Towards a New Generation of Learning Management Systems
Hubert Vogten
1
and Rob Koper
2
1
Faculty of Psychology and Educational Sciences, Open University of the Netherlands,
Valkenburgerweg 177, Heerlen, The Netherlands
2
Open University of the Netherlands, Heerlen, The Netherlands
Keywords: LMS, Learning Network, Distance Education, Liferay, Portal.
Abstract: In this paper we argue that a centrally governed Learning Management System (LMS) still has ample
legitimacy in an information society that is ever more adopting cloud computing services in daily life. We
argued that control over services and produced data is essential from the perspective of an educational
institute for reasons of accountability, quality control, legislation, privacy and reliability. However, the
current generation learning management systems are primarily geared to provide ‘additional’ online
learning. In ‘real’ online learning teachers and students almost never need to meet face-to-face. All
instruction, tests, communication and collaboration is organised using internet and mobile technologies. We
propose a paradigm shift for the next generation LMS, discarding the course as key concept in favour of the
social learning network concept. We argue that a generic social collaborative portal platform is a good
foundation for the development of this next generation LMS. We support our arguments by presenting a real
world case and we conclude that we can reuse 80% of the standard code.
1 INTRODUCTION
So, what to do when you are seriously planning to
provide ‘real’ online learning for students you will
never meet or see? Are the traditional LMSs that are
currently locally implemented as an add-on to
campus-based education usable for these purposes or
is it better to implement and use another solution?
The uptake of ‘real’ online learning has grown
massively in the past years driven by the public
attention and uptake of Massive Online Open
Courses (MOOCs). The term ‘real’ in the previous
sentence is contrasted to the term ‘additional’ online
learning, i.e. additional to campus-based teaching
and learning. In real online learning teachers and
students almost never need to meet face-to-face. All
instruction, tests, communication and collaboration
is organised using internet and mobile technologies.
In the last decade many universities and schools
have implemented a dedicated Learning
Management System (LMS) such as Blackboard or
Moodle to be used in addition to the regular campus-
based teaching and learning. Its use is mostly
restricted to the sharing of PowerPoint slides,
information about grades, examinations, classroom
changes, providing access to online books, papers
and other resources. The real work and instruction is
mostly done outside the LMS. Given that the main
users of LMSs are using the LMS as an addition, it
is likely that the requirements for the development
of these platforms are mostly driven by this
extended classroom paradigm and not by
requirements stemming from ‘real’ online learning.
In practice most universities that are using MOOCs
use different (or adapted) platforms, like Udacity,
Coursera or Futurelearn.
Using these MOOC platforms may be
satisfactory from the learner’s perspective; it is not
necessary desirable from the educational
institution’s perspective, for whom the online
platform is an essential, integral part of the core
business. Such educational organisation is obliged to
offer these services at an agreed level of quality and
availability. Furthermore, educational institutions
want to have access to all learning performance data
so the learning and teaching can be improved by
applying learning analytics technologies. Also
legislation, internal quality control and privacy
issues imply that an educational institute is
responsible and accountable for these data. An
educational institution simply cannot take
accountability for services in the cloud that are not
514
Vogten H. and Koper R..
Towards a New Generation of Learning Management Systems.
DOI: 10.5220/0004955805140519
In Proceedings of the 6th International Conference on Computer Supported Education (CSEDU-2014), pages 514-519
ISBN: 978-989-758-020-8
Copyright
c
 2014 SCITEPRESS (Science and Technology Publications, Lda.)

under their control. Hence, an integrated, centrally
governed infrastructure for online learning has still
got its legitimacy. However, such an infrastructure
requires a paradigm shift from a limited view on
online course support towards a more inclusive view
on ‘real’ online learning. This raises a question:
what is the best technical foundation for
implementing such paradigm shift. This question
will be addressed in the next sections by discussing
an implementation of such an infrastructure at the
Open University of the Netherlands.
2 THE LMS BEYOND
THE COURSE
In 1984 the Open University of the Netherlands
(OUN) was established as an institute for open
higher distance education for adult learners. Its
major goals were (1) to offer adult learners a second
chance to higher education and (2) to provide an
alternative route to higher education in order to
reduce the load on costly traditional, face to face
education. The Open University started in 2010 with
the development of the infrastructure for their next
generation LMS, which is called OpenU.
Hermans, Kalz, & Koper (Hermans et al., 2013)
distinct three types of online environments in which
adult learners act:
 The LMS, built around the course concept and
intended for formal instruction;
 The Personal Learning Environment (PLE),
governed by the learner;
 Social network sites and learning networks (LN)
for social and informal learning.
OpenU has the ambition to support all three
online environments through the same infrastructure.
