Int. J. Learning Technology, Vol. 1, No. 1, 2004 111
Copyright © 2004 Inderscience Enterprises Ltd.
Educational modelling language and learning design:
new opportunities for instructional reusability and
Hans Hummel, Jocelyn Manderveld,
Colin Tattersall and Rob Koper
Open University of the Netherlands,
Educational Technology Expertise Centre, PO Box 2960,
6401 DL Heerlen, The Netherlands
Fax: +31 45 5762 2907 E-mail: email@example.com
Abstract: Learning technologies offer new opportunities to meet the rapidly
growing demand for new, constructivist ways of learning (such as competency-
based, collaborative or adaptive learning). They have the potential to act as
catalysts for more effective exchange and reuse of learning objects to enable
personalised learning. This article examines the extent to which current
learning technology specifications contribute to educational change – to actual
sharing and reuse in educational practice. Furthermore, the article describes the
need for an Educational Modelling Language centred on learning activities to
give instructional meaning to learning objects.
To date, specifications for learning objects have primarily been designed to
ensure interoperability at a rather low infrastructural level (e.g., test items,
meta-data), focusing on technology issues and reuse of learning objects.
We argue that more widespread adoption of e-learning specifications and
standards calls for a pedagogical framework at a higher infrastructural level
(e.g., a complete course), focusing on the instructional value and reuse of
learning activities. Such a framework is offered by the new Learning Design
(LD) specification. LD enables the description of both learning content and
processes from a variety of pedagogical perspectives, both objectivist and
Keywords: learning activities; learning objects; personalisation; educational
modelling language (EML); learning design (LD).
Reference to this paper should be made as follows: Hummel, H.,
Manderveld, J., Tattersall, C. and Koper, R. (2004) ‘Educational modelling
language and learning design: new opportunities for instructional reusability
and personalised learning’, Int. J. Learning Technology, Vol. 1, No. 1,
Biographical notes: The authors work at the Educational Technology
Expertise Centre of the Open University of the Netherlands (OUNL) and were
involved in the design and development of EML, the Educational Modelling
Language (1998-2002), which formed the basis for the recently adopted IMS
Learning Design specification.
Dr. H.G.K. Hummel (1960) holds degrees in pedagogy and educational
psychology, with minors in Informatics and Orthodidactics (1985, University
of Leiden, Netherlands). He has coordinated innovative projects applying ICT
in primary and vocational education for both a research institute and a
112 H. Hummel, J. Manderveld, C. Tattersall and R. Koper
publishing company. Hans has worked at OUNL since 1987, co-developing
dozens of distance courses, and leading the development of interactive
computer programs in a variety of domains. He was involved in the
development of EML, in learning technologies standardisation workgroups
(IMS/LD, Prometeus/Pedagogies) and was OUNL-spokesman for corporate
communications related to innovation.
Dr. J.M. Manderveld (1973) holds a degree in educational psychology
(1997, University of Tilburg, Netherlands). She managed a number of
educational projects for the Dutch Railways before joining the OUNL in 1998,
where she has been involved in designing and developing flexible and rich
learning environments, and in the development of EML. Jocelyn was OUNL
project manager for the standardisation of EML and has participated in
standardisation workgroups including IMS/LD and CEN/ISSS.
Dr. C. Tattersall (1965) gained a degree in computational science
(1986, University of Leeds, United Kingdom) before conducting research into
text generation for intelligent help systems at Leeds University’s Computer
Based Learning Unit, resulting in his PhD (1990). He worked in both the
telecommunications and software industries before joining OUNL in late 2002,
where his work includes harmonising IMS Learning Design with other IMS
Prof. Dr. E.J.R. Koper (1957) holds a degree in educational psychology
(1986, University of Tilburg) and a PhD in educational technology. He worked
as a teacher in higher education and was director of a teacher training company,
joining OUNL in 1987. Rob became head of ICT application development and
now holds a chair in Educational Technology (1998). He was the programme
manager for R&D into Learning Technologies (1998-2002) that resulted in
EML, IMS Learning Design, and tools for authoring and publishing, and
currently leads the new program on Learning Networks (2003-2008). Rob
chairs and advises several (inter)national ICT initiatives, and is an IMS
Technical Board Member and is a Prometeus Steering Committee member.
