Case Studies of xAPI Applications

to E-Learning
Kin Chew Lim
SIM University (UniSIM), Singapore
kclim@unisim.edu.sg

Abstract - The purpose of this paper is to (Corbi & Burgos, 2014), explained how the
demonstrate the applications of the xAPI and LIME model are used in their
Experience API (or xAPI – Experience study which helped them to monitor their
Application Programming Interface, and students and make recommendations in
also known as Tin Can API) in a few their studies. In this case, the xAPI LRS is
e-learning examples. The xAPI is a new used as an e-learning monitoring engine.
specification for learning technology. It
allows one to capture data in a consistent Since the xAPI specifications were first
format about a person or group’s activities released in 2013, they have received
from many technologies. Different systems widespread industry acceptance. It is also
can communicate securely by capturing and expected that there will be many more xAPI
sharing the activity streams using xAPI’s applications for e-learning.
simple vocabulary. xAPI is regarded as the
next evolution to SCORM (Sharable Keywords - Activity Stream, Game Learning,
Content Object Reference Model). SCORM Recommender System, xAPI
is used for packaging e-learning content for
interoperability of LMSs. However, I. INTRODUCTION
SCORM is now considered obsolete. xAPI is
an open source API. This allows software When computers were first invented in the
programs to read and write experiential late 1940s and early 1950s, they were
data in terms of statements like “I did this”, developed and used mainly for the military,
or “actor verb object”. Learning activities government and large corporate users. It was
like “I attended Conference C”, or “I only in 1960 that the first computer- based
tweeted Tweet E to Twitter” are stored in training (CBT) program was introduced [1].
the LRS (Learning Record Store). This was the PLATO, or Programmed Logic
for Automated Teaching Operations [2]. It
The first case study involves tracking in was originally designed for the University of
the game (Oregon Trail). Tracking is useful Illinois students but ended up being used in
in order to spot trends and make schools throughout the area [3]. Technology-
judgments about what activities are actually based training (TBT) and teaching using
working to help people learn things. This technology accelerated after personal
case study shows how the various game computers were introduced by IBM in the
activities are translated to the xAPI early 1980s.
statements. These can then be automatically
recorded in the LRS. If the data can co-exist Subsequently, many courseware titles were
with other learning data (e.g. test results) developed and delivered via CD-ROMs and
then we can understand how the student laser disks. These gave way gradually to the
learn to play in the game. The second case learning management systems (LMSs) when
study is the LIME (Learning, Interaction, computer systems became more powerful and
Mentoring and Evaluation) model case could store more contents. It was the AICC
study. The authors, Corbi and Burgos [4] which released the first specification for

The Twelfth International Conference on eLearning for Knowledge-Based Society, 11-12 December 2015, Thailand
3.1
 Case Studies of xAPI Applications to E-Learning

