing
the real object at all, creating the illusion that the virtual object is actually real.

The line between AR and MR is not clearly defined, however, and there is no exact
description of what distinguishes the two. One way of thinking about it, based on the
previous quotes, is that AR tends to enhance the real world with virtual data, while MR
tends to mix real world data and virtual data to bring forth a whole new experience [10].
However, the usage of the term MR seems to become increasingly obsolete, in favour of
AR, effectively merging the two. The reason behind this is, as mentioned previously, that
the line between AR and MR is quite diffuse. Hence, people tend to call all forms of both
MR and AR, just AR [11].

The third concept is VR, which is more straightforward to explain. In VR, there is
no connection to the real world whatsoever. The user, of course, still exists in real space

3



and might use hardware controllers, motion-tracked or not, but this is not enough to
deter from the essence of VR. As long as the digital world that the user is immersed in
lacks real-time information about the real world, it is called VR [12].

2.2 Mobile Augmented Reality Technology

AR, the transferal of real-time data, from the real world, into the contexts of digital appli-
cations, has become popular in digital entertainment, and in computer science in general
during the last couple of decades [13]. Human-computer interaction has traditionally
been based on the mouse and keyboard setup, which is proven effective for navigating
through complex systems, managing data and playing games designed for the setup, but
which is also limiting in how and when it is used [14].

With the arrival of smartphones, some of the limitations regarding where and how digital
applications are used has been diminished, and this has opened up possibilities in what
these applications may do. However, many mobile applications still function similarly to
those on computers, with an on-screen keyboard for data input, and finger taps acting
as “mouse” clicks. But smartphones do have something that desktop computers, and
even laptops, do not have: a comparatively high level of portability and ease of access,
making them usable anywhere at any time without much effort from the user. Combined
with a wide array of technological features, including cameras, GPS, haptic feedback and
internet access, the smartphone is more versatile than a traditional computer [15, 16].

The versatility of the smartphone has opened up possibilities in designing digital ap-
plications, one particular branch being AR applications. The fundamental idea of AR
is to take real-time data from the real world, such as images or video, and project a
digital layer onto it, hence, bringing virtual information to the real world: augmenting
reality [13]. A classic GPS navigation system can be seen as an AR application, as it
presents the user’s position in the environment of the real world, with extra functionality
that aids the user in finding her way to a desired destination. So, although this concept
has been around for decades, it has never been as accessible as now. Some of today’s
applications use GPS and camera data for certain features, including navigation, public
transit travel planning and social networking. Furthermore, AR has also been utilized in
the development of games [17].

2.3 Augmented Reality in Mobile Games

One of the pioneers of the mobile AR games industry is Niantic Inc [18], which developed
the games Ingress [19] and Pokémon GO [3]. Ingress, a game in which the players act as
agents and battle each other in a global war for domination, has more than ten million
downloads on Google Play alone, which is a respectable number. But few, if any, games
can compete with the initial success of Pokémon GO. Only two months after its initial

4



release, Pokémon GO was reported by Niantic CEO John Hanke to have been down-
loaded over half a billion times [2]. Niantic approaches the AR concept from a new angle.
Rather than projecting a digital layer upon the real world in order to provide useful
information more effectively, they use the data in order to transform the real world into
something completely different. This is the case in both Ingress and Pokémon GO, where
the game world is set in the real world in terms of geography, but where the contents of
the game world are completely fictional. Suddenly, energy portals and magical creatures
are literally scattered across your neighborhood, and with it comes “real” secret Ingress
agents and Pokémon trainers that you can meet and interact with [3, 19]. The success of
Pokémon GO has proven that AR games have great potential, and that people in general
are open to the concept of playing a game in the “real world”. However, much is yet to
be discovered, as the genre is relatively young.

Another game genre that exists in a number of variations on a multitude of platforms is
the Tamagotchi, a digital pet simulator which features a virtual pet, with needs that the
player has to meet in order for the pet to survive and thrive [20]. This genre has also
been adapted in AR, vaguely in Pokémon GO which arguably shares some similarities
with the Tamagotchi, but also more directly, such as in Curiopets for iOS, a Tamagotchi-
Pokémon GO hybrid, which features real world location integration and projection of the
pet within the camera image [21].

The combination of the Tamagotchi and AR makes for a game concept with a num-
ber of variables that might affect the user experience and the motivations of its players.
The progression system for both player and pets, based on experience gathering, item
collection and the growth of one’s digital companions, is one such aspect. Competitive
features, such as battling in Pokémon GO or playing games with friends in Curiopets, is
another one. Social interactions with other people, that might occur when playing in the
field, is a third one.

Another possible one, which is the focus of this thesis, is the emotional investment and
attachment that the player has to his/her pet(s). How does this affect the player’s moti-
vation to keep playing the game? If the pet is at risk of becoming sick due to inactivity of
the player, will this motivate the player to take an extra walk around the block in order
to keep the pet healthy? What variables matter when the player decides to, either care
for the pet, or leave it be? Is there any emotional investment at all from the player’s
side in such a game, or is it just pure entertainment value that keeps him/her interested?
These are just a handful of questions that could be valuable for any game development
project with Tamagotchi influences.

2.4 Formal Game Analysis

Exploring already existing applications within the AR game genre was deemed important
in order to understand what has already been done within the field. Such background
research proved valuable insights throughout the Augotchi project, since it gave us some-

5



thing concrete to compare the prototype with, as well as a rough guideline when making
decisions regarding the features of the game.

The background research, in this case, consisted of a formal game analysis in which
the game features of Pokémon GO and Ingress were taken apart and analyzed in order
to identify their building blocks in terms of game play patterns and game mechanics.
This section will present the results of this formal game analysis. The patterns have been
derived from analyzing the different components of the two games. The different patterns
are described in a generalized way. In some cases, patterns are hard to describe without
using in-game terminology, and therefore, a vocabulary for these terms will be provided
for each game. There will be some design principles of both games that are identical, or
at least similar to a great extent, and these will be further discussed in the conclusions
section of the analysis.

2.4.1 Formal Game Analysis Conclusion

Pokémon GO and Ingress have a lot in common. This is no surprise, since the games
belong to the same genre and are made by the same people. In essence, the games are
the same in most aspects, but with two major differences. In Ingress, the player must
compete with other players in order to progress in the game. To do this more effectively,
the player has to interact with other players. This is something that is not necessary
in Pokémon GO, which allows the player to play on his/her own and still reach the late
game. The second major difference is the collecting aspect of Pokémon GO, where the
major part of the game revolves around expanding and improving the player’s personal
collection of Pokémon.

Both games could almost be described as having two main phases when it comes to the
GPS map game play, one which is largely identical, and one which differs a lot. The
common phase is the resource gathering phase, which has the player move around to
different portals, or pokestops, in order to gather items from them. Items have the same
role in both games, and act as fuel for the player, since most player actions require these
items. The second phase is related to how items are used. In Ingress, items are exclusively
used in order to either build up portals for the allied faction or destroy enemy portals,
and items are meant to be consumed and used as effectively as possible. In Pokémon GO,
some items work in a similar way, but the effects of some items are also long-lasting. For
example, In Ingress, you might build a strong portal with rare items, only to see it being
destroyed the next second. Ultimately, your consumed items did not have a significant
impact. In Pokémon GO, if you spend a lot of stardust, candy and other valuable items
in order to empower one of your Pokémon, it is a no-risk, long-term investment. Augotchi
might be a combination of the two, and it will be important to keep both aspects fun
and challenging:

• Since Augotchi will feature a pet which needs to be constantly taken care of, it can
be viewed as a constant battle against the system, where the system continuously

6



draws the pet closer to death, and the player continuously tries to protect it. This
is similar to the constant battling in Ingress, the difference being that the decline of
one’s pet’s health will be predictable, while the actions of enemy players in Ingress,
are not.