Therefore, OpenU should be able to support various
target groups in their formal and informal learning
needs in a distance education setting. These target
groups include master, bachelor and PhD students,
but also professionals wanting to keep up with latest
developments and trends in their area of expertise.
But also researchers should be enabled to showcase
and discuss the state of art in their topic of research,
i.e. to increase the impact of their research. OpenU
should allow learning networks to be established for
all target audiences, but it should also be possible to
cross these boundaries and bring students,
professionals and researchers into contact with each
other. Informed by a social constructivism view on
learning, emphasizing that the development of
knowledge and skills require intensive social
interactions (Schunk, 2012) OpenU should provide
ample social tools. Finally users should be allowed
to construct their own personal learning environment
and the self-directed learner should be encouraged to
explore formal and informal learning opportunities.
Therefore a substantial part, about 10%, of all
learning materials will be offered as Open
Educational Resource (OER). Students and
professionals should be able to receive credits for
their participations in these MOOCs, which can be
used in either their curriculum or for their
professional development. From these use cases we
derived a set of high level requirements for OpenU.
Social and Collaborative Requirements
 The system should allow grouping and
participation of users into communities. Each
community should have a virtual presence on the
web;
 It should be possible to set fine grained access
rights to these communities, based on the role of a
user in such a community;
 A user should be able to define an online identity;
 Various social tools should provide the social
cohesion in the system and should allow and
promote collaboration.
Content Management/Publishing Requirements
 Users should be able to collaboratively create and
manage digital content. This content can take
various formats, such as web pages, Wiki pages,
blogs and files;
 It should be possible to set detailed access rights to
this content, controlling who can create, edit,
delete and view the content;
 It should be possible to create web pages for the
created content including possibilities to control
layout and navigation.
Portal Requirements
 Users should have a single access point for their
learning. The system should provide an integrated
and consistent user experience.
 It should be possible to combine functionality in a
flexible manner to construct the different types of
learning networks as each type of learning network
has slightly different requirements.
 The system should be open allowing the seamless
integration of various external services.
LMS Requirements
 Learners should be able to keep track of their study
progress while performing the activities of a
learning design.
 It should be possible to upload, discuss and review
assignments.
TowardsaNewGenerationofLearningManagementSystems
515

 The system must support the creation and delivery
of self-assessments. It should be possible to
incorporate these self-assessments into a course
design.
 It should be possible to monitor study progress by
learners and tutors alike, including self-assessment
outcomes.
 Users should have their own portfolio and it
should be possible to create showcases based of
the data in the portfolio and traces of their
activities within the system.
 It should be possible to implement various
pedagogical approaches and course designs for the
different target audiences.
3 THE LIFERAY PORTAL
When reviewing the global requirements one can
argue that there are several types of systems that
provide a good foundation to achieve the desired
functionality. Content Management Systems (CMS),
Social Networking Systems (SNS), Web Publishing
Systems, Portals and traditional LMSs all provide
parts of the required functionality. However, no
single of these systems will meet all requirements.
Given that no system meets all of our
requirements, the following question arises; what is
the best foundation for developing the next
generation LMS assuming that it makes no sense to
start from scratch? One obvious strategy would be
starting with a traditional LMS, such as Moodle or
Blackboard, as starting point. However, the focus on
the extended classroom and course paradigm
probably would hinder a swift integration of social
and informal learning objectives in OpenU. An
alternative strategy would be using a more generic
system that is free from any pedagogical paradigms.
We argue that the latter approach is the better, more
flexible approach towards the next generation LMS.
In the next section we will discuss the OpenU case
and show how we have configured and extended the
Liferay portal environment (Liferay Portal, 2013) to
meet our requirements.
Liferay provides a major part of the required
functionality by integrating several relevant
subsystems into a single framework. Some
characteristics of Liferay are:
 Liferay is a full JSR-286 (Hepper, 2005) portal. It
will allow the creation of a singular user
experience through portlet technology.
 Liferay has an integrated CMS, complete with
workflow control. It supports various content
types.
 Liferay provides a social, collaborative
environment via a range of social portlets.
 Liferay is based on an extendable open service
architecture that allows bidirectional exchange of
data through well-defined and standardized
interfaces such as web service and JSON/REST.
However, Liferay is lacking specific LMS
functionality which therefore needs to be added. In
the next three sections we will describe how we used
and configured Liferay to meet our requirements and
we will describe what components we added.
3.1 Setup of Liferay as LMS
Figure 1: Basic Liferay functionality.
Figure 1 depicts a high level, layered functional
decomposition of the OpenU LMS, based on an out
of the box Liferay portal. We will explain how we
decided to map the main Liferay concepts to build
our learning networks. The top level represents the
four learning network types that OpenU currently
supports:
 The Personal Learning Network is owned and
largely controlled by the user. In effect this is the
OpenU implementation for a PLE. All users have
their own personal learning network which they
control;
 The Course Learning Network is intended for
delivery of formal, designed courses for the
bachelor and master curricula. Most members will
be course students, but prospective students and
professionals are allowed to participate as well,
although they will not receive support by any
Open University staff.