the LMS. This specification allowed students’ place. On the other hand, smartphones are not
scores to be tracked on the computer system always connected to the Internet. Finally, it is
he was using. When the Internet became a difficult to ascertain how much learning the
world-wide sensation in mid-1990s, Web- participant has done if he or she uses multiple
Based Training (WBT), virtual classrooms and devices to access information [7].
e-learning in general became fashionable. At
about the same time, several learning B. What led to the Development of the xAPI?
standards consortia were founded. These SCORM was first released in 2000 [8]. It
included the IMS Global Learning Consortium has served its purpose of achieving
[5] and the Advanced Distributed Learning interoperability in different LMSs. But since
(ADL) Project [6]. t hen t he landscape has changed
tremendously. Firstly, there is an extensive
SCORM (Shareable Content Object worldwide proliferation of mobile devices and
Reference Model) was released by the ADL the mobile app ecosystem. People are now
Project in the year 2000. using different mobile devices to receive
information, communicating, learning and
SCORM is the de facto specification for collaborating amongst themselves. At the same
packaging learning content is a standard time wireless and Wi-Fi coverage are
format which allows the package to work in increasing everywhere. People everywhere
different LMSs. However, SCORM is tied engaged in games, whether on the web, using
very closely with the LMS. It will not work the console or mobile devices. Applications in
outside of the LMS and the browser. augmented reality and simulations are
spreading not only on the desktop computers
A. Shortcomings of Present LMS-Centric and but on mobile devices like the iPADs. People
Content-Centric E-Learning are also communicating extensively using
So far, the approach in e-learning is to social media tools like Facebook, Twitter,
deal with how the content is to be structured, Instagram and blogs. Open source movement
packaged and moved from one system to is gaining widespread use with people
another. This is a very LMS- and content- everywhere [9].
centric model. SCORM is thus very LMS- and
content-centric and hence it has many A person might be texting one moment.
restrictions. Next moment, he used a desktop computer to
access an LMS to do an online quiz. After a
For example, multiple-choice quizzes are while he might be in a restaurant discussing
used widely in the LMS-centric model. These business deals with his client. For this, he
quizzes are usually of the single- answer used an iPAD. Later in the afternoon, he
assessments. Questions with single answers could be attending a 1-hour webinar using his
do not reflect real world situations in which Android smartphone. All these activities show
there might not be single-solution answers. that very little online learning happens on the
Learners also could guess the answers. The LMS! The LMS is used only as a repository
materials provided in the LMS are mostly of learning materials.
textual in nature although occasional video and
animation clips were used. The LMS-centric Subsequently, the ADL of the US
model will always have the teacher as the Department of Defense engaged Rustici, an
knowledge dispenser. Participants in an LMS- e-learning software company, to work on a
centric model do not share much. In addition, new proposal for the new generation of
contents from other devices (e.g. smartphones, e-learning specification. After extensive
tablets and social media) were difficult to be consultations with the e-learning community,
consolidated with those on the LMS. The Rustici developed the Tin Can API in 2013.
LMS must be connected to the Internet all the The ADL lat er renamed it xAPI, for
time in order for learning interactions to take Experience API. Version 1 of t his

The Twelfth International Conference on eLearning for Knowledge-Based Society, 11-12 December 2015, Thailand
3.2
Kin Chew Lim

specification was released in April 2013 [10]. called Oauth [19]. Another use of the xAPI is
The current version is at version 1.0.2 [11]. that of platform transition; e.g. an e-learner
Your paper must be in two column format starts e-learning on a mobile device and
with a space of 4.2mm (0.2") between finishes it on a comput er [19]. Ot her
columns. possibilities include those of tracking games
and simulations [13], tracking real-world
II. WHAT IS THE XAPI? performance [20], tracking team-based
e-learning [13] and tracking learning plans and
The Experience API forms part of the goals [21].
Training and Learning Architecture (TLA) [6]
that the Advanced Distributed Learning (ADL) A. xAPI Statements
Project is working on. This API (also known The xAPI is a web service. A web service
as the Tin Can API), is an open source supports applications on the World Wide Web
e-learning software specification [12]. The (WWW) and makes use of the HyperText
specification makes it possible to collect data Transfer Protocol (HTTP). As a web service,
about the learning experiences a person has the xAPI allows for statements of experience,
achieved either online or offline. Learning typically learning experiences, to be delivered
experiences are recorded in a Learning Record to and stored securely in a Learning Record
Store (LRS). LRSs can exist within traditional Store (LRS).
Learning Management Systems (LMSs) or on
their own [13]. The web service allows clients to read and
write experiential data in the form of
The Experience API is commonly “statement” objects. In their simplest form,
considered the successor to SCORM (Sharable statements take the form of “I did this”, or
Content Object Reference Model) [14]. Since more generally “actor verb object”. xAPI also
2000, SCORM has been the de facto provides facilities for more complex statement
e-learning standard for packaging e-learning forms [22].
content to be delivered to LMSs. (Training
Industry Magazine, 2014). However, there are
several drawbacks to SCORM [15].