• Since the goal of Augotchi will be to keep one pet alive and thriving, it is also a
long-term investment, like in Pokémon GO. As long as you do a good job in taking
care of your pet, there is no risk of losing it, or seeing it get hurt. The question
is then: what happens beyond just keeping the pet alive? How far up the positive
axis can you push it?

One noteworthy similarity between the games is that they both lack a win or lose state.
The reasons behind this are not obvious, and are subject to speculation. Ingress and
Pokémon GO have both been on the market for years, and some of the people that
started playing on day one still continue on the very same game session. The games are
designed in a way which makes them, in practice, impossible to fully complete. It seems
like Niantic has decided that the classic aspect of “beating” the game does not fit the AR
game genre, and that the developers want their players to be committed to the game in
the long term. Of course, when you develop a game which is free and relies on in-game
purchases, you want your players to never stop playing, so that you maintain a steady
stream of income. But it does not explain why there are no rounds in the game, for
example, by restarting the game state once every few months. If Augotchi is to feature
the death of the virtual pet, this design principle is lost, but it is hard to pinpoint what
it would mean for the players and the gameplay. In other words, two questions worth
contemplating are:

• Should it be possible for the pet to die in Augotchi, given what is seen in Niantic’s
games, where no end state exists?

• If death may occur, how much does it take before it happens?

The way items work in both Pokémon GO and Ingress is also worth contemplating. The
idea of having different rarities and power levels of items might be useful for Augotchi as
well, if items are to be featured. In any case, there has to be a reason for the players to
move around in the real world in order to justify the AR mechanics, and the collection
of necessary items is a seemingly effective way to do this. The random generation of
items is also an interesting aspect, since it makes the item gathering less predictable, and
more exciting. The limited inventory space is probably connected to the issue of making
people walk as well, since it makes sure that a player cannot intensely gather items for
a couple of days, in order to not gather items again for another month. The constant
character progress of both games might also be related to maintaining the motivation of
the players. The issue of motivating the players to move around in the world is one that
Niantic has obviously put a lot of thought into solving, and the Augotchi concept will
have to solve this issue as well. Hence, a couple of questions need to be answered:

• What are the fundamental motivators for a player to move around in the real world?

7



• How can the motivations of the player be maintained over time?

• What are the micro-goals and macro-goals of Augotchi?

Macro-goals are essential to any game; it is what the player ultimately strives to achieve
and defines the point of taking actions within the game. Ingress only has one identified
macro-goal, and Pokémon GO has a number of possible goals that might be interesting
to the player. An important structure to define for Augotchi is what macro-goals the
players might strive towards, as well as what micro-goals they need to pursue in order to
get closer to said macro-goals.

A full list of all design patterns found in the formal game analysis can be found in
appendix A.

8



3 Theory

The theory section of this thesis will be dedicated to presenting ways in which a person
may be emotionally affected by entertainment and games, and how this impacts said
person’s willingness to engage in certain activities. This section will also present already
concluded research regarding relationships forming between people and virtual creatures,
as well as how to design digital applications featuring a virtual creature in order to
enhance emotional engagement.

3.1 AR - Benefits and Problems

The concept of AR differentiates from traditional digital applications in a number of
ways. In this section, the possibilities and limitations of AR will be presented in order
to better understand the implications of such technology.

One benefit of AR applications that encourage physical activity, is the positive impact
they have on their players. According to the World Health Organization, physical inac-
tivity is the fourth leading risk factor for preventable death worldwide [22]. With physical
inactivity being such a major health problem worldwide, many search for answers for how
to solve this. LeBlanc and Chaput write:

”In July 2016, Niantic released what may be the most successful population
level physical activity program that we have seen in modern history.” [23]

This is an astonishing observation, not only as a credit to Niantic as the developer of
Pokémon GO, but to the AR game market as a whole, illustrating the potential of the
genre not only as an entertainment medium, but as a health benefactor. The interest
in Pokémon GO decreases over time, just like any newly released game. However, data
from Microsoft shows that the game may increase physical activity levels of its players
by about 25%, which is a substantial increase [23].

While increasing the level of physical activity on an individual level, AR games are
beneficial on a societal level as well. One of the areas in which the AR game genre might
be helpful is with the activation of suburban areas that were previously not utilized to
their intended potential:

”Urban planners and designers have spent the last 50 years trying to activate
unused public spaces, create walkable cities and encourage sociability through
urban design. Pokémon GO has succeeded, almost overnight, to entice people
of all demographics into the streets of cities around the world. In fact, many
previously underutilized public spaces have suddenly become hot spots for
all demographics, playing Pokémon GO and other similar augmented reality
games (ARGs)” [24].

9



This means that the players of AR games are much more likely to wander areas where
ordinary people would not often visit at all. This is beneficial for several reasons, the two
most important reasons being that people see more of the city, and maybe visit stores that
would otherwise suffer from low visitor counts. The second reason is that it might have
people taking different routes, which means that the main routes that might otherwise
be cramped and overused, will get some relief.

Even though there seems to be benefits to AR games, they might also pose problems.
A new form of advertising illustrates the potential issues of AR. Due to players of AR
games searching for virtual items at set locations, companies that benefit from a higher
customer count may contact the game developers and make deals with them, in order to
have them create in-game sites at the company’s grounds, directing players to the shop
through the game. Liberati writes the following:

”We should not underestimate this contaminations between realities because
it is a way to change our world indirectly. For example, this kind of hybridisa-
tion between realities could be used in order to change the relevance of places
in the paramount reality by someone. A big brand could make agreements
with the designers of these games in order to add value to specific places
they are interested in like stores. By making a store a valuable place where
Pokémon often appear, the importance of that place in the paramount reality
changes according to the new value acquired in the game. McDonalds is one
of the sponsors of Pokémon GO in Japan and they have agreement in turning
every McDonald store into a “valuable” place for the game” [25].

While not an issue in itself, it may become an ethical issue if the user does not realize that
this is happening. More problems that could arise in both a personal and grander scale
is if users who are immersed in AR games do not pay attention to the world around them
while playing. Even if most AR games have heavy warnings for this kind of behaviour,
it is still non-negligible that games like these will cause accidents. Studies estimates that
the AR game Pokémon GO caused over 100,000 traffic accidents during its first year [26].
The notion that AR games can be dangerous is important to keep in mind. However, in
a game that features GPS integration, people will inevitably be active in trafficked areas,
which inevitably will lead to accidents.

3.2 Emotional Engagement in Entertainment Media

Emotional engagement is a vague term with a number of possible meanings. The emo-
tional engagement of Augotchi refers to the virtual pet and its ability to emotionally
engage the player. Hence, the term will, for the sake of this thesis, be defined as:

The involvement of an agent, through the occupation and attraction of said
agent’s attention, by affecting that agent emotionally.

