uAdventure: The eAdventure reboot Combining the experience of commercial gaming tools and tailored educational tools

Abstract

Educational games (aka serious games, SG) are powerful educational contents. However, they are costly to develop, and once developed, SGs become dependent on software and hardware combinations that may become obsolete, such as Adobe Flash or Java Applets. Addressing these problems would allow a much greater use of SGs in education. The eAdventure authoring tool, developed by the e-UCM research group, addressed high development costs, and resulted in the creation of multiple SGs in collaboration with different institutions. However, eAdventure's Java Applets have become increasingly difficult to run due to platform obsolescence. To maintain the benefits of the eAdventure platform and user base, we have created new platform called uAdventure: an SG editor built on top of the game engine Unity that allows for the creation of educational adventure games without requiring programming. Since Unity is supported on a majority of platforms (including mobile). By developing SGs with uAdventure, the games become future-proof, as they can be updated and retargeted for new platforms as required. In this sense, uAdventure improves the lifecycle of SGs by reducing both authoring and maintenance costs.

Full text

uAdventure: The eAdventure reboot
Combining the experience of commercial gaming tools and tailored educational tools
Ivan J. Perez-Colado, Victor M. Perez-Colado, Ivan Martínez-Ortiz, Manuel Freire-Moran,
Baltasar Fernández-Manjón
Dept. of Software Engineering and Artificial Intelligence, Facultad de Informática,
Universidad Complutense de Madrid Madrid, Spain
{ivanjper, victormp}@ucm.es, {imartinez, manuel.freire, balta}@fdi.ucm.es
Abstract—Educational games (aka serious games, SG) are
powerful educational contents. However, they are costly to
develop, and once developed, SGs become dependent on software
and hardware combinations that may become obsolete, such as
Adobe Flash or Java Applets. Addressing these problems would
allow a much greater use of SGs in education. The eAdventure
authoring tool, developed by the e-UCM research group,
addressed high development costs, and resulted in the creation of
multiple SGs in collaboration with different institutions. However,
eAdventure’s Java Applets have become increasingly difficult to
run due to platform obsolescence. To maintain the benefits of the
eAdventure platform and user base, we have created new platform
called uAdventure: an SG editor built on top of the game engine
Unity that allows for the creation of educational advent ure games
without requiring programming. Since Unity is supported on a
majority of platforms (including mobile). By developing SGs with
uAdventure, the games become future-proof, as they can be
updated and retargeted for new platforms as required. In this
sense, uAdventure improves the lifecycle of SGs by reducing both
authoring and maintenance costs.
Keywords—serious games; Unity 3D; learning analytics;
geolocalization
I. INTRODUCTION
Educational simulations, gamification approaches and
educational games in general (hereinafter referred to as serious
games, SGs) have been attracting lot of attention by education
and training communities due to their intrinsic motivating and
engaging characteristics [1]. SGs have been applied in different
educational settings and application domains, most notably in
business management, health, and military. SGs are used to
facilitate training of abilities tactics that are specifically suited to
be acquired in role-playing educational settings, such as conflict
management or business negotiation skills. Moreover, SGs can
be used to acquire or refresh a required ability that is infrequent
or rare in real life (medicine) or poses security or cost difficulties
(military). In addition to education and training purposes, SGs
are also used in other domains such as public health or
environmental protection to improve awareness or to affect
behaviors.
As a result of the increasing use and acceptance of SGs, a
niche market is being created around their development and
1https://en.wikipedia.org/wiki/Zoo_Tycoon
maintenance. This market is expected to grow over 16%
annually between 2015 and 2020 [2]. For instance, CodinGame,
an online SG to learn programming, has around half a million
registered users, two thousand of which are online at any given
instant[3].
The process of developing and procuring SGs can be
classified using the following categories: i) Commercial Off-the-
Shelf (COTS), ii) Pre-existing Serious Games and iii) Tailored
Serious Games. This classification is based on the alignment of
the learning goals of the educators against the actual learning
goals (if any) that are part of the SG’s actual design. Generally,
the greater the degree of personalization, the better the game will
suit educators' needs. As we get closer to a fully-personalized
game, the development process of the game will become more
complex and costly (per user).
In some cases, an already available commercial game would
suit the educator’s needs. These COTS games, when applied in
an adequate educational context, can have a relevant impact. For
example, Zoo Tycoon 1is a game focused around building and
running successful zoo scenarios where players learn concepts
related to economics or business management; and has been
used as a SG to help teach these concepts. The economical
advantage of this approach is clear: the game does not need to
be developed, only licensed. However, the effective application
of this approach also requires an additional effort to design the
pedagogical scenarios and a degree of teacher support while the
game is deployed as an SG.