A growing body of professional educators feel the urge to improve the effectiveness of
educational processes. Moreover, there is a pressing need for personalised and flexible
learning without constraints of time and place. These demands are caused by societal
trends such as life-long learning, the diminishing gap between working and learning, the
increasing globalisation of education, and most of all by the possibilities of new
technologies [1,2]. New technologies provide the means to integrate teaching and
learning into every aspect of each person’s life. There is a demand for new ways of
learning, often based on constructivist principles [3-5]. Examples include collaborative
learning [6,7] where discussion plays an important role in learning, competence-based
learning [8,9] and problem-based learning [10,11], where knowledge is constructed by
individual learners in solving real problems in realistic situations.
A major problem in realising new ways of learning is that educational changes tend to
take place in isolation and are not always documented. Individual teachers and institutes
choose their own issues and products, such as assessment forms, innovative textbooks,
interactive media and Learning Management Systems (LMS). Teachers strive for
Educational modelling language and learning design 113
maximum flexibility and freedom in customising their learning material (or learning
content, as learning technologists prefer to call it) and learning processes. Teachers (and
students) often have specific pedagogical demands which require customised approaches,
as opposed to off-the-shelf solutions offered by external suppliers [12,13].
Customisation, then, seems a desirable element of innovation. However, the lack of
transparency in professional practice and the lack of collaboration between teachers and
between institutions hinder structural innovation. A good starting point would be to look
for collaborative agreements between learning content specialists, for collaborative
(specialised) development or reuse of learning content. More agreement on these issues
would give the possibilities of reuse, interoperability and personalisation more depth and
Although technological standards impose demands on education, they can also be
very supportive in realising new ways of learning. In terms of standardisation we can talk
about general requirements and features of learning content and learning processes
without having to restrict in any way specific pedagogical views (of individual teachers
or institutes) on what learning should be about. Standardisation, then, does not restrict but
facilitates customisation. Agreements on these requirements and features are recorded in
specifications and taken up in the standardisation process. The field of learning
technologies covers the development and recording of these specifications and standards.
Having introduced the potentially fertile relationship between new ways of learning
and learning technologies, the central question to be answered in the remainder of this
article is: To what extent do current learning technologies specifications already
contribute to new ways of learning in educational practice, and what else is required?
In order to answer the question we first describe learning technologies specifications
and the reasons for their development. We focus on (re-usable) learning objects, and
possibilities for building learning objects to support personalised learning. We evaluate
the extent to which current specifications support the building of personalised learning.
Following on from this evaluation we argue the need for a pedagogical framework which
relates learning objects to an instructional context (e.g. a complete course). Our
conclusion contains conditions and expectations for the effective, worldwide uptake of
such a pedagogical framework and for future implementations (for example in LMS).
2 Learning technology specifications
Learning technology (LT) is an area with many names but with few definitions .
Oliver and Bradley  define LT as the use of technology to support innovations of
teaching and learning. We more narrowly define LT as ‘specifications of methods and
techniques which support the realisation of e-learning’. Examples of specifications are:
• formats and rules for the design of a didactic approach
• competency profiles and assessment models (e.g., portfolios)
• personalisation models (e.g., flexible study arrangements)
• architectures and user interfaces.
The essential feature of specifications such as these is that they are independent of
hardware and software.
114 H. Hummel, J. Manderveld, C. Tattersall and R. Koper
The design and development of learning technology specifications is a global issue,
and there are several initiatives underway:
• Several industrial consortia are developing learning technology specifications. IMS
(Instructional Management Systems ) is probably the best known, a consortium
of companies, universities and institutes. IMS is ‘open’, but membership requires a
(substantial) annual subscription. Membership includes BlackBoard, WebCT, IBM,
OUNL, and others.
• Expert-based initiatives include the IEEE LTSC (Learning Technology Standards
Committee ) and ADL (Advanced Distributed Learning ). Another good
example of initiatives based on the consensus of experts from universities and
companies, is Prometeus (PROmoting Multimedia access to Education and Training
in the EUropean Society  which is supported by the European Commission.