10



According to Bartsch & Viehoff [27], there are a number of reasons why people seek
gratification by consuming media entertainment, reaching from simple emotional arousal
to more abstract motivations such as self-improvement or relationship-forming. Some of
the motivators on their list are:

Mood-management - Influencing one’s mood, for example, by watching an action movie
when bored, or simply keeping yourself distracted from negative thoughts.
Excitation transfer - The notion that high suspense, created by fear of bad outcomes,
results in a gratifying feeling of relief when the ending is, in fact, happy.
Sensation seeking - Similar to mood-management and excitation transfer, but with-
out any specific goal. Any sensational experience that triggers a particular emotional
response to a high degree may be entertaining in itself.
Intrinsic motivation - The capabilities of the consumer of any media, including com-
puter games and watching movies, should be challenged by the media in a balanced way,
since this is essential to the consumer’s willingness to continue the activity.
Relationship functions of entertainment - Entertainment has an ability to constitute
different forms of relationships, such as parasocial relationships; a one-way relationship
between audience and media figures, that act as an extra source of social and emotional
gratification.

Bartsch & Viehoff [27] further classify said factors into two categories which are related
to emotions and functionalities respectively. Making friends and letting out emotions
that do not fit in everyday life situations, are examples in the functional category, with
the emotional category containing the sensations of fun, thrill and being moved. These
are not the only emotions that might be influenced by media, however.

Riva et al. [28] found, in their study of the link between emotions and presence in
VR, that feelings of sadness, happiness, anxiousness and relaxation may be substantially
affected by the nature of an environment in VR. They also found that a test subject’s
presence in the simulation increased with the magnitude of the emotional effect that the
environment had on them. In other words, a neutral environment did not affect people
emotionally, and therefore, they felt less present in the simulation. On the contrary, an
anxious environment had a significant impact on both emotions and feeling of presence.

One way of getting people involved in media entertainment is the connections they share
with the characters therein. Moyer-Gusé [29], in her analysis of entertainment-education
in story-driven media, lists parasocial interaction and liking as two of five ways in which
people get involved with the characters in media. Liking refers to the initial phase of
involvement, where the audience member judges a character and determines, inter alia,
whether the character is likable or not. Likeable characters may then be subject to
parasocial relationships, which, as mentioned previously, refers to a kind of bond be-
tween a person and a media figure. It has been shown that people have the capability to
perceive figures, with which they have formed parasocial relationships, as role-models, or
even friends. Furthermore, she proposes that a character, with which a person has formed

11



a parasocial relationship, increases the persuasive power of entertainment-education, and
hence, its ability to influence, amongst other things, the behaviour and attitude of said
person.

Choi et al. [30] further strengthen the notion that bonding with characters is impor-
tant when trying to achieve a higher level of emotional engagement, and that human
beings tend to treat characters as social actors if said characters are showing signs of
emotion, either by social cues or facial/body expressions. They further claim, in their
analysis of facial expressions and engagement in decision making games, that:

“These findings are tantalizing, because they reinforce more general findings
that people treat computers as social actors when they include appropriate
social cues.” [30]

These findings are also in line with a theory known as the ”Media Equation Theory”
[31], which states that people tend to treat characters and places in media as if they were
real people or real places. They further claim that people view computers, running even
relatively simple code, as social actors. This relates to the concept known as anthropo-
morphism, which is presented further in Section 3.3.

In addition to the social aspects of consuming media entertainment, some of the fac-
tors of Bartsch & Viehoff’s list are described as vital, for virtual games specifically, by
Boyle et al. [32]. Challenge, exploration of virtual worlds and opportunities to form
relationships are some of the pathways to game enjoyment, with challenge, followed by
emotional arousal and excitement, being rated the most important by young people aged
18-24. Other motivations mentioned are information seeking and social interactions, as
well as an increased enjoyment from knowledge of the subject of the game, as when play-
ing a sports game while having extensive knowledge of said sport. Coulson et al. [33]
compiles a list of motivations, when playing virtual games, in their analysis of emotional
attachment and interpersonal attachments in video games. The list can be narrowed
down into the following three categories:

Social motivations - Such as wanting to chat with other people, engage in activities
with friends or even find new friends. This kind of person tends to be drawn to games in
order to interact with others.
Achievement motivations - The desire to achieve something in games. Be it further
developing a character by gaining levels, gathering equipment or completing a difficult
task, such as defeating a hard boss, gathering all stars in a level or beating a worthy
opponent in player versus player combat.
Immersion motivations - A willingness to immerse oneself in the world and story de-
velopment of a game, as well as a desire to continue playing in order to see more of the
story and take part in the plot. Immersion could also be fueled by curiosity, which makes
the player want to further explore the world of the game and discover everything it has
to offer. Another motivation is to avoid stress or problems the user experiences in real
life.

12



The authors continued by studying the social and emotional attachments the players
form with the characters of a game over time, and monitored the statistics. In their anal-
ysis, they conclude that researching relationships within virtual environments is, indeed, a
legitimate endeavour [33]. It is important to note, however, that within the game that the
authors used in their analysis, the characters that the test subjects formed attachments
to were all human-like characters. Since the virtual pet in the Augotchi prototype will
not necessarily be a humanoid creature, it is important to establish what the implications
of a non-humanoid, and even non-living, pet would be.

3.3 Relationships with Virtual Companions

In 2006, in her research on mobile phones and their use, Jane Vincent [34] found that
users may develop emotional attachments to their phones. She could identify an increase
in gratification and satisfaction of users when they had their phones with them, and she
could see an increase in anxiety, and even panic, when the users were apart from their
phones. She further claims that phones have become part of people’s lives, and that
people share their problems, and even happiness, with their phones. It is interesting to
note that this research was conducted long before the big breakthrough of smartphones
[35]. Today, the number of smartphone users have increased significantly [36], and it is
also easy to imagine that more powerful and versatile phones lead to an even greater
attachment. It highlights the dominant role of technology in our lives, going from being
a tool to being a substitute for human relationships [7]. The notion that people can form
emotional bonds with an inanimate object, and not just humanoid beings, is valuable
information. However, bonding with characters in entertainment, and bonding with a
useful, technological device, cannot be assumed to depend on the same factors.

In any case, it has been established that emotional bonds may appear between human
beings and characters in entertainment, as well as inanimate objects. A digital compan-
ion would be some kind of hybrid, an inanimate object which is simulated to express the
behaviour of living things. What are the specific characteristics of human interaction
with such a hybrid?

When human beings look at things they do not fully understand, they tend to apply
whatever attributes they recognize in similar things, to the unexplainable objects. In
extension, when looking at digital or robotical, animated objects, there is a tendency
for human beings to relate what they see in the behaviour of the object to what they
have seen prior in living things, such as animals, or even other human beings [37]. This
method of applying known patterns of living things, to behaviour human beings per-
ceive, but cannot explain, in non-living things, is called “anthropomorphism” [38], and it
is fundamental to being able to form relationships with inanimate objects. An extreme
example of anthropomorphism, is the “pet rock”. In one study, 106 participants were
each given a rock, and were asked to decorate it [39]. Approximately half of the partici-
pants painted a face on the rock, and about half of the participants claimed to have given
the rock a personality, to some degree. The authors try to explain this outcome in the

13



following quote:

“A possible explanation for the self extension results could lie in the form
factor of the rock itself. That is, perhaps the spherical rock suggested a
head, which in turn led participants to anthropomorphize the rock and attach
personal meaning to it that they would not have attached to another less
humanlike product“ [39].