Pre-existing Serious Games are commercial SGs that are
usually supported by training and consultancy companies that
may offer a customization layer. This customization layer, is
usually offered as an a la carte service, where the organization/
instructor can choose a set of learning goals and the company
provides a set of SGs or a single SG that cover the required
competencies or learning goals. This approach can be
considered as a Games as a Service, where the learning
organization/instructor does not have to either develop or
maintain the game, for a fee. Other advantages include
streamlined integration with the learning process, or assessment
tools for monitoring and behavior analysis. An example of this
genre is Classcraft2, where the classroom is transformed into a
2https://www.classcraft.com
2017 IEEE Global En
g
ineerin
g
Education Conference (EDUCON)
978-1-5090-5466-4/17/$31.00 ©2017 IEEE 25-28 April 2017, Athens, Greece
Page 1754

role-playing game that improves students' collaboration and
teamwork.
Finally tailored serious games, despite having the greatest
potential effectiveness, also tend to require large,
multidisciplinary development teams and, therefore, have high
production costs. However, for simpler SGs, it is feasible to
provide educators with an easy to use platform, with limited
features, that can allow them to create their own SGs; or at least
to modify or localize it to their needs.
The eAdventure (eA) platform [4] provides an authoring tool
that allows for SG development without requiring advanced
computer skills (i.e. programming). In addition, the eA platform
allows for the reuse and repurposing of preexisting eA games.
This allows educators to customize the games for a specific
educational setting. For example, changing the students'
evaluation included in the game or simply changing graphic
assets of the game elements to customize their aspect to the
specific learning context.
However, the eA platform was launched in 2007, and
although it has received several updates, its underlying
technology has run into technical limitations. In particular, eA’s
targeted the Java Plugin, then present in most browsers.
However, as described in JEP 2893, the Java Plugin is to be
deprecated in newer versions of Java; and, as of this writing,
most browsers already ship with this component either absent or
disabled. To address these concerns, and to streamline the
development of new educational features, we have designed
uAdventure (uA), an evolution of eA.
The rest of the paper is structured as follows; section II
provides insight into eA analyzing the limitations that required
an evolution from both technical and educational perspectives.
Section III and IV describe uA and the new features and
educational opportunities that it offers. Finally, section V
provides some conclusions and future lines of work.
II. FROM EADVENTURE TO UADVENTURE
A. eAdventure
Developing games, and especially SGs, requires addressing
and balancing both educational and technological aspects. One
of the first aspects to consider is the platform or technology used
to develop the game. For example, using a professional game
engine such as Unity or Unreal Engine requires a game
development team with good programming background; and
this will prevent most teachers from participating in the SG’s
core development team.
Another aspect the adequate balance between engagement
and education [5]. On the one hand, if the game is too focused
on the educational facets it may not be engaging enough for the
student. On the other hand, if the game is too narrowly focused
on the fun facet, the student may not learn enough, yielding only
limited educational value.
3 February 2016’s JEP 289: http://openjdk.java.net/jeps/289
4http://www.adventuregamestudio.co.uk/
Within games there are several genres, and not all of them
are equal from an educational perspective. Dickey [6] and
Amory [7] have identified point-and-click adventure games as
one of the most suitable genres for SG, due to the presence of
elements such as a slow pace, reflection, study of the
environment, and problem-solving [8]. There are specific
commercial editors for creating these kind of games, such as
Adventure Game Studio4, Adventure Maker5or 3D Adventure
Studio6 [9]. However, these platforms do not usually include
support for SG educational elements (e.g. evaluation).
eA is a platform for simplifying the development of 2D
point-and-click adventure serious games [4]. In eA, educators
can actively participate in the development of the SG side-by-
side with the game developers or even develop the SG
themselves without requiring programming skills. This way,
educators are at the core of the development process taking care
of the educational aspects that fits for their own needs.
eA has four characteristics that make it especially suited for
the creation of educational SGs: i) game development can be
achieved with small budgets, ii) maximizes the Return on
Investment (ROI), iii) customizes the learning experience and
iv) facilitates both distribution and evaluation. One of the most
expensive aspects of game development is the creation of
graphic assets. With eA, it is possible to use pictures taken with
any digital camera to create the graphic assets. eA games are
editable, meaning that existing games can always be opened
with the editor to either customize them for specific learning
scenarios, or to update them for new requirements.
In addition, eA games can be packaged as an e-learning
standard content package that includes the game as a Java applet
making possible to distribute it using a Learning Management
System (LMS). Furthermore, if the LMS it is compatible with
the ADL SCORM [10] e-learning specification the eA games
can send back assessment results to the LMS.