 The Professional Development Network focuses
on latest trends and developments that are
especially of interest to professionals desiring to
keep up with latest developments in their
profession. Although these networks are intended
to address continuing professional development,
regular university students can participate in them
as well. Both professional and students can receive
credits for their participation.
 Topic/research networks focus on the state of the
art on main research topics. The intended
audiences are researchers, students and
CSEDU2014-6thInternationalConferenceonComputerSupportedEducation
516

professionals alike.
A learning network is implemented via the
Liferay ‘community’ concept. A Liferay community
consists of a number of web pages and each of these
web pages can contain portlets in a specific layout.
A portlet occupies a part of the screen estate and its
functionality can range from something very simple
such as displaying a piece of web content, to a full
fletched interactive group wall application. One can
compare these portlets with Lego bricks. Each type
of Lego brick provides a basic building block, but
only the combination certain bricks result in a
desired model. Similar, only after combining a
particular set of portlets this will result in the desired
learning network type.
Besides determining which portlets should be
combined on a page, we also must define who is
member of a learning network and what rights each
user should have within the network. In Liferay this
authorization is handled through a role based
permission system. A standard role, for example, is
the community member role. But it is possible to
define as many roles as required. The actual
permissions that can be set vary per portlet. Typical
permissions are view and edit permissions.
With these standard Liferay concepts it is
possible to construct a basic learning. However,
configuring each learning network from scratch
would be impractically at best. Liferay provides a
template mechanism for automating this process. We
defined a template for each type of learning network.
The bottom layer of Figure 1 is formed by the
Liferay portlets, the equivalents of the Lego bricks.
This includes content management portlets as well
as social and collaborative portlets. With this
approach and the standard portlets it is only possible
to create very basic learning networks. Advanced
learning networks will require additional LMS
functionality that is not available out of the box.
3.2 Extending Liferay
Liferay provides an open API and SDK for
developing new portlets which can be either build
from scratch or can reuse services from existing
portlets. We extended Liferay with a number of
educational portlets that are required to turn Liferay
into a LMS using both aforementioned portlet
development strategies.
Figure 2 depicts the new high level functional
architecture of Liferay including these extensions.
The following additional portlets have been
developed:
 Assignment portlet: this portlet allows the sharing
and discussion/rating of assignments with tutors;
 Assessment portlet: this portlet allows the creation
of assessments and provisions these assessments to
learners;
 Sequencing portlet: this portlet builds on the
Liferay Wiki and allows the construction of
learning designs that are very similar to IMS-LD
level A. The course author can create and sequence
learning activities. On a role basis, the course
author can determine who should perform which
learning activities;
 Showcase portlet: allows users to share evidence
of their learning progress via a showcase. The
showcase portlet was initially developed as part of
the EU lifelong learning programme project
TRAILER (Brouns, Vogten, Janssen, & Finders,
2013);
 Monitor portlet: portlet that provides tutors with
information about the progress of students in the
learning network;
 Groupwall portlet: a portlet build on top of the
Liferay forum resembling the Facebook wall
feature, but owned by the network community.
With these additional portlets in place, we were
capable of implementing all four learning network
types. However, populating these networks with
actual users is very labor intensive and therefore also
error prone, especially when the number of users
increases. Additional software is required to manage
these subscriptions.
Figure 2: Extended Liferay functionality.
3.3 Subscription Management
Figure 3 depicts the functional architecture with an
additional subscription management layer. Informed
by the ‘separation of concerns’ design principle, this
management layer hides all specific Liferay
configuration details.
This is achieved through the introduction of a
semantic neutral artefact called ‘product’.
The UML class diagram of Figure 4 represents
this product artefact and its relationship with the
Liferay concepts. The standard Liferay artefacts all
have a grey background color. A product represents
TowardsaNewGenerationofLearningManagementSystems
517

Figure 3: Subscription management layer.
Figure 4: UML class diagram for product artefact.
a learning network via its association with a
community. A product also defines which Liferay
roles are associated with a subscription instances. A
subscription instance has a state of either accepted,
expired or cancelled. A user can have a subscription
for a product, which is therefore either accepted,
expired or cancelled.
Setting or changing a subscription will trigger a
recalculation of the associated Liferay role
assignments for the subscription user. This
guarantees that a user always has the correct Liferay
permissions in accordance with the product
subscriptions. A product subscription can be set
from various sources. For example, in the case of
OpenU, a proprietary student administration system
acts as source for all subscriptions of the bachelor
and master students. Product subscriptions can also
be set as a result of a purchase in web shop or
simply via as a result of an open registration. The
latter is typically when accessing the open
educational resources. Regardless what the source of
the subscription is, all role assignments will be
automatically calculated without any further need
for manual intervention and as a consequence the
correct Liferay permissions are granted to the user.