This hints that all it takes for a human being to anthropomorphize an object, is for it to,
for example, have a round headlike shape. Indeed, the authors also claim that:

“We showed that the simplest of activities, creating a design on a ‘pet rock,’
can lead to feelings that the object symbolizes the self” [39].

If a rock can be perceived as something more than just an inanimate piece of matter,
may it also qualify for relationship forming with human beings? Jen Wrye [40] argues
that, what constitutes a pet is not simply the relationship status between human being
and living entity, such as a man owning a dog, but that an object’s “petness” is based on
emotional investment into said object. Furthermore, she claims that there is no inherent
petness to any object. Hence, being a dog does not qualify more as a pet than being
a rock with a painted-on face. In fact, the opposite could be true, since a rock with a
painted-on face has an obvious connection to human investment, while an arbitrary dog
might have had no contact with human beings whatsoever. In short, viewing human-pet
relationships as a social construct, rather than a strictly defined relationship setup, makes
it possible to view any object, living or dead, as a pet.

However, there are differences in how different objects are perceived, which might af-
fect an object’s potential to become, for example, a pet. A study made by Kahn et al.
[41] shows that children treat a stuffed dog differently than an animated robotic dog.
A particularly interesting finding was that, during the interviewing of the children, the
results were very similar in how they thought they treated the robot and the stuffed dog.
However, when they compared the actual recorded data, there were clear differences. For
example, the children were much more prone to mistreat the stuffed dog. One explanation
for this could be that the children have previous experience with soft and hard objects,
and know that soft objects, in general, are harder to break. Additionally, the children
could have been taught to be more careful around electronic devices. Although treating
the robots more carefully, when asked, the children answered that they believed both dogs
could hear to the same extent, but when actually looking at the data, the children were
significantly more prone to give the robot verbal commands. Regarding these findings,
the following statement was made, “AIBO” referring to the robotic dog:

“Children also appeared to believe that AIBO was the sort of entity with
which they could have a meaningful social (human–animal) relationship. Specif-
ically, over three-quarters of the children said that they liked AIBO, that
AIBO liked them, that AIBO likes to sit in their lap, that AIBO can be their
friend, and that they could be a friend to AIBO“ [41].

14



They continue by concluding that, while most of the children knew that the robotic dog
was not a real living dog, it did not stop the children from conceiving and treating the
robot as such. This further proves that when robots and computers show even minimal
signs of social cues, humans tend to treat the computer as if it was a social agent [41].
That a minimal amount of social cues from a virtual object can affect the relationship
forming between itself and its owner is further strengthened by Mival et al. [6]. They
describe the interactions of the Tamagotchi as basic, since they only consist of pressing
a few buttons, and point out that strong emotional relationships can still form between
these virtual pets and their owners:

“There have been cases of people treated for depression following the death
of their creature, highlighting how a simple software algorithm based on four
simple rules can truly engage people to form a relationship when correctly
packaged.“ [6]

This claim is further supported by Bloch & Lemish [7], who claim that the Tamagotchi
community has constructed the notion of a “virtual temple”, where the players can go
and leave offerings, or prayers, to accommodate those whose Tamagotchis have virtually
left them, and that several support-chat groups exist for both owners of live and deceased
virtual pets. This might seem absurd to some degree, and indeed there are skeptics that
argue that real love can never be obtained from such a bond [42]. Nevertheless, it seems
clear that there are strong emotional feelings involved in these human-virtual creature
relationships.

3.4 Virtual Creature Graphical Design

One aspect of human beings that makes work more difficult for designers, is that they
are really good at distinguishing what looks real and what does not. Furthermore, a
phenomenon which makes this work even more complicated, is a concept known as the
uncanny valley [43]. What this means is that, if something is very close to realistic,
but only slightly diverge from what is to be expected, human beings often view the
thing as appalling [33]. Some reasons why the uncanny valley problem appears may be
that the character, while seeming alive, reminds them of death. Another reason may be
the portraying of humanoid robots in popular culture, that are often viewed as having
superhuman strength and evil intentions. Finally, it may be that they lack emotions and
empathy, which reminds them of the fact that they are not alive, even though it looks
like they are. Koschate et al. [44] further proposes that a solution to the problem of the
uncanny valley effect is to increase the level of emotion expressed by, in this case, a robot.
They also claim that empathy is a core element of humanness and therefore is important
to simulate if one wants to create a character and make it as human-like as possible. It
is to be noted, however, that the study was made using humanoid robots rather than
animals, so a good question might be, does this apply to animals as well?

15



Figure 1: Mori’s uncanny valley graph [1]

In a study, looking into the uncanny valley problem in virtual animals, Schwind et al. [45]
shows that the uncanny valley problem is ever present in zoomorphic characters as well.
However, they argue that the curve might look different. They found that photorealistic
rendering gave a high level of familiarity, but also that toon shading (or cel shading) gave
good results in familiarity, aesthetics and realism. They conclude that the cel-shaded cat,
which was one of many cats rendered in different ways, is within the safe zone of Mori’s
uncanny-valley graph (Figure 1). To support familiarity, they also conclude that:

“For game developers, this means that an animal character should not possess
human expressions, emotions, or speech if a realistic animal character should
be fully accepted.” [45]

This means that the human expectation of how a cat should behave needs to be satis-
fied, and that adding human attributes to the cat in order to achieve a higher level of
anthropomorphism only further increases the uncanny valley effect, since it is not what
humans would expect of a cat. For example, adding facial expressions to an otherwise
ordinary cat, is therefore bad practice, which stands in contrast to the research of Choi et
al. [30], who claim that facial expressions in virtual characters are important in evoking
emotional engagement in the user. This illustrates that there are indeed differences in
how to effectively design a virtual animal or virtual humanoid character.

Schwind et al. [45] argue that there are two preferred ways of moving from the val-
ley. The first way is to increase the level of graphical quality to a photo-realistic level,
whilst maintaining the correct body ratios and matching the human’s expectations of
the animals (in behaviour and sound for example). The other method is to stylize or

16



abstractify the character to a level where you end up on the other side of the valley. Yet
another note here is that, in the study, only cats were studied. However, it is argued that
it is reasonable to assume that the same rules apply to other animal species.

Avoiding the uncanny valley is important in order to not create virtual characters that
are directly visually off-putting to an observer. While important, it is only one topic of
the theories on why characters are liked or disliked. Coulson et al. [33] claims that there
are three critical factors involved when a player decides on if a character is likable or not:
Physical attractiveness, Social attractiveness and Task attractiveness. The first property
is simple, and refers to the character’s visual appeal. Social attractiveness revolves around
the personality of the character, and how it acts socially. Task attractiveness is arguably
the most interesting one, since it has not been touched upon previously in this theory
section. Task attractiveness measures how useful the character is in a situation, or in the
game/setting as a whole. If a player finds a character useful, he or she is more inclined
to like that character. When considering the personality of the character, the authors
describes it as typically conceptualized in terms of five big personality traits, also known
as the five factor model [46]. The traits of this model are: Openness, conscientiousness,
extraversion, agreeableness and neuroticism. This theory does not apply specifically to
virtual characters, but is used in describing human personality in general. However, it
will not be further elaborated on here.

Choi et al. [30] points out the importance of emotional expressions, on the virtual charac-
ter’s part, when trying to evoke emotional engagement from the player. While relevant,
this paper only talks about the connection in a humanoid-user relation and how the char-
acter’s humanoid facial expressions influence the user. It is assumed that this could be
applied to animals as well, but probably not to the same extent. Facial expressions in
animals may be very hard to actually read as a human, but to animals, body language is
more vital than to human beings. This means that, if the body language of the animals
is recreated accurately when expressing the desired feeling, results resembling those of
the authors’ research might be acquired [30].