B. Why does eAdventure need a refurbishment?
eA design started in 2007, and the first version was delivered
in early 2008 as an open source project. The project has been
maintained since then by different contributors, mainly by
researchers that have been actively using the platform. These
contributors have often added features requested by educators
participating in the eA community. Nevertheless, there were
some architectural and technological decisions that were correct
at that time, but are now hindering the maintenance and
evolution of the platform.
eA was built with Java technology to take advantage of the
“build once run anywhere” motto. Java technology powered
both the eA editor (used in game development) and the games
themselves, which could be deployed in different platforms
without changes. This way it is possible to alleviate, or
completely avoid, compatibility issues between the computer
used by the educator (or the development team) to create the SG
and the usual heterogeneous environment that can be found at
school labs (including security restrictions). In addition, this
5http://www.adventuremaker.com/
6http://3das.noeska.com/
2017 IEEE Global En
g
ineerin
g
Education Conference (EDUCON)
978-1-5090-5466-4/17/$31.00 ©2017 IEEE 25-28 April 2017, Athens, Greece
Page 1755

multiplatform nature and the ability of eA games to be
customized fostered the collaboration between educators in a
community of practice by sharing SGs.
Although Java’s “build once run anywhere” motto may still
hold for the desktop platform, desktop computers, once the main
platform in school, have since disappeared from multiple
institutions, displaced by tablets. Desktop Java is generally
unavailable on tablets and smartphones: although Java as a
language is still relevant (due to the Android platform),
development in Java is not longer providing effective
multiplatform support. Moreover, mobile devices offer new
ways to interact with the device (touch, accelerometer,
geolocation, etc.) that open new opportunities to create
ubiquitous educational experiences that should be used in SGs.
eA SG deployment relies on Java applets (i.e. Java programs
that run inside the browser) to facilitate SG distribution and
integration into existing e-learning platforms. When eA was
created, e-learning content interoperability was mainly focused
on distribution of web content. The use of Java applets in eA
facilitates the reuse of e-learning standards allowing SGs to be
packaged as an IMS-CP content [11]. This allows students to
access and play the game just like any other web content. The
eA platform could also take advantage of ADL’s SCORM e-
learning standard to communicate back with the LMS, mainly to
store SG progress and assessment information inside the hosting
Learning Management System (LMS).
However, due to the continuous security problems [12] in
Java Applets in particular and in browsers' plugins in general,
most of the browsers have already deprecated the Java applets
or are in the process of deprecating them [13].
The eA platform, besides being an open source point-and-click
editor, complements the usual game platform features with
educational features (e.g. assessment, e-learning standards
support) that make eA still unique and relevant (see Fig. 1).
Although only two of eA’s components have been specifically
7https://unity3d.com/unity/multiplatform
8http://eu.battle.net/hearthstone/en/
9http://www.pokemongo.com/
designed for educational purposes, considerable effort is
required to maintain the platform as a whole. This imbalance it
is dangerous for the maintenance and survival of an open source
research project with limited resources (developers). In
particular, keeping up with technological evolution makes it
harder to add and update the educational features that teachers
expect and demand.
Nonetheless, we consider that eA’s main ideas are still valid
and relevant to the educational community. Therefore, we
decided build a new SG platform, which we describe in the
following section.
III. UADVENTURE
uA development aims to achieve the following goals: i) to
address the technical issues of the eA platform, ii) to face the
project management/survival issues, iii) to provide a solid base
platform to build new educational features and iv) maintain
compatibility with eA (at least importing previous eA games).
uA is built on top of Unity mainly because it has become the
most popular game development platform [14]. One of the
reasons of this popularity is its multi-platform support7
including desktop (Windows, Mac OS and Linux), consoles
(PlayStation, Xbox, PSVita or Nintendo 3DS), web and mobile
(Android and iOS). Unity has attracted attention both from big
companies like Blizzard with “Hearthstone”8, Nintendo with
“PokemonGO” 9or Ubisoft with “Grow Up!”10, and is also one
of the top choices for so-called “indie” developers11. The “indi e”
label is generally used to refer to low-budget games, and in this
sense encompasses most SG developments. Compared to the
other choices, Unity stands out for its simplicity and
extensibility, rich documentation, and a very active community
of developers.
In addition, the choice of Unity was an answer to project
management issues. uA can take advantage of a preexisting
strong platform and community. Thus uA development time can
be focused on the development of those characteri stics that made
eA unique. Moreover, due to the use of a well-known platform,
finding contributors or hiring developers to create new features
should be greatly simplified.