This API is stewarded by ADL. xAPI

focuses on how the activities people do are
evidence of a learning experience. It is a
Representational state transfer (REST) web
service. As for the data format, it uses the
JavaScript Object Notation (JSON). The web
service allows software clients to read and Fig 1. The Basic Elements and Structure
write experiential data in the form of of an xAPI Statement
“statement” objects. Statements are in the
form of “I did this”, or more generally “actor In the example, “Andrew experienced ‘Solo
verb object”. [16]. More complex statement Hang Gliding’”, we recognize that “Andrew”
forms can be used. is the actor, “experienced” is the verb, and
“Solo Hang Gliding” is the activity. The
With the xAPI, e-learners can take statement object itself would take this structure
e-learning outside of the browser [17]. In in JSON (JavaScript Object Notation) format:
addition, xAPI allows e-learning to execute in
native mobile applications [18]. Thus, there is "actor": "Andrew", "verb":
more control over the learning content should "experienced",
the xAPI specification be used. Not only that, "object": "Solo Hang Gliding"
there is better security using a technology }

The Twelfth International Conference on eLearning for Knowledge-Based Society, 11-12 December 2015, Thailand
3.3
 Case Studies of xAPI Applications to E-Learning

This is a simple example. How do we know B. Learning Record Store (LRS)

which Andrew we mean? Which ‘Solo Hang A Learning Record Store (LRS) is a place
Gliding’ activity was it? Was it one that was to store learning records. The LRS is a new
part of military training, or the one from a system that goes together with the xAPI. As
commercial enterprise, or something self- xAPI-enabled activities generate statements,
directed? Here is a valid xAPI statement: they are sent to an LRS. The LRS is simply a
repository for learning records that can be
{ accessed by an LMS or a reporting tool. An
"actor": { LRS can live inside an LMS, or it can stand on
"name": "Andrew Downes", its own. LRSs record all of the statements
"mbox": mailto:andrew@example.com made. An LRS can share these statements with
}, other LRSs.
"verb": {
"id":
"http://adlnet.gov/expapi/verbs/experienced",
"display": {"en-US": "experienced"}
},
"object": {
"id":
"http://example.com/activities/solo-hang-
gliding",
"definition": {
"name": { "en-US": "Solo Hang Fig 2. Learning Record Store (LRS)
Gliding" }
} The data stored in an LRS can be accessed
} by LMSs, reporting tools, or other LRSs. The
} data can be stored as individual learning
records and/or entire transcripts. An LRS can
A structure has been added here to ensure limit who can read and write learning records.
that we can uniquely identify the component
SCORM and other e-learning standards
parts. This helps to correlate statements
only store a certain amount of learning data.
about the same person, activity, or verb. There
xAPI allows for the LRS to store nearly
is also a st ruct ure added t o provide
everything. This means better reporting and a
information about the objects, like name.
much more accurate picture of learners.
Other descriptive fields are available
(Experience API, 2015). An LRS can live in an LMS and use the
LMS’s reporting tools to make meaning of the
We can also add a lot more to a statement,
LRS’s data. Or it can live on its own with its
in the form of statement context (“Andrew
own reporting tools.
completed ‘Solo Hang Gliding’ in the context
of ‘Army Training Level 1’” or “Bob LRSs can share data amongst themselves,
completed ‘Truck Driving Training Level 1’ so learners and data can be transferred from
on his Android phone, under the instruction of one organization to another. Statements can
Dan”) or you can attach results to a statement also be sent to multiple LRSs (e.g. “I want to
(“Bob passed ‘Truck Driving Training’ with record my training in my own personal LRS as
score 90%”). You can even declare custom well as my employer’s LRS.”)
fields on a statement, in the form of extensions
(Rustici Software 4, n.d.).

The Twelfth International Conference on eLearning for Knowledge-Based Society, 11-12 December 2015, Thailand
3.4
Kin Chew Lim

The game do not always get played in a

browser. It is not always being played on a
“connected” device. It was definitely not
played in an LMS. However, sometimes it is
played as a mobile app. These things do not
work with older standards like SCORM and
AICC.