• Learning technology specifications are also developed at a national or regional level.
In some countries (such as France or The Netherlands) standards are referred to as
‘norms’. In the United States, ANSI  is producing learning technology
specifications. At the European level CEN  does the same. Specifications from
several countries and expert bodies can eventually become ISO (International
Organisation for Standardisation) standards. It can take as long as five to ten years
for specifications to gain worldwide acceptance, but once they become ISO standards
they are ensured a long life. One of their Joint Technical Committees (JT1),
subcommittee 36 (SC36), is currently responsible for standards on learning
technology, but until now no official learning technology standard has been
The form and structure of a specification varies considerably. Some specifications (such
as those of IMS) use XML (eXtensible Markup Language), which is an ‘open’ modelling
language and ensures independency of media and interoperability . The essence of a
specification however is provided by the information model that can be uniformly
represented in UML schema. It is important that the specification is ‘open’, as this will
enable other institutions to reuse and apply the material. Specifications that are developed
within a ‘closed’ community tend be used only within that community or company.
These developments tend to progress more rapidly, but at the same time they run the risk
of addressing only one specific situation.
LT specifications have to be recorded in a clear, uniform, abstract and formal way.
This is not only important for reuse and interoperability, but also for recognition by
standardisation bodies. Standardisation leads to a more effective exchange and (re)use of
learning objects , and we feel that widely adopted, open and accredited standards are
a necessary requirement for revolutionary changes to occur in education. This has been
demonstrated in other domains – in the case of electricity, it was the standardisation of
voltage and plugs, for railroads, it was the standard gauge of the tracks and for the
internet, it has been the common standards of TCP/IP, HTTP, and HTML .
We believe e-learning standards will offer a common language for sharing ideas
without restricting customisation. Interested parties will be able to use the standards,
exploiting their in-built flexibility to implement and adapt them to their own
environment. Structural innovation in e-learning will be hampered until such standards
are in place.
Educational modelling language and learning design 115
3 Reusable learning objects
To date, the focus of LT has been on developing specifications for learning objects. The
learning objects movement has grown over the past few years, and is becoming
increasingly mainstream. Several specifications and a standard for learning objects exist,
and there is much interest in meta-data and packaging. Thinking in terms of learning
objects has been triggered by the object-orientation approach in engineering, which
values the creation of components (called ‘objects’) for subsequent reuse on a variety of
platforms and in a variety of contexts .
3.1 Definition of learning objects
What is a learning object? Wiley [26, p.6] simply defines a learning object as ‘any digital
resource that can be reused to support learning’. This definition is more specific than the
strict LTSC  definition of a learning object that also includes non-digital resources
such as persons, ideas, … at any time or place. Note that the IEEE/LTSC was founded in
1996 to develop LT standards, primarily to facilitate the widespread adoption of this
(learning) object-orientation. We further refine the definition of a learning object as ‘any
digital, reproducible and addressable resource used to perform learning or support
activities, made available for others to use.’
This definition excludes many things, e.g. non-digital materials, non-reproducible
unique exemplars, non-addressable resources (i.e. when not connected with a URL and
metadata for access). It also excludes courses (being a composite of learning objects and
learning activities) and ‘persons’, ‘activities’ and ‘services’. Reuse is the central element
of learning objects, as generativity, adaptivity (e.g., personalisation), learning activity and
other activities are all facilitated by the properties of reuse . However, reuse is also a
weakly defined concept, but can be narrowed down by following our definition of a
3.2 Instructional design literature and learning objects
The majority of literature trying to explore the instructional value of learning objects was
written by M. David Merrill and his team at Utah University . Merrill departs from
Instructional Transaction Theory (ITT), and distinguishes four types of learning objects:
entities (objects in the world like devices, persons, places); properties (attributes of
entities); activities (actions the learner takes on objects in the world); and processes
(events that change properties, triggered by activities). Merrill’s more recent studies on
learning objects leans heavily on Gagné’s Conditions of Learning. The assumption of all
instructional design theories and models such as ITT, Conditions of Learning, and others
such as the Four Components / Instructional Design (4C/ID) Model of Van Merriënboer
 is to some degree an objectivist one, and can be associated with the metaphor of the
“Mind as a Computer”. According to this view, when executing (complex) tasks the
human mind manipulates information in the same algorithmic way a computer
manipulates digital data. Both learning content, and (the sequence and combination of)
learning processes can be designed in advance for all students, and solutions or answers
are either right or wrong, as in, for example, a multiple-choice question.