Emphasis should not be put solely on understanding the virtual characters, however.
Coulson et al. [33] stresses the importance of understanding the player observing the
character, as well as the character being observed, and that through understanding, the
mapping between real and virtual people will severely assist in developing more effective,
believable and entertaining virtual characters. The authors conclude that:

“..,virtual characters evoke strong feelings, and with increasing realism of
both graphical and psychological characteristics, designers and researchers
need to focus not just on virtual characters’ appearance, but on their utility,
and how both of these are filtered through the lens of player motivation and
personality.” [33]

This further illustrates the importance of understanding what a player likes to see in a
character, in order to make it likeable, or dislikeable, depending on the context.

17



3.5 Researching Emotional Engagement

One way to investigate human–virtual-entity interaction is described by Serholt & Baren-
dregt [47]. They studied robots tutoring children, and the child-robot interaction that
arose in such a setting. The study shows that children tend to reply to the robot as
if it was a human being, for example, by using gestures, nods and facial expressions.
The authors reached these conclusions by observing children interacting with a virtual
tablespace and the tutoring robot, while recording the actions taken by the children. The
observed behavioral cues of the children, such as gestures and where their gaze were, were
then sorted and categorized. The study was carried out over a period of three and a half
months, with most sessions ranging between 10-15 minutes, but there were also longer
sessions of around 40 minutes. It is worth noting that Serholt and Barendregt’s study
was focused on children aged 10-13, which is a considerably different target group than
the one intended for Augotchi. For Augotchi, the testers are planned to be at least young
adults, and the different ages might vary. Another major difference is that a robot is a
different kind of entity than a digital creature.

Direct observation and analysis of behaviour is one way to find out how effective an
artifact is in impacting people emotionally. Riva et al. [28], when measuring the emo-
tional response of people experiencing different VR environments, used two completely
different methods in order to draw their conclusions. The first being questionnaires and
the second being questions during observation of the participants. In their case, they mea-
sured differences in mood and immersion by asking the participants to fill out numbered
scales. For example, the participant was asked to answer, on a scale of 1-7, to what ex-
tent the virtual environment was experienced as reality for the participant. Even though
they could reach some conclusions, the authors point out that the sample space was quite
small and that the usage of only self report questionnaires might be a liability.

18



4 Methodology

The Augotchi project was divided into a number of phases. The methodology changed
from phase to phase, and the project consisted of four main phases:

• Concept development

• Minimum Viable Product (MVP) development

• Iterative process

• Project evaluation

Each of these phases consisted of various amounts of software implementation and design
work. The concept development was pure design work, and consisted of various methods
for refining the concept of Augotchi and determining what the application was supposed
to feature and how it would look. MVP development was purely technical, and consisted
of programming the first version of the application, which embodied the concept created
in the concept phase. The iterative process phase was a mix between the two, where
design work and technical implementation were performed in parallel, in order to refine
and expand the application and the concept. Finally, a brief project evaluation took
place, which was the analytic phase, where data gathered during interviews and co-
design workshops, were analyzed and discussed. In this chapter, the methods that were
planned to be used for the different phases are described.

4.1 Concept Ideation

4.1.1 Formal Game Analysis

The formal analysis of game play, which results were presented in 2.4, was based on the
method described by Lankoski & Björk [48]. Such an analysis is conducted by identifying
the objects, player actions and goals of a game, in order to identify its game play patterns
and then analyze what the roles of the identified patterns are. Thoroughly describing the
primitives (Objects, actions and goals) is key to discovering the implications that they
have on the game as a whole. It is also important that the researchers have extensive
knowledge of the game being analyzed, which requires the game to be played many times,
or in the case of the games studied for this thesis, a substantial amount of time. The
formal game analysis was expected to result in a deeper understanding of the two games,
and specifically to make it possible to draw conclusions regarding Niantic’s philosophy
on the design of AR games featuring GPS map integration.

19



4.1.2 Quick online surveys

Online surveys is a method designed to result in data about a selected topic, and depend-
ing on the needs of a project, the topic can be anything. However, ”Online surveys” is
a broad term. ”Internet or Intranet surveys” is one category of online surveys that are
short and quick, as described by Sue and Ritter [49].

4.1.3 Brainstorming

One of the most important methods when conducting concept creation is brainstorming
[50]. However, brainstorming is a broad term which includes a lot of different methods.
One thing that is important, no matter what brainstorming method is being utilized, is
to give the participants a question to answer. Brainstorming should always be a task
where the participants are looking for an answer to a question, as it is severely harder
to come up with good ideas if no question is present. These questions are commonly
referred to as ”focus questions”, and they are were defined prior to each workshop, in
order to gain the desired results. During the project, several forms of brainstorming [51]
can be used:

• Traditional brainstorming - This is the general brainstorming method. It is
designed to have a lot of interaction and encourage a lot of responses. Traditional
brainstorming is done by asking the participants a focus question and giving them
some time to answer out loud, while recording the answers.

• Stickies - This method utilizes post-it notes, on which the users write their answers
to the focus question on, before putting the responses up on display for all to see.
The upsides of this compared to traditional brainstorming is that it feels safer for
the participant to not have to speak out loud in front of an often unfamiliar group.
Additionally, the stickies method helps with neutralizing a dominant member of
the group that could otherwise give a disproportionate number of answers. The
drawback is quite small, being only that it slows down the brainstorming somewhat
since it takes longer to write than it does to speak.

• Grids - This method, as the name suggests, starts with the host drawing a grid on a
board. The grid may be 3x3, 4x4 or larger. The host then asks a focus question and
the participants get some time to answer it. When they are done with the answers,
the answer is written both in the header of column x and row y, meaning that the
first person’s answers are placed in column 1, row 1, and so forth. The task for
the participants is then to cross the ideas in the rows and the columns, producing
even more ideas. The positive things with this method is that it produces a lot of
answers from only a few initial ones, and it keeps the group motivated to continue
filling in the grid until all cells are occupied. However, the highly structured nature
of this method can be intimidating to some participants, and additionally, if the
focus questions are poorly designed, it might result in crazy ideas that might not

20



be so useful.

• Paper swap - In this method, the users take turns and collaboratively write several
series of answers. The host reveals the focus question, and all participants write
an answer on a sheet of paper. The paper is then sent one step to the left in a
ring, so that all participants have swapped papers. Then the focus question is told
again and the participants have to answer it a second time, while building on what
is already written on the paper. This process is repeated until all participants have
given input on all papers, or after a fixed number of steps. The positive aspect of
this method is that it is easier for the participant to come up with more ideas if
they have a red line to follow. The drawbacks are that you miss the spontaneous
energy that comes from traditional or full group brainstorming.

In addition to the methods mentioned above, many more methods of brainstorming were
described by Miller, in his book on brainstorming [51].

Before conducting any of the above mentioned brainstorming methods, it is really im-
portant to go through the basic rules of brainstorming with the participants, and during
the process it is important to make sure none of the participants violate these rules. The
rules are specified by Miller:

1. Focus on quantity, not quality - During the brainstorming round, it is important
to gather a lot of ideas. To improve the quality of those ideas is supposed to be
done in a latter stage.

2. Withhold evaluation - It is important not to critique an idea during the brain-
storming round, since this might lead to people not wanting to share bad ideas
(which is counterproductive, since quantity is key).