Why would anyone want to record the

learning experiences created in the game?
Fig 3. Sharing Data among LRSs and LMSs Furthermore, why would we use an e- learning
standard to record the learning experiences?
III. CASE STUDIES IN ELEARNING Well, just like any other learning data, tracking
it is useful to spot trends and make judgments
A. Oregon Trail Game about what activities are actually working to
Background - The Oregon Trail is a help people learn things. If this data could be
comput er game [23]. It was originally tracked such that it could live side by side with
developed by Don Rawitsch, Bill Heinemann, other learning data (test results, for example)
and Paul Dillenberger in 1971. It was then the big picture of teaching and learning is
produced by the Minnesota Educational easy to see.
Computing Consortium (MECC) in 1974. This
game was designed to teach school children The xAPI (or Tin Can API) makes the game
about the realities of 19th century pioneer life much more useful. Firstly, it makes it possible
on the Oregon Trail. The player assumes the to record all the learning data that is generated.
role of a wagon leader guiding his or her party Secondly, that data lives side by side with all
of settlers from Independence, Missouri, to of the other data that being generated by
Oregon's Willamette Valley on the Oregon students and teachers.
Trail via a covered wagon in 1848. The game
is the first entry in the Oregon Trail series of
games. It has since been released in many
editions by various developers and publishers.
The Oregon Trail was extremely successful. It
sold over 65 million copies [24].

The Details - Oregon Trail is an e-learning

game. First developed in 1971, it has since
then been played on many platforms such as
Apple computers, Windows computers, iOS
devices, Nintendo, and even Blackberrys. In Fig 4. The Oregon Trail Game
fact, it has been sold 65 million copies.

As an e-learning game, it teaches people by

making learning fun and interactive for the
learner. Unfortunately, most of the learning
activities that are created do not get recorded.
These learning activities create valuable
learning experiences for the learner. But they
are often not measured.

Older e-learning standards did not work for

recording the experiences created in the game.
Fig 5. Oregon Trail – e.g. Alan Choose to be a Carpenter

The Twelfth International Conference on eLearning for Knowledge-Based Society, 11-12 December 2015, Thailand
3.5
 Case Studies of xAPI Applications to E-Learning

allows the learning data to exist in the same

system as all of the other learning data of the
learner. Instructors can see the big picture of
learning and teaching. It also helps them to
better figure out effective ways to teach.

B. LIME Model Case Study

The second case study is the LIME
(Learning, Interaction, Mentoring and
Evaluation) model case study. According to
the authors, Corbi and Burgos [25], this case
Fig 6. Oregon Trail – “You have Crossed Kansas River” study helped them to monitor their students
and make recommendations in their studies.
The xAPI statements, as played by the
character, Alan, are as follows:

Alan’s
xAPI Statement
Experience
Alan chose to be Alan completed “choose
a carpenter. occupation” with a result of
“carpenter.”
Alan added Alan completed “add member
Ryan to his to your party” with result of
party. Ryan.
Alan added Alan completed “add member Fig 7. Examples of using the xAPI
Mike to his to your party” with result of
party. Mike. 1. Recommender Engines: In recent years,
Nicole got Alan experienced “party recommender engines have become very
dysentery, and popular. You can now find recommendation
member getting dysentery.”
Alan chose to Alan experienced “ignoreengines in books, movies, music, news,
continue his dysentery and continue on
products, research articles, search queries, and
journey rather journey.” social tags in general. These engines deliver
than take action. suggestions based on the collected information
Alan made it to Alan completed “arrival at on preferences, general user behavior and even
the Kansas Kansas River crossing”. items bought or content searched. Students
River. depend on recommendations from their peers
and professors in order to do their research.
There are two ways to track the learning in Lately, the research community is paying
the Oregon Trail game. The first way is to much attention to recommender engines.
access the game’s program codes and
incorporate the xAPI statements. However, Fig. 8 shows, on the Y-axis, the number of
this might not be possible if the game’s codes cited papers from each year as of 2013. There
are not accessible to other people. The second is a peak of interest around the year 2009.
way is to develop the simple program which
allows the various xAPI statements to be
recorded directly into the LRS. This is the
preferred way.