Finally, we have implemented some registration
business rules that allow the definition of products
dependencies. With these business rules it is
possible to define flanking products for a source
product, meaning that a user will be automatically
subscribed to these flanking products whenever a
user is subscribed for the source product.
4 IMPLEMENTATION
To get an impression about the efforts required to
extend Liferay we performed some code metrics
analysis. We have used the CLOC 1.6.0 (CLOC,
2014) for this purpose. Table 1 represents the result
of running the code metrics on the sources of the
standard Liferay 6.0.12 EE product.
Table 1: CLOC statistics for standard Liferay Portal.
Language files code
Java 17581 1881276
HTML 4190 409408
JSP 873 68249
XML 227 44102
CSS 39 5809
Javascript 117 34715
XSD 21 18250
SQL 32 5821
XSLT 5 169
SUM 23085 2467799
Table 2 has the code metrics for the extensions of
the Liferay code that we have developed. This
includes the educational portlets as well as the code
for subscription management layer as described in
the previous sections. It also includes some very
specific code for the integration with the Open
University infrastructure, such as the identity
management services. However, these specific
additions have only a minor impact on the statistics
because they are relative small in size.
Although we didn’t use all available
functionality of the standard Liferay Portal, we used
a very large portion of it. Therefore we may
conclude that roughly 80% of the code needed to
implement OpenU was realized with standard
Liferay code and roughly 20% had to be developed
from scratch.
The source code of all extensions is available as
Open Source through SourceForge at
http://sourceforge.net/projects/openu/
CSEDU2014-6thInternationalConferenceonComputerSupportedEducation
518

Table 2: CLOC statistics for the Liferay Portal extensions.
Language files code
Java 2291 374844
HTML 7 423
JSP 150 10298
JSF 295 24841
XML 306 18739
CSS 69 15541
Javascript 31 4909
XSD 1 230
SQL 23 1016
XSLT 1 77
SUM 3174 450918
5 CONCLUSION
AND DISCUSSION
So, what have we learned? Most LMSs have
dedicated code for teaching and learning, but this is
built on a more generic functionality layer that can
also be found in most portal and CMS software
platforms. In our situation we found that only 20%
of the code is specific to an LMS. Furthermore, the
educational model of a distance teaching university
like the OUN requires a specific set of teaching and
learning modules that are hard to find in a standard
LMS. At the pedagogical level there are many
choices to make. For instance whether or not to
implement informal and social learning as described
in this paper. Many users of traditional LMSs need
to adapt and configure the platform substantively in
order to fit the specific local educational
requirements or when delivering ‘real’ online
education at substantive scale. In this effort they
could be hindered by underlying restrictions in the
models applied by the developers of the LMS. For
instance, most LMSs embrace the course and
extended classroom concepts in their core, which
can be difficult to change towards more generic
concept like learning networks.
The real issue in selecting and implementing an
infrastructure for teaching and learning is to find a
suitable platform that fits into (and connects to) the
existing infrastructure, is secure, flexible, stable and
scalable, provides generic CMS, communication and
collaboration tools out of the box and is extensible
and adaptable. But of course, this is in our opinion
the best way to proceed when one is serious in
implementing ‘real’ online learning.
REFERENCES
Brouns, F., Vogten, H., Janssen, J., & Finders, A. (2013).
E-portfolios in lifelong learning. In F. J. García-
Peñalvo (Ed.), Proceedings of the 2013 Conference on
Technological Ecosystems for Enhancing
Multiculturality (pp. 535–540). Salamanca Spain:
TEEM2013.
CLOC. (2014). CLOC. Website of CLOC. Retrieved
January 09, 2014, from http://cloc.sourceforge.net/
Hepper, S. (2005). JSR 286: Portlet Specification 2.0.
Website Java community process. Retrieved March 10,
2013, from http://jcp.org/en/jsr/detail?id=286.
Hermans, H., Kalz, M., & Koper, R. (2013). Towards a
learner-centred system for adult learning. … -Wide
Information Systems. Retrieved from http://
www.emeraldinsight.com/journals.htm?articleid=1709
8107&show=abstract.
Liferay Portal. (2013). Website of Liferay. Retrieved
March 10, 2013, from http://www.liferay.com/
products/liferay-portal/overview.
Schunk, D. (2012). Learning Theories: An Educational
Perspective. Learning theories: An educational
perspective. (Sixth Edit.). Pearson Education Inc.
TowardsaNewGenerationofLearningManagementSystems
519