3. Encourage wild, outlandish ideas - Simply put, it is easier to tone down a crazy
idea, rather than making dull ideas interesting.

4. Combine or build on ideas from others - An easy way to get more ideas is to
combine two or more ideas into new ideas.

The expected results from all brainstorming methods are ideas. How many ideas depend
heavily on the number of participants, their mood, the time given to them and the
methods used.

4.1.4 Participatory Design and Brainstorming

When developing new technology it is important to adapt the technology to the user, not
to have the user adapt to the technology. One prime way to achieve this is by involving
the potential users early in the process of designing the technology. This is an aspect

21



of participatory design [52] and revolves around having the thought user group assist in
coming up with the technology. This is most often done in the ideation and conceptu-
alization stages of production, especially if the majority of the developers working on a
technology is not the intended end user. Exploratory Design Games [53] is one interesting
approach to participatory design.

4.2 Concept Evaluation

Ideas that are generated have to be filtered, ranked or improved in order to finally pick
what ideas to proceed with. Several methods exist for evaluating the ideas that are found
during the ideation phase of a project:

1. Investing in ideas - Taken from Miller’s book on brainstorming, this method
revolves around having the participants spend makeshift currency on the ideas
they think are the best. The idea with the most currency invested into it is then
deemed the best idea. The upsides with this method is that it is intuitive, and that
participants can individually rank the ideas while still producing a collaborative
end result. The drawback is that it might seem a bit childish to some participants.

2. Pros and cons - This method, also taken from Miller’s book, is designed for
participants to rank the ideas based on pros and cons separately. The goal is to
rank the cards from strongest to weakest while not revealing what specific ideas
the pros and cons are connected to. When done, the cards are flipped over and
the results are revealed. The method helps participants see the benefits and risks
of every option, but it has some drawbacks, one being that the participants may
remember what card they liked and vouch for that regardless.

3. Cost vs. benefit matrix - This method was taken from a Power-point presen-
tation by Health Quality Ontario [54]. The benefit cost matrix aims to rank every
option against each other on two axis - namely cost and benefit. After all options
are placed into the matrix, the ones with the lowest cost and highest benefit are
selected. In some cases, the cost is most important, when time is short for example.
Sometimes however, cost matters little, and then it is appropriate to select ideas
with higher cost, if they have higher benefits. It changes on a case by case basis
and is something that needs to be discussed afterwards. A positive aspect of this
method is that it results in a graphical overview of how the ideas relate to one
another. The cons with this method is that it might be hard to place ideas into the
matrix at the correct points, or that it might be hard to pick an idea in the end
anyway.

22



4.3 Concept Creation

One useful method for creating a concept is to define what is called an MVP. One way
to do so is to filter out a set of critical features that need to be implemented in order to
realize the fundamentals of the concept. Based on the result of the previously mentioned
value-cost matrix the following categories can be derived. Each category loosely repre-
sents one area of the matrix. This idea of creating feature categories from the matrix is
our own invention.

Critical features - Without these features, the game will not be playable. They are
critical to the application and constitute the requirements for the MVP. These features
will be implemented first, and the category will not be relevant after the MVP has been
developed. This category does not only contain game features, since it also consists of
tasks for programming the technical foundation of the game.

Important features (Highest benefit, high-low cost) - These features are important
for the application in order to make it enjoyable, however, they are not crucial, which
means that the application will still work even if they are not present. Important features
have the highest priority when deciding what to implement in each iteration.

Desired features (High-medium benefit, medium-low cost) - These are features
that are considered to have high value, but that do not belong in the important features
category. High priority, additional content, will be of this feature type, and may be con-
sidered as the “cherry on top”. The application will be fully playable and, hopefully,
enjoyable even without these features.

Additional features (Low benefit, high-medium cost) - Features that will only
be implemented if there is extra time. While not considered bad ideas, these features
have been determined to have little to no impact on the playability or enjoyability of the
application.

Minor features (Any benefit, low cost) - Features in this category are considered
to be quick and relatively easy to implement. The overall application will not be heavily
impacted if some of these are implemented. However, they are considered to be clear
improvements to the game. Furthermore, a number of smaller features might lead to a
greater user experience overall.

Discarded features (Lowest benefit or highest cost) - These are rejected features.
They have been determined to not be implemented, either due to them being considered
bad ideas or because they have a very low benefit-cost ratio.

23



4.4 MVP Development and Vertical Slicing

To define an MVP is a good way to establish the first goal of the project. In this thesis
an MVP is defined as:

A product, featuring the least amount of features, while still being runnable
and qualify as an AR game.

Product-wise and implementation-wise, this MVP is the worst acceptable result of the
thesis. A vertical slice, on the other hand, is a more refined version of a product which
aims to accurately represent an eventual final product. A vertical slice in this thesis is
defined as:

A refined product, which contains a set of features which makes it qualify as
releaseable, and which accurately represents the potential final application.

Vertical slicing is a concept derived from agile principles, and is described by Ratner
& Harvey [55]. While not a method per say, defining a vertical slice is a solid way of
establishing a long-term goal which transcends the short-term goals of each development
iteration, which aids in keeping the tasks of each iteration relevant to the goals of a
project.

4.5 The Agile Development Process

One of the most common software production approached is the utilization of Agile
development methodology. Agile development is a group of applied software development
methods that originates from the Lean manufacturing methods [56]. There are a lot of
different frameworks for Agile development, such as Extreme programming, Kanban and
SCRUM. Arguably, the simplest of the agile development methods is SCRUM, since
SCRUM is designed with small teams in mind. It is to be noted, however, that SCRUM
contains a plethora of methods, which means that not all of them have to be utilized
[57].

In SCRUM, there are three different roles that are assigned to various team members,
each with its own responsibilities. There is the product owner, who is the person owning
the project or product that the team or company is trying to make. The next role is
the Scrum Master. The responsibilities of the Scrum Master is to guide and coach the
team to achieve the desired result in the desired time frame. This person is also the one
managing the daily meetings, and he or she is also the one who mainly communicates
with the product owner. The final role is the Team member. The team member is
simply part of the team that is trying to finish the project under the SCRUM Master’s
guidance.

A central concept of SCRUM is fix length iterations that repeat [57]. Sprints are usually

24



between one and four weeks long, but shorter sprints are preferable. Another important
concept of SCRUM is to have a sprint meeting once per sprint (either in the end of a
sprint, or at the start of a sprint). During the sprint meeting, the SCRUM Master reviews
what has been done the previous sprint and the desired results of the upcoming sprint
are discussed. The third important aspect is daily meetings. This is to prepare the team
for the day and make sure that everyone are always sure of what they are supposed to
work on. The daily meeting is traditionally also held by the SCRUM Master. The final
core concept of SCRUM is the SCRUM Board. A SCRUM board consists of cards that
represent different tasks. Depending on their position in the production chain, they are
placed in certain areas on the SCRUM board. This board is in place mainly to grant the
team, as well as the SCRUM Master and product owner a clear visual representation of
how far the project has progressed, and if anyone is stuck and needs help.