Games are an immersive and effective way

for people to learn, but they have previously
not been well tracked. xAPI allows one to
track whatever one wants to in a game. It also

The Twelfth International Conference on eLearning for Knowledge-Based Society, 11-12 December 2015, Thailand
3.6
Kin Chew Lim

LRS implementation was open-sourced by

ADL (Advanced Distributed Learning). It is
based on the Python computer language and on
the Django web framework.

Fig 8. No. of Papers & Workshops Versus Year

Corbi and Burgos (2014) proposed using

the LIME model [26] to develop the Fig 10. LRS Software Stack & Interaction
recommender engine. The LIME model is
based on four vectors: When students interact with educational
content using different systems or tools, they
i. What every learner does based on his/her
will leave traces in the LRS. Each tool will
own contribution (L = Learning) provide a totally different actor/user ID to
ii. What the learner does to support
ensure anonymity.
interaction-based learning and the
relation with others, in addition to group The verb element is a key part of an LRS
interaction (I = Interaction) communication. This is because it describes
iii. What t eachers/expert s value the action performed by the student. In an
(M=Mentoring) elearning environment, a verb could be
iv. A transversal vector is applied to the somet hing like: “writ e”, “read”,
three previous vectors, focused on “experiment ed”, “passed”, “failed”,
evaluation (E=Evaluation) “experienced”, etc.

The object/activity part of the statement

refers to “what” was experienced in the action
defined in the verb. It usually corresponds to
the learning activity (e.g. twitter, webinar,
wiki, chat room, forum, mail message, etc.).
Objects/activities must also embody a URL
(Universal Resource Locator) pointing to their
rationale. This can include other information
such as the learning activity’s description,
verbs that can apply, possible results and usage
suggestions. The result component provides
Fig 9. Examples of the LIME Model the outcome to the statement. It includes score,
level of success and completion fields.
2. xAPI as an E-Learning Monitoring
Engine: The LRS (Learning Record Store) is The context part adds more details to the
the core of the xAPI. The LRS is a specific overall statement, like the relationship of the
module for data storage that allows an LMS to activity with other activities, its order in the
report tracking information on the learning learning stream, or the teacher’s name.
experience. At any time, an LMS can send
collected data over the network to an xAPI An LRS must also implement REST calls
web service. An LRS is a wrapper or API for data transfer (PUT, POST, GET and
software layer to a SQL database. This free DELETE). The xAPI can make use of either

The Twelfth International Conference on eLearning for Knowledge-Based Society, 11-12 December 2015, Thailand
3.7
 Case Studies of xAPI Applications to E-Learning

OAuth or HTTP Basic Authentication when

communicating securely with the outside
world.

One key aspect of the LRS architecture is

that it can be implemented in shared cloud
ecosystems. This allows communications
from very different elearning platforms and
academic institutions. In other words, Fig 11. Categories & Settings in the LIME Model
monitoring data can be uniformly stored. This
allows rapid, vast and democratic access to In the LIME model each input (action
learning analytics information. There are some
performed by a student in the eLearning
free LRS hosting services but mainly for platform or Social Network) is attributed a
testing and technology promotion purposes. category and a weight, assigned by the
One such service is by the ADL instructor. An example of model configuration
(http://lrs.adlnet.gov/xAPI) and another one by
for a specific site can be found in Fig. 12.
Rustici Software
Based on these components, the lecturers-
(http://tincanapi.com/prototypes/). tutors can manually define and parameterize
recommendation rules. These rules will only
3. The LIME Model and the LRS:
trigger a message to the student if conditions
Lecturers-tutors must design a strategy for regarding categories, inputs and settings are
each of his/her courses. The model codifies met.
this strategy for a course or class group by
using settings and categories.