Furthermore, the expansion of mobile technologies has
created new opportunities to apply them in education. As we
move into mobile technologies, new forms of “anytime and
anywhere” learning are possible. Learning can happen at any
moment with a range of devices: at home with the PC, at school
with a tablet, or just outside during the free time with the
smartphone.
uA have been created in a context where previously
developed eA games have serious deployment problems. This
new uA framework, built on top of Unity, allows eA projects to
be imported, allowing them to be tested and mo dified on the new
execution core. This extends, supports, and simplifies their
lifecycles. After opening a game in uA, and using the Unity
10 https://www.ubisoft.com/en-GB/game/grow-up/
11 https://unity3d.com/public-relations
Game Engine
Java Platform
eAdventure
editor
.eap
eAdventure
Packa
g
e
eUCM Maintained packages
Java Maintained packages
eAdventure
engine
eAdventure
game model
Educational
Features
Education-
oriented features
Fig. 1 eAdventure’s main component s
2017 IEEE Global En
g
ineerin
g
Education Conference (EDUCON)
978-1-5090-5466-4/17/$31.00 ©2017 IEEE 25-28 April 2017, Athens, Greece
Page 1756

compilation tools, it can be exported as a standalone game for a
range of different platforms and operating systems.
A. uAdventure Architecture
eA is composed of several layers (see Fig. 1), such as
editor, game player, representation core and project builder. In
tightly-coupled, complex software solutions developers need to
cope with all the complexity; and in this case, developers
wishing to extend eA needed to be familiar with all of these
layers. In contrast, uA is composed only of two main layers:
Interpreter and Editor. Most of the previous ad-hoc eA layers
have been replaced by Unity’s standard layers. Thus, uA will
be easier to maintain and evolve.
The uA module architecture is presented in Fig. 2. It shows
how the graphical editor (GUI), the interpreter and the emulator
are built on top of Unity. New features are highlighted in Fig. 2
with dashed lines: the xAPI tracker, extensions such as location
support, and a future Unity-like GUI.
Compared to the previous architecture, it may seem more
complex, since there are more modules. However, this is
misleading: even though there are more e-UCM maintained
modules, they are organized into only two layers: Editor and
Interpreter. The rest of the e-UCM maintained modules are
smaller and easy to maintain compared to the eAdventure’s
packages (Editor and Game Engine) shown in Fig. 1, whose
cores have been replaced withthe Extensible Editor Engine and
the Extensible Game Engine provided by Unity.
Therefore, the two main layers that compose uA are:
Interpreter and Editor.The Editor is focused on creating a
game specification that, in conjunction with the required game
resources, will make it playable. It uses the eA data model plus
some new additions, like new map scenes and new assessment
specification [15]. The Interpreter relays on Unity’s
representation coreand translatesall eA model components into
Unity game elements.
B. The uAdventure editor
The Editor allows users to modify the game specification,
manage and create new game resources. Some key aspects of
eA have been recreated in uA, such as the main graphical user
interface, including menus, windows and general appearance.
This way, it will be easier for the current eA users to migrate to
Fig. 3: Fade transition from players appearance view on uAdventure (left) to previous players appearance view on eAdventure (right). The ed itor is quite simil ar
in both systems.
Fig. 2: uAdventure modules architecture, including Unity provided modules.
2017 IEEE Global En
g
ineerin
g
Education Conference (EDUCON)
978-1-5090-5466-4/17/$31.00 ©2017 IEEE 25-28 April 2017, Athens, Greece
Page 1757

uA. However, uA will also provide a brand-new Unity-Unified
GUI. This new GUI will be based on the idea of tighter
integration with the Unity editor, and is inspired by the most
used Unity windows, such as the Scene preview and object
(model) hierarchy, Unity animator, or element inspector. This
enables greater modularity and allows game developers
familiarized with Unity to take advantage of uA’s added value
for educational videogames. A comparison between the old eA
GUI and the new uA GUI is presented in Fig. 3. In this case, an
eA user that is just starting out with uA can quickly become
productive: the interaction with the GUI is very similar to what
they already know, making for a smoother learning curve.
As uA evolves inside Unity, it will be able to benefit from
other Unity features, such as 3D assets (downloadable from the
Unity Asset Store), the physics engine, navigation meshes for
movement, or native animation state machines. Games will
result richer and developers should need less time to produce
them. In addition, Unity’s native features are already well
documented, can be reused between multiple games and
distributed as packages, and will be improved and maintained
as Unity evolves.
One of the improvements of uA is multimedia support with
the usage of the FFmpeg project for export-time video
conversion. Multimedia resources including video and audio are
complex to deal with when deploying a game in different
platforms. uA addresses those problems, not only when playing
the game (using Unity’s multimedia players), but also during
game creation. It uses FFmpeg for converting videos on-the-go,
increasing, for example, the number of supported video formats,
from AVI (DIVX 4, XVID, DIVX3low, S-Mpeg 4 V2, DIVX5)
and MPG (MPEG1), to almost every video format available.