116 H. Hummel, J. Manderveld, C. Tattersall and R. Koper
In sharp contrast to these cognitive information processing approaches are alternative
perspectives which stress the flexible dissemination and use of content, such as Cognitive
Flexibility Theory , and the personalisation and contextualisation of learning
processes, such as Situated Cognition . These approaches advocate that instructional
design cannot be algorithmic and should take into account multiple perspectives on
content and not rely on a single schema . Moreover, knowledge is continuously under
construction and evolving for every student, activity and situation. Rather than acquiring
knowledge as self-contained, abstract entities, the emphasis is on acquiring useful
knowledge through enculturation (understanding how knowledge is used by
3.3 Learning objects versus learning activities
Interoperability has been the dominating element in specifying learning objects, mainly
because vested interests in commercial applications are huge. Learning objects are likely
to become the instructional technology and the world will be flooded with learning
object-based tools. Vendors therefore have stressed the importance of recognition,
adoption, and the potential for future support.
However, technical standards and venture capital are important but not enough to
promote learning. In order to promote learning, technology-enabled learning should also
be guided by instructional principles. Without attention to the process of instruction,
interoperability and reusability of learning objects will not materialise. Educational
designers must first establish how individual students could be studying most effectively.
There is a growing feeling of uneasiness, a feeling that the primacy of reusable learning
objects is leading to e-learning as page-turning, that the people-to-content model leads to
“static, fossilized, dead [content], low learner motivation [and] engagement, impersonal
[and] isolating environments” . Software vendors and standards bodies offer products
that are presented as ‘instructional theory neutral’ (e.g. in the information model of the
IMS Content Packaging specification). However, we feel most of the commercially
available LMS nowadays reflect old ways of learning embedded in objectivist views on
learning. In the worst case their possibilities are limited to ‘clip-art slide shows’ on the
web, not allowing for any active role of the learner.
The recently approved Learning Design (LD) Final Specification  provides a
counter to the trend towards designing for lone learners reading from screens. It guides
staff and educational developers to start not with content, but with learning activities and
the achievement of learning objectives. It recognises that learning can happen without
learning objects, that learning is different from content consumption, is highly personal
and that learning comes from being active. It recognises, too, that learning happens when
learners cooperate to solve problems in social and work situations. In all this, it stresses
that we must focus on the learning in e-learning, and it is this focus which makes it
important for staff and educational developers. Before describing the Educational
Modelling Language, developed by the Open University of the Netherlands and the basis
for LD, we will first turn to the issue of adoptability and personalisation.
Educational modelling language and learning design 117
4 Personalising learning
Building individualised learning activities to support personalised instruction in an
adaptive environment is a big challenge for the future of e-learning. The web offer the
perfect technology and environment for individualised learning, since learners can be
uniquely identified, content can be specifically personalised, and learner progress can be
monitored, supported and assessed. The greatest benefit of learning personalisation is the
system’s ability to make complex instruction and learning easier. According to Martinez
[34, p.156] this is achieved by
“… presenting only the specific information that a particular learner wants or
needs in the appropriate manner and at the appropriate time. Each time you
personalise, you learn and store a little more about a learner’s unique set of
In order to further clarify the concept of personalisation, Martinez  distinguishes five
levels of increasing sophistication:
1 name recognition
2 self-described personalisation (study preferences based on e.g. a pre-quiz)
3 segmented personalisation (different sets of content for learning groups)
4 cognitive-based personalisation (e.g. text/audio or linear/hypertext presentation of
5 whole-person personalisation (making predictions about the delivery of the content).