The agile cycle for SCRUM is composed of four steps, and repeats once every sprint:

• Test the latest changes made in the previous sprint

• Evaluate changes made in the previous sprint and the application as a whole

• Establish new requirements based on the prior evaluation

• Implement or modify features as suggested by the new requirements

4.5.1 Result Evaluation

Conducting interviews Cote & Raz give some advice on how to recruit participants
for interviews related to game research [48]. They argue that game research is relatively
easy to conduct, since most games have a well-defined community organized on specific
forums on the internet, which makes the target audience easy to find. However, they
also warn that most people who are active on forums do not necessarily represent the
average gamer, since they are more likely part of a minority of highly dedicated fans of
a specific game. Once interview participants are found, there are a number of ways in
which interviews may be conducted; ranging from structured questionnaires to more open
discussions. What interview method to use depends on the needs of the project, which
means that it is important to establish what information is sought out prior to planning
the interview sessions.

TLX forms Nasa TLX forms [58] is a useful tool when evaluating a user’s perception
of a certain feature. Unlike interviews, Nasa TLX forms result in quantitative data
which measure a tester’s experience while performing a given task. These forms could
potentially be used in order to compare different features of a game with one another, or
even to similar features in other games.

The fly-on-the-wall principles One way of conducting evaluation during the iterative

25



process is to meet with the testers and observe them as they play the game. This is
known as the fly-on-the-wall principle [59]. If the AR setting causes issues in conducting
such testing, the players could be observed in a manner that resembles the approach
of Serholt & Barendregt [47], in which specific interactions between the virtual entity
and the players is observed and recorded. Emotional expressions and body language are
examples of events to monitor which could be used in order to measure how emotionally
engaged the players are while playing the game.

4.5.2 Forming New Requirements

During the sprint meeting, either at the end of a sprint or at the beginning of one, the
new requirements or tasks of the next sprint are set. There are a number of approaches to
decide on what features should be brought into the next sprint. For example, instead of
selecting a number of features for each sprint, as one would normally do, one overarching
concept can be selected instead. Exactly how the system is to be implemented is then
decided throughout the sprint, in a manner that resembles rapid prototyping [60]. In any
case, it is important to evaluate the product beforehand and take the evaluation findings
into account when planning the next sprint. Evaluation data in this case does not need to
be directly gained from external testers, but could also consist of bug reports formulated
by the developers themselves.

When estimating how much time a selected feature requires, several different methods
can be utilized. One of them is planning poker [61], with which the team ranks the tasks
in terms of time cost, by using cards that have values in a Fibonacci series of numbers.
Each task is brought up, and the team members then play a cost card, estimating how
much time they estimate the task requires. If someone ranks the card abnormally high
or abnormally low, he or she has to motivate their decision. After discussing the matter,
the group either agrees on how long the task takes, or the mean of the selected numbers
is selected.

4.5.3 Implementation

After new requirements have been established, it is time to implement them. This marks
the beginning of a new iteration cycle.

During the sprints a SCRUM board is used. The exact structure of a SCRUM board,
however, is free to vary, and it can be either physical or digital. The physical SCRUM
board is probably preferable if the whole team is located in the same physical place.
Otherwise, an electronic scrum board may be used. There are many sites online that
provide this type of service. One such is the website Trello [62], which provides a free,
easy to use online SCRUM board.

26



4.5.4 Iteration Stop

When the planned time for the iterative phase comes to an end, the loop is halted and
development is considered done. The reason for choosing an approximate stop date for the
iteration phase is that the product is improved with each cycle, and thus more iterations
equal a better product. The goal is therefore to conclude as many sprints as possible in
a given time frame, rather than setting the total number of finished sprints.

4.6 Interview Data Analysis

In an article by Cote & Raz [48], it is described how you would analyze and interpret data
gained from interviews. Before the analysis of the data can commence, however, the data
has to be transcribed. A question to be answered is at what level to put the transcribing,
but it is relatively straightforward, since the concepts of the game are well known. After
the data is transcribed, Cote & Raz continue to explain the analysis process. Mainly the
task consists of breaking down the data into patterns and putting those patterns into
a theoretical framework in order to draw conclusions. When designing the interviews,
considerations will be made regarding how to best extract data that is related to the
goals of the interview, namely the entertainment value of the game.

4.7 Ethical Considerations

When developing AR games featuring GPS map integration, a long line of concerns
arise. Some of them are connected directly to the users and their well being, other
concerns revolve around their personal information and integrity, and lastly, there could
be concerns regarding such technology’s impact on society as a whole.

One concern is that virtual companions, in the not-to-distant future, could gain the
ability gather highly personal data about users, in order to respond to said users more
effectively, and this type of data is considered to be highly sensitive [63]. As long as the
project does not conflict with GDPR, it will not give rise to any issues, at least from a
legal perspective. In any case, the pet featured in Augotchi was never planned to have
an advanced AI that could, for example, analyze, encode and store data which represents
the feelings of its owner.

AR applications could also have questionable impacts on society. Miguel Sicart expresses
one such concern, namely the implications that AR applications have in regards to com-
mercial rules and legislation in public spaces [64]. He argues:

“GO is a connected network of corporate interests, from the Pokémon Com-
pany to google and fast-food companies that have discovered that in AR, pub-
lic space regulations are not necessarily applicable. Public space is threatened

27



by an interface that is proprietary, and by the lack of regulations and codes
of practice.” [64]

This is certainly something that will have to be kept in mind when developing any kind
of AR application. But as of now, no such codes of practise exist.

One of the most important ethical issues of the Augotchi project is the immediate danger
connected with playing AR games like Pokémon GO. This is illustrated by the Pokémon
GO application itself, which has warnings that are put in place to remind the player of
these dangers. When starting the application, while the game is loading, the following
text appears:

“Remember to be alert at all times. Stay aware of your surroundings.”

Furthermore, every time the game is loaded and opened, a popup is displayed with an
“OK”-button and one randomly selected warning, see the central image of figure 2 below.
Warnings include texts such as:

“Do not trespass while playing Pokémon GO.”

There are also conditional warnings implemented in Pokémon GO. For example, when
the player is moving too fast, obtrusive warnings pop up that remind the player of not
driving and playing the game simultaneously.

Figure 2: Player warnings in Pokémon GO

These warnings illustrate the concerns that the developers have regarding how people
play their game, all of which are related to the players physically moving around in

28



public areas. Hence, these are ethical issues specifically relevant to AR games that do
not arise when playing games, sitting down, in the comforts of one’s home.

For Augotchi specifically, it will be important to emphasize these dangers when bringing in
testers. When bringing in testers, these have either come from communities of people who
are accustomed to the dangers of AR, or they have been people that the team has met and
informed in person. Hence, obtrusive popups throughout the game will not be necessary,
as long as the participants have been properly informed, or have previous experiences
with similar games. If a feature of Augotchi differs from similar games in a way that
could make it more risky, this will have to be conveyed to the testing participants. An
example of such a difference would he how the game’s content is placed in the world.

29



5 Planning

In order to reach the goals of Augotchi, this project is divided into a number of phases.
These phases are the conceptualization phase, the MVP development phase, the iterative
development process, and finally, the evaluation phase.

5.1 Conceptualization

The primary goal of the conceptualization phase is to formulate a list of critical features
and/or requirements that will be carried forward to the MVP development phase. This
phase is also planned to explore the possibilities of AR games and virtual pets in general.
Hence, ideation is the core of this phase, which consists of various brainstorming activities
and idea evaluation techniques. A co-design workshop with external participants will be
conducted in order to produce a wider array of ideas.