This will greatly simplify the inclusion of new assets in games.
Therefore, using Unity as a support layer has not only
allowed us to include new features; but also has improved the
development and maintenance of uA. On one hand,
development requires a deep knowledge of the Unity platform
and experience in extending. On the other hand, some of the
most time-consuming maintenance aspects are no longer
required (e.g. coping with changes in mobile platforms,
multimedia playback) as they are provided by Unity. This allows
e-UCM’s resources to focus on improving and increasing the
educational features and the added value offered by uA.
C. The uAdventure interpreter: an emulator for eAdventure
The Interpreter loads game description file and makes it
playable. Taking advantage of Unity’s features, the Interpreter
extends the Extensible Game Engine, developing behaviors for
every element that is visible or plays a role in the game.
Furthermore, the Interpreter also manages the GameState,
including the state of the flags,variables,global-states, and
other game elements such as the inventory or the current scene.
Finally, it also orchestrates, relying on Unity’s MonoBehaviour
lifecycle, the execution of the game.
The Interpreter is not only responsible for representation,
but also for handling user interaction. Video games developed
with uA can be exported to a variety of platforms, which raises
a serious issue: different platforms imply different interaction
methods. uA supports different interaction methods, including
traditional mouse and keyboard as well as touch screen
environments. The framework provides developers with some
tools for making the game easier to play depending on user’s
runtime environment. For example, when a game is exported for
PC/Mouse-Keyboard system, the game will be presented like
older eA games, with different cursors, floating names, and lists
of answers at the bottom of the screen. On other hand, when a
game is exported for us in a Tablet/Touch system, interactive
elements glow periodically to capture the player's attention,
because tablet players can no longer “mouse-over” scenes to
find interactive elements.
uA also includes a multi-platform eA game emulator that
can take a “.jar” executable eA game (an already-exported,
ready-to-run eA game) and run it on every platform available
within uA. This emulator runs the previous games, but supports
new uA features (provided by the Interpreter). This includes
better animations, more effects, improved and adapted GUI, or
the ability to play eA games on mobile devices. The emulator
allows users to browse their filesystem, and easily import those
games they want to play on their preferred platform. Previously
launched games will be saved in the emulator with its
configuration that will enable learners to use old games on pre-
configured devices.
D. Migrating SGs from eAdventure to uAdventure
eA’s SG life cycle has been greatly improved with the
implementation of the uA Interpreter as most of the games
developed with eA in the past are now supported by uA. This
prevents the considerable effort that went into designing,
implementing and testing these games from being lost. As the
game is now difficult to deploy or does not run in new
environments it would have otherwise required a complete
redevelopment to achieve the same degree of support. That is,
the use of uA avoids the high cost in time and money that would
otherwise be required to reimplementation these older games in
new platforms.
There is an actual case study of adapting a game to a new
game engine, where an experienced game developer of the e-
UCM research group reproduced the First Aid [16] game inside
Unity. The production took almost 50 hours of development, and
even with that effort, only 85% of the game was supported (some
game situations were ignored). Based on this data, we can
estimate that 60 to 80 hours of development are necessary to
fully rebuild a simple eA game into a new game engine,
assuming that the developer responsible has access to the initial
detailed design document, and without factoring extensive user
testing, which may reveal additional problems. This is a
significant cost for the maintenance of a previously produced
SG, where in many cases there is no budget at all for on-going
maintenance. Indeed, the 6 educational games developed by the
educational organization CATEDU[17] using eA are in this
situation. Thanks to uA, this cost can be reduced to the few
minutes it takes to import the project into uA and generate new
versions. Indeed, we have taken the First Aid [16] game and
done exactly that; and are currently piloting it again in different
2017 IEEE Global En
g
ineerin
g
Education Conference (EDUCON)
978-1-5090-5466-4/17/$31.00 ©2017 IEEE 25-28 April 2017, Athens, Greece
Page 1758

schools in Madrid to test the learning analytics features
described in the next section.
IV. NEW EDUCATIONAL FEATURES IN UADVENTURE
In terms of educational features, there are few differences
between uA and eA. This section will focus on description of the
two main educational features that are being implemented in uA:
i) learning analytics as an assessment tool and ii) geolocated
games.