Modelling the educational process will bring new opportunities and benefits for
personalised learning. Current developments with Life Long Learning are leading to an
increase in heterogeneity within the total population of learners, who demand more
adaptive and customised learning environments. Example 1 describes a course (the unit
of study) on ‘Learning to listen to jazz’ that was created using Educational Modelling
Language (EML). It demonstrates personalising at level (3) for both the learning content
and the learning processes. Personalisation will have even more potential at the
curriculum level (a collection of units of study), where both teachers and students are
able to define their own study arrangements according to their own prior knowledge,
preferences or intentions.
4.1 Example 1: personalisation in the jazz course
The course “Learning to listen to jazz” was designed and modelled in EML as a
self-study course that can be taken individually. The objective is to learn how to
distinguish between rhythm and melody when listening to jazz, and is constructed so as
to give students individualised learning pathways. Based on an intake assessment, the
student will be given advice about the learning pathway, depending on their previous
knowledge and learning style (previous knowledge test and study approach test). The
teaching method chosen is based on self-assessment. This means that students judge for
themselves their grasp of the subject matter and choose whether or not to follow the
advice on which learning pathway to follow. This design means that students not only
learn about listening to jazz, they also get an insight into the way they learn and whether
118 H. Hummel, J. Manderveld, C. Tattersall and R. Koper
a particular type of course material suits the needs of the student. The scenario of the
whole course is summarised in Figure 1.
Figure 1 Didactical scenario of the jazz course
After choosing the most suitable route in the ‘orientation’ on the basis of the test results,
either the historical or the thematic route can be followed. Halfway through the course,
students are once again offered the option to change the route for the rest of the course.
Using the thematic route, they can jump from style to style. In the historical route, the
sequence presented is recommended to be followed. After finishing a whole route, a final
reflection is made available to the student. The whole course was tagged in EML, using
Educational modelling language and learning design 119
Framemaker+SGML in combination with the EML DTD. To give you an idea of the size
of the source XML file, the printed document runs to 122 pages.
4.2 Further requirements
We now return to the central question raised in this article: To what extent do current LT
specifications already contribute to new ways of learning in educational practice, and
what will be required further? Most design efforts for learning objects and learning
technology specifications have avoided critical instructional design issues. As a result the
need for a pedagogical framework to achieve instructional objectives has been ignored. In
addition to the need for personalised learning content (illustrated in Example 1), there is
also a need to define how various learning processes relate to one another, and how these
can be used together according to particular didactic approaches for complete learning
The current specifications for learning objects address various components of a
learning task, but none is fit to model ‘whole-tasks’ (or units of learning). Relatively few
learning objects can be described, being mainly restricted to samples of learning material
and test items. Due to their origins, most object-oriented learning systems and learning
technology specifications have focused on interoperability issues, such as attributes, data
interchange protocol, tool agent communication, meta data standards and the technical
architecture of the system  and to a lesser extent on reusability issues. As a result of
this focus on technical and technological issues, giving the various learning objects
instructional meaning has been neglected. For the future development of educational
technology and technology-enabled learning, it is now crucial to first give some thought
to using learning objects for new ways of learning before implementing this technology
on a large scale (e.g., in LMSs). As long as learning objects lack instructional meaning,
we will not be able to use them effectively, which will hinder the structural innovation of
education. Educational Modelling Language and Learning Design provide an
instructional framework to model both learning content and processes (e.g., to personalise
learning objects), and also describe the didactic / instructional relations between various
learning objects. This framework was researched, designed, developed and implemented
by the Open University of the Netherlands.
5 Educational Modelling Language and Learning Design
Educational Modelling Language (EML) is a notational system developed by the Open
University of the Netherlands (OUNL) in the late 1990s and intended to describe a wide
variety of instructional models (for example, Competency Based Learning, Problem
Based Learning). At the heart of the specification is a model which underlies many
different behaviourist, cognitive, and (social) constructivist approaches to learning and
instruction. The model revolves around describing ‘units of learning’, atomic or
elemental units providing learning events for learners, satisfying one or more interrelated
Once described in EML, these models are able to be interpreted (or ‘played’) by an
EML-aware software component (or ‘player’), analogous to the way HTML is interpreted
by a browser. A prototype EML player has been used at OUNL and partners throughout
the world for the past couple of years and a production-quality player is currently
120 H. Hummel, J. Manderveld, C. Tattersall and R. Koper
undergoing final field trials. So far, thousands of study hours of learning material in a
variety of instructional models has been created and is still ‘up-and-running’.