A course setting is the balance between

formal and informal scenarios. In this context,
formal means a regular academic program
with regular evaluation means (e.g. graded
exams); informal means continuous evaluation
and user activity inside the LMS and every
tool linked to it (e.g. Social Networks or
repository). The system collects specific inputs Fig 12. Sample Configuration of the LIME Model
from both settings, keeping an overall balance for a Specific Course Site
of 100%. For instance, if the designer requires
just a formal setting, the balance should be LIME is a lecturer-tutor-crafted, rule-based
Informal: 100% - Formal: 0%. recommender syst em for learning
environments on the cloud. LIME’s goal is
Furthermore, a learning scenario must be
simply to improve learning efficiency, and to
defined as the balance between the Learning,
facilitate the learning journey of every student
Interaction, Mentoring, and Evaluation, in
by a personalized recommendation set. LIME
combination with the Formal and Informal
can be fed from learner inputs in many ways.
settings categories. In the LIME model, every
However, this model can also be initialized
category and setting are assigned with a
with tracked data stored in an xAPI LRS
specific weight (wi), keeping an overall
instance/server.
balance of 100%.
How can LIME inputs be built out of
information stored in the LRS? A LIME model
input has to define an action and a context in
which a learner performs this action:

The Twelfth International Conference on eLearning for Knowledge-Based Society, 11-12 December 2015, Thailand
3.8
Kin Chew Lim

 participate in chat
 answer in main forum thread
 message to tutor
 resolve a problem set
 formally broadcast mail to mates

xAPI verbs and objects, taken in an isolated

way, are not sufficient. However, a joint entity
composed of a verb plus an xAPI object makes
more sense. Fig 15. Predicate Filtering in LIME

As LIME was developed as a Basic

Learning Tool Interoperability (Basic LTI)
[27] application, this equivalency list can even
Fig 13. LIME Inputs from xAPI Statements be stored in the LMS database through the
IMS LTI Settings API specification, part of
It is up to the implementer to define which IMS LTI 1.0 and above. The LIME model thus
verbs and objects best represent the scenario to remains free from external configuration files
be tracked and monitored. Let us take a look at or own database management.
the sample verbs and activities available on the
official xAPI sit e (i.e. It is important to notice that LMS must be
ht t p://adlnet .gov/expapi). LTI compatible and support the Settings API
protocol.
In Fig. 15 are listed all the verbs and
activities the LRS can store. We also have
their possible combinations to build a
meaningful and compatible LIME input.

Fig 16. Aggregation of LRS Sentences

Fig 14. xAPI Verbs and Objects to LIME Inputs

These aggregation operations are covered

Each input should be assigned a weight
by the xAPI standard. The xAPI provides a
(wi), a category and a setting. These
query language to easily data-mine an LRS.
parameters should reside on the LIME
For instance, the following code collects all
system’s own configuration repository. In
the times the user “Alan” has tried an exam,
other words, LIME administrators should
and returns an aggregated result:
maintain an updated equivalency list between
LRS vocabulary and LIME inputs. These stmts.where (
inputs will then interplay with rules (Figure 'actor.name = “Alan” and ('+
11), which are, in turn, based on predicate 'verb.id =
filtering. “http://adlnet.gov/expapi/verbs/passed”'+
'or '+

The Twelfth International Conference on eLearning for Knowledge-Based Society, 11-12 December 2015, Thailand
3.9
 Case Studies of xAPI Applications to E-Learning