A. Learning Analytics for SGs
The evaluation of a formative activity that uses a SG poses
significant challenges. Usually the teacher has little control over,
or information of, the learner’s activity through the process of
playing a serious game [1]. This also happens usually when
using other eLearning technologies in the classroom, such as
MOOC or LMS. Some research results show that the more
teachers know about the learners’ behavior usinga learning tool,
the better teachers can control that learning tool [18].
eA’s assessment feature was based on conditions that tracked
player actions inside the game such as “The player has failed test
X” or “Player refuses to accept help from Non-Player Character
Y” [11]. These conditions, although useful for evaluation, were
one of the most complex features to master for developers of eA
games. This requires mastering conditions and, to some extent,
to have computational thinking abilities to understand how to
link the game states that have a correlation with the educational
learning goal; or with the skill that is put into practice while
playing the game. Moreover, the eA editor allows authors to
create multiple assessment profiles (sets of conditions) [19]. The
evaluation outcome was a textual report, which had to be
interpreted in terms of actual learning outcomes by an instructor.
In addition, evaluation based on pre-set conditions implies
that, the more the authors want to evaluate, the more time they
must spend creating conditions. Still, there are some aspects of
the learner’s behavior that are lost because conditions can only
track scenarios that are devised before the learning activity takes
place. Finally, this approach works for small groups of learners,
but as the groups become larger, the difficulty of supporting
corrective interventions during the learning activity increases:
more learners result in more the individual documents that must
be read and analyzed.
A complementary approach to condition based assessment,
is based on the application of big data and analytics techniques
where learners’ gameplay data is harvested, processed and
presented in a dashboard to the educator. This approach can be
used to provide metrics that will help the instructor to assess
learners or even discovering game design errors [20], [19].
In order to apply big data and analytics techniques, all
gameplay interactions must first be collected. Instead of creating
a custom format to represent theses interactions, we use an
existing e-learning specification called xAPI (eXperience API)
[21]. In xAPI, traces of interactions (called statements) are
composed of: i) an actor, that defines who performed the action,
12 http://www.rageproject.eu/
13 http://beaconing.eu/
ii) a verb, that defines the action between the actor and the
object, iii) an object, that defines the thing that was acted on, iv)
and optional properties such as result, representing a measured
outcome related to the statement, or a timestamp of when did
that happen [22]. Under this umbrella it is indeed possible to
represent gameplay actions; but in order to facilitate the
development of analytic tools that automate insights of the
learners' gameplay the e-UCM team, in collaboration with the
ADL initiative (developers of x-API), authored a specific xAPI
application profile for use with serious games [15].
The reasons for using xAPI are twofold. On the one hand,
this allows uA SG to interact with existing e-learning tools that
are compliant with the xAPI specification and also with, for
example, a Learning Record Store (a storage service for xAPI
statements). On the other hand, and even more important,
gameplay sessions will be compatible with future xAPI tools, for
example, it will be possible to apply new learning analytics tools
to gameplay sessions recorded in the past.
For example, with the current xAPI support (although
incipient) it is possible to integrate the SGs created with
uAdventure with the Learning Analytics application created as
part of the H2020 RAGE12 and BEACONING13 projects.
B. Embracing Ubiquity: geolocated games
One of the main advantages of the uA and its supporting
platform is the ability to create SGs for mobile devices.
Moreover, these devices usually include a plethora of sensors
that are unique to the mobile platform.
Getting into mobile allows games to be played anywhere and
anytime. This opens new opportunities to use SGs in different
educational scenarios, and using alternative interaction
mechanisms. For example, MIT’s MitAR14 is a serious game
platform that uses augmented reality and geolocation [23];
although it was never successfully ported to today’s major
mobile platforms. More recently, 2016 was the year of the
engaging AR mobile game “PokemonGo”, which not only
attracted large numbers of users, but also indirectly had an
impact in the players' wellness [24].
uA is starting to take advantage of the GPS, compass and
camera sensors that are already available to Unity games. This
will allow the creation of geolocated games that were beyond
the capabilities of eA, such as gymkhanas, interactive tours, or
cultural visitor guides. This new feature has been implemented
adding a new type of scene to the uA data model. Until now, a
game scene was a metaphor to represent a place and
environment where players or their avatars interact with the
game world. This new type of scene actually is a virtual
representation of the real environment where the game takes
place. This new map scene (and its companion editor) provides
access to a map to define POIs (points of interest), regions, or
even routes to link all the existing mechanics to the geolocation
gaming paradigm.
Finally, in order to be effective indoors, QR code scan
support was added. QR codes are easily generated, and can be
14 http://web.mit.edu/mitstep/projects/mitar-games.html
2017 IEEE Global En
g
ineerin
g
Education Conference (EDUCON)
978-1-5090-5466-4/17/$31.00 ©2017 IEEE 25-28 April 2017, Athens, Greece
Page 1759

placed on interesting locations to be scanned by those playing
on a mobile device; however, their use requires access to the
device’s in-built camera from within the game. In addition to
enabling indoor use, support for QR codes also allows games to
be changed or adapted to new locations by moving the
(physically placed) QR codes around. For instance, one of most
well-known games in the eA community is the “Fire Protocol”
game15, originally created as a demonstrator of eA’s
capabilities, and also as a drill for the actual fire protocol
procedure in the Computer Science building at Complutense
University of Madrid. Although the already uses pictures of
actual locations inside the building, forcing the players to
physically move around the actual locations depicted in the
game provides a much greater degree of immersion.