5.2 MVP-development

During the MVP development, the first version of Augotchi is planned to be developed.
The features that were selected from the previous conceptualization phase will be ranked
according to the ranking system discussed in section 4.3, and then implemented. The
MVP-development phase of the project is essentially what is called the ”initial phase” of
SCRUM, which aims to provide a solid foundation for the future iterative process.

5.3 Iterative development

The iterative development phase will consist of a number of sub-phases, divided into
iterative cycles known as sprints. These phases are the planning, implementation, testing
and evaluation phases. During the later iterative phases, when testers have been recruited,
evaluation will be conducted in order to evaluate the changes made during each sprint.
The results of the evaluation are then used in order to plan and execute the next sprint.
The goal of each sprint is simply to improve the prototype and to gather data which may
lead to insights regarding the development of a game like Augotchi.

5.4 Evaluation

During the evaluation phase, the testers will be interviewed in order to evaluate the
prototype as a whole. Data gathered through evaluation sessions during the iterative
process, as well as the data from these interviews, will then be analyzed in order to draw

30



conclusions regarding the development of a game which aims to emotionally engage its
players through a virtual pet.

31



6 Process and Execution

6.1 Preparation and Research

The preparation phase of the project included a number of steps, including a literature
study, constructing a time plan, as well as exploration of available technology and possible
methodology. A formal analysis of game play was also conducted, which was presented
in section 2.4

6.1.1 Establishing Roles

Prior to the thesis work taking place, the skill sets and goals of us, the students, were
identified in order to plan a thesis that would be well executed in a manner which would
satisfy both parties. While we both had a background in computer science engineering,
our ambitions and interests varied. Kevin worked as a 3D environment artist for Far
North Entertainment, and Simon mainly as a game designer and developer.

This difference has been important when planning every phase of the project, and a major
part of the iteration planning consisted of balancing the division of labor, so that both
students had an approximately equal workload even though the tasks were of entirely
different natures. During the prototype development, this division has primarily been
between programming and graphic assets creation/design. These role differences have
also been utilized during theory research, to make sure that each topic was studied by
the person for whom it was most relevant. Whenever the workload turned out to be
uneven during an iteration, the freed up student worked on tasks belonging to a shared
pool of tasks, such as UI creation and planning.

6.1.2 Theory

With both students having little knowledge regarding emotional engagement and rela-
tionship forming between digital artifacts and people, a literature study was conducted
in order to gain a better understanding of said topics. The supervisor of the project
supplied us with some video material and articles that could be relevant, which was a
good starting point for the study. We split the material in two and studied the contents
respectively. The knowledge gained was then shared between the two of us in a small
literature seminar.

The main problem with most of the videos and articles were that they all discussed
contexts in which the digital artifacts were useful to someone in some sense. Most of the
studied videos and literature focused on how to improve certain aspects of human life,
such as elderly care, or the education of children. Augotchi was meant to be a game,

32



hence a product supposed to entertain, which made it hard to find other projects with
the same direct goal in mind. The literature which revolved around digital characters
in games and other entertainment media was helpful for the design of the game and
its character, but did not connect the digital character design to the deeper emotional
engagement that Augotchi would hopefully lead to. Similarly, when studying some of the
literature directly referencing the Tamagotchi, it was established that there were clearly
strong bonds forming between the digital pet and its owner, but nothing was established
about why this was the case.

The studied literature was divided into different areas of research and presented in the
theory chapter with these areas as section titles. Each section could vaguely be described
as answering a specific question:

• AR - Benefits and Problems - What are some general considerations when working
with AR applications?

• Emotional Engagement in Entertainment Media - Why do people care about phys-
ically unreachable characters?

• Relationships with Virtual Companions - What has already been studied regarding
relationships between people and virtual companions?

• Virtual Creature Graphical Design - How should a creature’s looks and behaviour
be designed in order to make it effective at creating such relationships with the
user/player?

The goal of the thesis, following the literature study, became to design a virtual pet
in a game that would adhere to the findings of the studied literature, and then test if
emotional engagement of the player could be achieved. This is the missing link that the
thesis first aimed to answer. However, due to reasons described later in sections 6.7.4,
6.7.6 and 6.7.8, the focus question was changed to revolve around the design of a virtual
companion in the specific AR setting.

6.2 Utilized Technology

Before developing the software prototype, it was deemed important to explore what
features and tools could be gained from already existing solutions on the market. By
using Google, forums and the Unity asset store [65], tools and solutions to most technical
challenges would be found, evaluated and tested, in that order. The process of finding a
solution to a specific problem was divided into 5 steps:

1. Establish what needs to be solved.

2. Find solutions to the problem.

33



3. Evaluate the solutions by looking at descriptions, user reviews, price and examples
of where the solution has been used before.

4. Compare all solutions and rank them based on the needs of the project.

5. Test the highest ranking solution in Unity, declare whether it is viable or not. If it
is not viable, remove it from the list and repeat this step for the next item in the
list.

6.2.1 Platform

When deciding which platform to use for Augotchi, the main criterion was its potential
for AR functionality. The number of device types that were deemed suitable in this
regard was small, not necessarily because the excluded platforms could not do AR at
all, but because of the fact that the good options outmatched the less optimal ones.
A desktop PC, for example, is heavy, unportable and does not usually come with GPS
integration or even a camera built in. On the other side of the spectrum, there is the
Microsoft Hololens, which is a headset-like device specifically developed with AR in mind
[66]. There are various alternatives to the Hololens as well [67]. However, none of these
AR devices are, even remotely, in a price range that would fit the budget of the project.
A platform that is more available had to be found, and the smartphone was therefore the
de facto solution for Augotchi, a choice for which there has been little concern.

The smartphone, as an AR device, has a number of strengths, as described in section
2.2. In addition to the technological capabilities of the device itself, it also opened up
possibilities during user testing, since a majority of people in Sweden already own such a
device. The availability of the platform is also important from a commercial standpoint.
Furthermore, both of us have previous experience in developing for Android, and own An-
droid phones, both reasons for developing for Android specifically. The phones that were
used during the development were a Samsung Galaxy s5 [68] and a Samsung Galaxy s8+
[69]. These devices are three years apart in terms of technological advancement, which
is useful when developing an application that will, preferably, work on a wide variety of
devices.

6.2.2 Software Solutions

Before heading into the initial design phase of the project, it was deemed important to
have a good grasp of what is possible and not in technical terms. In order to not create
a concept that would be hard to realize within the time frame of the project, a variety of
available software solutions were researched, and a set of applications and APIs [70] were
selected as the foundation for Augotchi. This improved the odds of correctly predicting
the time that the prototype would take to develop, hence improving the workflow and
the chance of the project being a success as a whole. Testing the different components

34



individually prior to the prototype development also minimized the risk of delays due to
technical issues.

Game Engine
A game engine was required for the prototype to be finished within a reasonable time
frame. It provided that which is required in any game, such as rendering graphics, han-
dling input, playing sounds and managing all the objects within the game. The two game
engines that were considered for the development were Unity and Unreal Engine. Both
would be free to use throughout the project and they are, arguably, the most popular
public game engines of small and big game development companies alike [71].

The game engine that was chosen for developing the Augotchi prototype was Unity [72].
Our previous work experience was one of the reasons for choosing Unity over its contender,
Unreal Engine [73]. The development within the engine would mainly be performed by
an experienced Unity developer, which made Unity an obvious choice, since development
time is sparse and learning a new game engine takes time and effort. Furthermore, no
critical reasons were found that pointed to Unreal Engine being preferable to Unity, even
when experience is not taken into ac