EML has about 400 different elements and is implemented in XML. The highest
level, a unit-of-learning, could be a whole course, a module within a course, and so on.
There is no predetermined notion of how large a unit-of-learning should be. This is a
powerful concept, since every unit-of-learning can consist of smaller units-of-learning,
enabling complex structures. Such a unit-of-learning is defined as ‘a systematic
aggregation of learning activities that are necessary to reach certain learning objectives,
including the environments and resources that are needed for executing those activities.’
The environmental resources can be used in several learning activities and units-of-
learning . The EML unit of learning model is presented in Figure 2.
Figure 2 Semantic information model of a unit of learning expressed in UML
Educational modelling language and learning design 121
In its approach to modelling both learning content and learning processes, EML
innovates in the world of learning technologies. It can be used to create adaptable and
flexible personalised learning experiences, and is able to support all five levels of
personalisation as described by Martinez  and a wide variety of didactic approaches.
EML contains a pedagogical metamodel  making it possible to design education from
a variety of different pedagogical approaches: from more constructivist approaches to
more objectivist views on learning.
EML was selected as the basis for IMS Learning Design 1.0, which was approved as
an official IMS Final Specification on 10 February 2003. As a result, EML is no longer
maintained or updated and OUNL’s attention is now focused on IMS LD (a description
of the differences between EML and LD goes beyond the scope of this article). EML and
LD share the same philosophy and aim: In a unit of learning, people act in different roles
in the teaching-learning process, working toward certain outcomes by performing
learning and/or support activities within an environment, consisting of learning objects
and services to be used during the performance of the activities. The approach separates
learning objects and services (modelled outside LD) from the educational method and
learning activities used in the unit of learning (modelled inside LD). These physical
entities represent the actual content used within a unit of learning. These can be files or
objects (tool, knowledge and test objects). Figure 1 describes the unit of learning model.
The unit of learning consists of two packages: learning design and physical entities.
Learning design basically describes the relationship between roles, activity-structures and
environments. The core concept is that learners perform learning activities in an
environment. This is elaborated in various ways:
• learner and staff are organised in roles which can be nested
• activities are organised in activity-structures which can be nested
• environments consist of learning objects and services
• performing an activity creates an outcome, which can be stored in the environment.
The outcome of an activity triggers a notification, which has consequences for the
pedagogical design of a unit of learning
• a unit of learning is designed towards certain learning objectives and prerequisites.
The flow of activities which happens during the learning process is modelled as a
theatrical play consisting of a series of acts
• the flow of activities represented can be influenced by notifications and conditions.
Two examples of the use EML/LD in relation to personalisation are presented in this
article. Example 2 gives an impression of the LD specification on the level of complete
units-of-learning or ‘whole tasks’ (e.g., a course on LT), and its possibilities for
personalised learning processes.
5.1 Example 2: complete ‘unit of learning’ modelled in IMS LD
A complete, though very simple course on LT, modelled in IMS LD, is now given to
emphasise the structure and relations between components and to illustrate the potential
to personalise learning processes. To achieve the learning objective, this course consists
of three learning activities. The student can choose between studying with or without
122 H. Hummel, J. Manderveld, C. Tattersall and R. Koper
examples. So every student gets different sets of content. Components in the (learning)
‘environment’ are required for execution of the second activity. This article is the only
‘knowledge object’ that can be studied within this environment.