V. CONCLUSSIONS AND FUTURE WORK
The eAdventure (eA) platform has reached its
technological limit due to Java platform’s limitations
supporting mobile and web-based deployments. However, the
main goals and advantages of the eA platform are still relevant
for the SG community. eA’s SG life cycle has been greatly
improved with the implementation of uAdventure (uA) as most
of the games developed with eA in the past are now supported
by uA.
The uA project addresses these technological limitations,
and provides new educational capabilities that were not
envisioned when eA was created. In order to address current
and future technological challenges, uA is built on top of the
Unity platform, which provides solid technical underpinnings,
together with a strong user base and support in the game
development community.
The uA tool has been created around two main elements:
the Editor and the game Interpreter. The editor provides easy-
to-use game authoring tool for non-programmers that is very
similar to eA user interface, but includes some new features.
Some of them are: touchscreen interaction support, Learning
Analytics support, and geo-positioned game support with or
without maps. Touchscreen support includes features that make
playing point-and-click adventure games easier on
touchscreens. Learning analytics support is provided by an
integrated xAPI tracker that generates traces of what the learner
does during gameplay. Finally, geo-positioning support uses
mobile GPS and wireless capabilities, and new map scenes. Use
of Unity has also allowed us to make multiple game
representation improvements, giving games a better look and
feel. Furthermore, by taking advantage of Unity’s geolocation
capabilities, uA will allow the creation of games such as
gymkhanas or historical city tours.
The interpreter has been built using the very same model
used in eA, in order to maintain the compatibility with games
created with eA and facilitate the migration to the new platform.
In the case-study presented in Section III-D, it can reduce up to
85% percent of the maintenance and migration costs associated
15 http://e-adventure.e-ucm.es/course/view.php?id=29
with making a pre-existing SG capable of running on new
platforms.
Integration of uAdventure with the real-time Learning
Analytics platform used in the H2020 RAGE and
BEACONING projects is ongoing. This LA platform can
populate learning dashboards from incoming flows of xAPI
traces, and can display alerts and warnings when learners enter
risky areas or evidence anomalous behavior.
Although uA still requires significant work before public
release, we believe that uA with the capabilities described in
this paper will be a worthy heir of eA’s legacy. Hopefully, the
present work and the work done in the H2020 projects RAGE
and BEACONING (where the e-UCM team participates) will
facilitate a more widespread adoption of SGs in tomorrow’s
classrooms.
ACKNOWLEDGMENT
We would like to thank Piotr Marszal for his coding
contribution on the first beta of uAdventure. This research has
been partially financed by the Regional Government of Madrid
[eMadrid S2013/ICE-2715], by the Ministry of Education
[TIN2013-46149-C2-1-R] and by the European Commission
[RAGE H2020-ICT-2014-1-644187, BEACONING H2020-
ICT-2015-687676].
REFERENCES
[1] M. Nichols, “A theory for eLearning,” Educational Technol ogy and
Society, vol. 6, no. 2. pp. 1–10, 2003.
[2] “Serious Game Market worth $5,448.82 Million by 2020.” [Online].
Available:
http://www.marketsandmarkets.com/PressReleases/serious-
game.asp.
[3] “CodinGame: General Leaderboard.” [Online]. Available:
https://www.codingame.com/leaderboards/global?column=codingpo
ints&value=all.
[4] J. Torrente, Á. Del Blanco, E. J. Marchiori, P. Moreno-Ger, and B.
Fernández-Manjón, “e-Adventure: Introducing Educational Games
in the Learning Process,” in IEEE Ed ucation Engineering
(EDUCO N) 2010 Conference, 2010, pp. 1121–1126.
[5] M. Prensky, “Digital game-based learning,” Comput. Ent ertain ., vol.
1, no. 1, p. 21, Oct. 2003.
[6] M. D. Dickey, “Game Design Narrative for Learning: Appropriating
Adventure Game Design Narrative Devices and Techniques for the
Design of Interactive Learning Environments,” Educ. Technol . Res.
Dev., vol. 54, no. 3, pp. 245–263, Jun. 2006.
[7] A. Amory, “Building an Educational Adventure Game: Theory,
Design and Lessons,” J. Inte ract . Learn. Res., vol. 12, no. 2, pp. 249–
263, 2001.