'verb.id = learning+ informal learning = 100%). The

“http://adlnet.gov/expapi/verbs/failed”'+ ')')lecturers- tutors can then manually define and
parameterize recommendation rules. These
IV. CONCLUSION rules will only trigger a message to the student
if conditions regarding categories, inputs and
This paper traces the background of the settings are met.
Experience API, or xAPI in short.
These two case studies are just some of the
Where previously much attention was increasing number of xAPI applications being
placed on structuring the e-learning contents, launched rapidly. The author feels that with
the trend is now moving towards measuring widespread support from the industry,
learning experiences. This comes about as government and academic bodies, there will be
there is realization that new technologies like even more and better xAPI applications in the
mobile, virtual and multimedia technologies future.
are constantly changing. However, it is
learning that all of us are more interested in. REFERENCES
xAPI has been designed to store use data in (Arranged in the order of citation in the
a simple, centric, standard, client agnostic and same fashion as the case of Footnotes.)
powerful way. A central part of this is the
Learning Record Store, or LRS in short. With [1] Bersin, J. (2004). “The Blended
this record store, one can measure the learning Learning Book: Best practices, proven
component in any elearning activity. methodologies and lessons learned”.
Pfeiffer.
The first case study is about measuring the [2] Lombart, P. (2011). “PLATO
learning that can take place in the Oregon Trail (Programmed Logic for Automatic
game. This is an adventure game in which a Teaching Operations)”.
person can take on a role and goes on an <http://whatis.techtarget.com/definition/
adventure trail called the Oregon Trail. The PLATO-Programmed-Logic-for-
gamer can assume various roles in his journey Automatic-Teaching-Operations>.
across the USA. He can cross a river, eat Accessed 11 May 2015.
some food and even fall sick during his [3] Epignosis, LLC. (2014). “E-Learning
journal. xAPI statements can be drafted and concepts, trends, applications”.
stored as indications of his or her learning <http://www.talentlms.com/elearning/ele
experiences. These statements can be stored arning- 101-jan2014-v1.1.pdf>.
on a Learning Record Store (LRS). The second Accessed 20 April 2015.
case study is about using the xAPI statements [4] Aviation Industry CBT Committee
to make recommendations for the student. (AICC). (n.d.). “Aviation Industry
Recommender engines or systems are Computer-Based Training Committee”.
becoming popular in many activities like <https://en.wikipedia.org/wiki/Aviation_
ordering books, reading research papers, Industry_ Computer-
booking hotels, going to restaurants, choosing Based_Training_Committee>. Accessed
movies, buying cars and making investments. 27 November 2015.
The core of the second case study is the LIME [5] IMS Global Learning Consortium.
(Learning, Interaction, Mentoring and (2015). “About IMS Global Learning
Evaluation) model. With this model, it is Consortium”.
possible to capture the formal and informal <http://www.imsglobal.org/background.
learning processes going on. By using the html>. Accessed 11 May 2015.
LIME model, every category and setting [6] Advanced Distributed Learning. (n.d.).
assigned with a specific weight (wi). An “Advanced Distributed Learning:
overall balance of 100% is kept (i.e. formal Capabilities: Training and Learning

The Twelfth International Conference on eLearning for Knowledge-Based Society, 11-12 December 2015, Thailand
3.10
Kin Chew Lim