Figure 3 Example 2: complete ‘unit of learning’ on learning technologies
<imsld: learning-design identifier=”Course-LT-Specs”>
<imsld: learner identifier=”Student”>
<imsld: locpers-property identifier=”P-availability-examples”>
<imsld: datatype datatype:”Boolean”/>
<imsld: learning activity identifier=”Preparation”>
<imsld: item identifierref=”R-preparation” identifier=”I-preparation”/>
<imsld: property-ref ref=”
<imsld: learning activity identifier=”Assignment-1”>
<imsld: item identifierref=”R-assignment-1” identifier=”I-assignment-1”/>
<imsld: learning activity identifier=”Assignment-2”>
<imsld: item identifierref=”R-assignment-2” identifier=”I-assignment-2”/>
<imsld: activity-structure identifier=”AS-assignments” structure type=”sequence”>
<imsld: learning activity-ref ref=”Preparation”/>
<imsld: learning activity-ref ref=”Assignment -1”/>
<imsld: learning activity-ref ref=”Assignment -2”/>
Educational modelling language and learning design 123
<imsld: environment identifier= "E-study-resources »>
Study resources</imsld: title>
<imsld: learning-object identifier= “LO-article”>
<imsld: item identifierref=”R-article” identifier= “I-article”/>
Course on LT</imsld:title>
Act course on LT</imsld:title>
Role part learner</imsld:title>
<imsld: role-re ref=”
<imsld: activity-structure-ref ref=”AS-assignments”/>
with examples </imsld: property-value>
<imsld: class class=”C-examples”/>
<resource identifier="R-article" type= "imsldcontent"/>
< ! – - the resource R-article contains this article. This article includes examples, such as these. In the first activity
students can decide if they want to study with or without examples. The examples are bracketed by a DIV-element
in XHTML and the DIV-element has the class-attribute “C-examples”. In the play’s conditions the class attribute is
set to either hide or show the examples.- – >
<resource identifier="R-Preparation" type= "imsldcontent"/>
< ! – - the resource R-article contains the description of the activity “Preparation”. – - >
<resource identifier="R-Assignment-1" type= "imsldcontent"/>
< ! – - the resource R-article contains the description of the activity “Assignment 1”. – - >
<resource identifier="R-Assignment-2" type= "imsldcontent"/>
< ! – - the resource R-article contains the description of the activity “Assignment 2”. – - >
6 Conclusion: conditions and expectations
Learning technology specifications have been primarily developed to ensure the
interoperability of learning objects at rather low levels of granularity (like test items),
124 H. Hummel, J. Manderveld, C. Tattersall and R. Koper
focusing on the technological value and use of learning objects. What is also needed to
contribute to new ways of learning in educational practice is a pedagogical framework
that structures the relations between various learning objects, and redirects attention to
the instructional value and use of learning objects. The real challenge is to reach
agreement on what constitutes a meaningful combination of learning objects in real
learning activities, to be aggregated at higher levels of granularity (like complete
courses). Such a pedagogical framework preferably is general enough to support both old
(objectivist) and new (constructivist) ways of learning to ensure that specifications will
be widely (re)used. LT specifications that have been developed so far are not fit for being
used by teachers and designers to innovate education, often because the LMS they are
implemented in still reflect an objectivist view on learning that does not allow for new
ways of learning.
EML and LD are LT specifications that describe both learning content and processes
within ‘units of learning’, or whole tasks (like a course). They contain a pedagogical
meta-model (or framework) that supports a large variety of didactical approaches (both
objectivist and constructivist). We included two examples of personalisation in courses
modelled in EML and LD in this article, and argued why such a pedagogical framework
could support new ways of learning.
Since IMS LD separates learning approaches and activities from the learning objects
and services used, new opportunities for reuse are raised:
• Individual learning designs can be applied across different domains. Each time,
different content is coupled to the same activities of the learning design.
• Learning objects can be used in different educational models. Each time, different
activities are associated with the same content.
EML has already been implemented within and outside the Open University of the
Netherlands and is ‘up and running’ in a large variety of higher education courses and
training programs. A commercial version of the Edubox player is under construction.
We are currently faced with two major concerns or conditions for further uptake.
Since it is only a matter of weeks since the IMS LD specification was approved, no IMS
LD player yet exists. As a result, an important part of the benefit of IMS LD cannot yet
be reaped – it is not yet possible to author an XML file coupling activities to resources
and services as described by the specification and have this interpreted in an IMS LD-
aware software environment for learners. However, we are confident that this situation
will soon change as Learning Management System vendors familiarise themselves with
the opportunities afforded by the specification. We are also exploring ways in which the
available EML players might be ‘upgraded’ to become IMS LD aware.