[8] R. Van Eck, “Building Artificially Intelligent Learning Games,” in
Games and Simulations in Online Learning , IGI Global, 2007, pp.
271–307.
[9] P. Moreno-Ger, J. L. Sierra, I. Martínez-Ortiz, and B. Fernández-
Manjón, “A documental approach to adventure game development,”
Sci. Comput. Program., vol. 67, no. 1, pp. 3–31, Jun. 2007.
2017 IEEE Global En
g
ineerin
g
Education Conference (EDUCON)
978-1-5090-5466-4/17/$31.00 ©2017 IEEE 25-28 April 2017, Athens, Greece
Page 1760

[10] J. Torrente, P. Moreno-Ger, I. Martínez-Ortiz, and B. Fernandez-
Manjon, “Integration and deployment of educational games in e-
learning environments: The learning object model meets educational
gamin g,” Educ. Technol. Soc., vol. 12, no. 4, pp. 359–371, 2009.
[11] P. M. Ger, “eAdventure: Serious games, assessment and
interoperability,” in 2014 Int ernational Symposium on Computers in
Education (SIIE), 2014, pp. 231–233.
[12] G. McGraw and E. W. Felten, Java Security: Hostile Applets,
Holes&Amp;Antidotes. New York, NY, USA: John Wiley &
Sons, Inc., 1996.
[13] “NPAPI deprecation: developer guide.” [Online]. Available:
https://www.chromium.org/developers/npapi-deprecation.
[14] F. Messaoudi, G. Simon, and A. Ksentini, “Dissecting games
engines: The case of Unity3D,” in 2015 International Workshop on
Network and S ystems Support for Games (NetGames), 2015, vol.
2016–Janua, pp. 1–6.
[15] Á. Serrano-Laguna, I. Martínez-Ortiz, J. Haag, D. Regan, A. Johnson,
and B. Fernández-Manjón, “Applying standards to systematize
learning analytics in serious games,” Comput . Stan d. Interfaces, vol.
50, pp. 116–123, Feb. 2017.
[16] E. J. Marchiori, G. Ferrer, B. Fernandez-Manjon, J. Povar-Marco, J.
F. Suberviola, and A. Gimenez-Valverde, “Video-game instruction in
basic life support maneuvers,” Emergencias, vol. 24, no. 6, pp. 433–
437, 2012.
[17] CATEDU, “E-ADVENTURES. Download page.,” 2011. [Online].
Available:
http://www.catedu.es/webcateduantigua/index.php/descargas/e-
adventures.
[18] A. Ravenscroft, “Designing E-learning Interactions in the 21st
Century: revisiting and rethinking the role of theory,” Eur. J. Educ.,
vol. 36, no. 2, pp. 133–156, Jun. 2001.
[19] W. L. Wong, C. Shen, L. Nocera, E. Carriazo, F. Tang, S. Bugga, H.
Narayanan, H. Wang, and U. Ritterfeld, “Serious video game
effectiveness,” in Proceedings of the interna tional conference on
Advan ces in compute r entertainm ent technolog y - ACE ’07, 20 07, no.
January, p. 49.
[20] Á. del Blanco, J. Torrente, P. Moreno-Ger, and B. Fernández-
Manjón, “Enhancing Adaptive Learning and Assessment in Virtual
Learning Environments with Educational Games,” in Intelligent
Learning Systems and Advancements in Computer-Aided Instruction,
IGI Global, 2013, pp. 144–163.
[21] M. Manso Vazquez, M. Caeiro Rodriguez, and M. Llamas Nistal,
“Development of a xAPI application profile for self-regulated
learning requirements for capturing SRL related data,” in 2015 IEEE
Global Engineering Edu cation Conference (EDUCON), 2015, pp.
358–365.
[22] “xAPI-About @ github.com.” [Online]. Available:
https://github.com/adlnet/xAPI-Spec/blob/master/xAPI-
About.md#partone.
[23] B. Coulter, E. Klopfer, J. Sheldon, and J. Perry, “Discovering
Familiar Places,” in Games, Learning, and So ciety, C. Steinkuehler,
K. Squire, and S. Barab, Eds. Cambridge: Cambridge University
Press, 2016, pp. 327–354.
[24] T. Althoff, R. W. White, and E. Horvitz, “Influence of Pokémon Go
on Physical Activity: Study and Implications,” J. Med. Internet Res.,
vol. 18, no. 12, p. e315, Dec. 2016.
2017 IEEE Global En
g
ineerin
g
Education Conference (EDUCON)
978-1-5090-5466-4/17/$31.00 ©2017 IEEE 25-28 April 2017, Athens, Greece
Page 1761