Architecture”. Project Tin Can – The Next Generation

<http://www.adlnet.gov/index.html>. of SCORM. Float Mobile Learning.
Accessed 11 May 2015. <http://floatlearning.com/2012/04/project
[7] Advanced Distributed Learning (ADL) -tin-can- the-next-generation-of-scorm/>.
Co-Laboratories. (2012). “An ADL Accessed 19 April 2015.
Perspective on Next Generation SCORM [15] Whitaker, A. (2012). “An Introduction to
Requirements as Derived from Project the Tin Can API”. An Introduction to the
Tin Can”. Tin Can API. The Training Business.
<http://www.adlnet.gov/wp- <http://www.thetrainingbusiness.com/so
content/uploads/2012/01/NEXTGEN- ftwaretool s/tin-can-api>. Accessed 19
SCORM- requirements- April 2015.
20120130_v1.pdf>. Accessed 10 May [16] Tillett, J. (2012). “Saltbox Developers
2015. Discuss Tin Can”.
[8] Glahn, C. (2013). “TinCan and The <http://floatlearning.com/2012/07/saltbo
Confusion About the Next x-developers-discuss-tin-can/>.
Generation of SCORM”. <http://lo- Accessed 10 May 2015.
f.at/glahn/2013/05/the-confusion-about- [17] eLogic Learning. (2012). “eLogic
the-next-generation-of-scorm.html>. Learning Partners with Rustici Software
Accessed 10 May 2015. to be an Early Adopter of the Next
[9] Hruska, N. (2013). “The Experience Generation of SCORM Standards
API”. <http://www.adlnet.org/wp- Known as the 'Tin Can API. (2012)”.
content/uploads/2013/04/The_Experienc <http://www.prweb.com/releases/SCOR
e_API_i n_Practice.pdf>. Accessed 21 M/e-learning/prweb9610860.htm>.
April 2015. Accessed 19 April 2015.
[10] GitHub: Experience API. (2015). [18] Brandon, B. (2012). “Making History:
“GitHub Web site”. mLearnCon 2012 Rocks Attendees”.
<https://github.com/adlnet/xAPI- Making History: mLearnCon 2012
Spec/blob/1.0.3/xAPI.md>. Accessed 9 Rocks Attendees. Learning
May 2015. Solutions Magazine.
[11] Experience API. “At version 1.0.2. <http://www.learningsolutionsmag.com/
(2015)”. articles/95 8/>. Accessed 19 April 2015.
<https://github.com/adlnet/xAPI- [19] Project Tin Can Phase 3 - the future of
Spec/blob/master/xAPI.md>. Accessed e-learning is now. (2012). “We Need
27 November 2015. Security/Authentication”.
[12] Bowe, M. (2013). “The Open Source <http://scorm.com/project-tin-can-
Landscape”. phase-3-we-need-
<http://tincanapi.com/2013/07/11/the- securityauthentication/>. Accessed 19
open-source- landscape/>. Accessed 19 April 2015.
April 2015. [20] Gautam, A. (2012). “Tin Can: My First
[13] Brusino, J. (2012). “The next generation Impressions From mLearnCon 2012”.
of SCORM: a Q&A with Aaron Silvers”. Tin Can: My First Impressions From
American Society for mLearnCon 2012. Upside Learning.
Training and Development. <http://www.upsidelearning.com/blog/in
<https://www.td.org/Publications/Newsl dex.php/2 012/06/21/tin-can-my-first-
etters/Lear ning-Circuits/Learning- impressions-from-mlearncon-2012/>.
Circuits- Archives/2012/06/The-Next- Accessed 19 April 2015.
Generation-of-SCORM-a-Q-and-a-with- [21] Downes, A. (2012). “I Want This: Tin
Aaron-Silvers>. Accessed 19 April Can Plans, Goals and Targets”.
2015. <http://tincanapi.co.uk/pages/I_Want_Th
[14] Tillett, J. (2012). “Project Tin Can is.html>. Accessed 19 April 2015.
– The Next Generation of SCORM”. [22] Rustici Software 4. (n.d.). “Tin Can API:

The Twelfth International Conference on eLearning for Knowledge-Based Society, 11-12 December 2015, Thailand
3.11
 Case Studies of xAPI Applications to E-Learning

Statements 101”.
<http://tincanapi.com/statements-101/>
Accessed 10 May 2015.
[23] Oregon Trail. (2014).
<https://archive.org/details/msdos_Oreg
on_Trail_ The_1990>. Accessed 27
November 2015.
[24] Rustici Software. “A Game's Story”.
<https://tincanapi.com/a-serious-games-
story/>. Accessed 27 November 2015.
[25] Corbi, A. and Burgos, D. (2014).
“Review of current student-monitoring
techniques used in elearning-focused
recommender systems and learning
analytics”. The Experience API & LIME
model case study. International Journal
of Artificial Intelligence and Interactive
Multimedia, Vol. 2, No 7.
<http://www.ijimai.org/JOURNAL/sites/
default/files/files/2014/09/ijimai20142_7
_6_pdf_27449.pdf>. Accessed 8 May
2015.
[26] Burgos, D. (2013). “L.I.M.E. A
recommendation model for informal and
formal learning, engaged”.
<http://www.ijimai.org/journal/sites/defa
ult/files/files/2013/06/ijimai20132_2_11
_pdf_25682.pdf>. Accessed 27
November 2015.
[27] Learning Tools Interoperability. (2014).
<https://www.imsglobal.org/activity/lear
ning- tools- interoperability>. Accessed
27 November 2015.

The Twelfth International Conference on eLearning for Knowledge-Based Society, 11-12 December 2015, Thailand
3.12