Another major concern in the implementation will be the teachers’ perspective and
the uptake of LD in educational practice. Teachers will only use LD if it allows them to
teach in the way they want, and be rewarded for applying it. Moreover, use requires good
tools, such a user-friendly yet flexible authoring environment, and a powerful and reliable
player. As a result of these needs, a large group of institutes and companies have started
to work together as the so called ‘Valkenburg group’ to develop a user-friendly authoring
system. This system will help to realise EML and LD’s potential while maintaining
possibilities for teachers to use different pedagogical models when designing their units
Educational modelling language and learning design 125
Nonetheless we feel staff and educational developers can already benefit from the
philosophy of IMS LD by focusing on learners’ activities and objectives, and designing
e-learning environments with this philosophy in mind. The vision towards which we are
working sees educational best practices available as reusable learning designs, able to be
downloaded and customised by staff and educational developers, coupled to (reusable)
learning objects and interpreted by IMS LD aware environments, giving learners the
stimulating, active, challenging and exciting experiences they deserve.
1 Fernandez-Manjon, B. and Sancho, P. (2002) ‘Creating cost-effective adaptive educational
hypermedia based on markup technologies and e-learning standards’, Interactive Educational
Multimedia, Vol. 4, April, pp.1–11.
2 Retalis, S. and Avgeriou, P. (2002) ‘Modelling web-based instructional systems’, Journal of
Information Technology Education, Vol. 1, No. 1, pp.25–42.
3 Duffy, T.M. and Cunningham, D.J. (1996) ‘Constructivism: implications for the design and
delivery of instruction’, in D.H. Jonassen (Ed.) Handbook of Research for Educational
Communication and Technology, Simon & Schuster/MacMillan, New York, pp.170–198.
4 Wood, D. and Wood, H. (1996) ‘Vygotsky, tutoring and learning’, Oxford Review of
Education, Vol. 22, No. 1, pp.5–10.
5 Brown, J.S., Collins, A.L. and Duguid, P. (1989) ‘Situated cognition and the culture of
learning’, Educational Researcher, Vol. 18, pp.32–42.
6 Dillenbourg, P. and Schneider, D. (1995) ‘Mediating the mechanisms which make
collaborative learning sometimes effective’, International Journal of Educational
Telecommunications, Vol. 1, pp.131–146.
7 Kim, A.J. (2000) Community Building on the Web: Secret Strategies for Successful Online
Communities, Peachpit Press/Addison Wesley Longman, Berkeley, CA.
8 Kirschner, P.A., Vilsteren, P.P.M. van, Hummel, H.G.K. and Wigman, M.C.S. (1997) ‘The
design of a study environment for acquiring academic and professional competence’, Studies
in Higher Education, Vol. 22, No. 2, pp.151–171.
9 Schlusmans, K., Slotman, R., Nagtegaal, C. and Kinkhorst, G. (1999) Competentiegerichte
Leeromgevingen [Competency-based learning environments], Lemma, Utrecht,
10 Savery, J. and Duffy, T.M. (1995) ‘Problem based learning: an instructional model and its
constructivist framework’, Educational Technology, Vol. 15, No. 5pp.31–38.
11 Jonassen, D.H. (1997) ‘instructional design models for well-structured and ill-structured
problem-solving learning outcomes’, Educational Technology: Research and Development,
Vol. 45, No. 1, pp.65–95.
12 Rogers, E. (1995) Diffusion of Innovations, Free Press, New York.
13 Surry, D.W. and Farquhar, J.D. (1997) ‘Diffusion theory and instructional technology’,
Journal of Instructional Science and Technology, Vol. 2, No. 1.
14 Oliver, M. (2000) ‘An introduction to the evaluation of learning technology’, Educational
Technology & Society, Vol. 3, No. 4.
15 Oliver, M. and Bradley, C. (1999) ‘Examples of best practice in the use of multimedia in
higher education, EXE Report No. 2, University of North London, London.
16 Available from: www.imsproject.org
17 Available from: http://ltsc.ieee.org
18 Available from: www.adlnet.org
19 Available from: www.prometeus.org
126 H. Hummel, J. Manderveld, C. Tattersall and R. Koper
20 Available from: www.ansi.org
21 Available from: www.cenorm.be
22 Bray, T., Paoli, J. and Sperberg-McQueen, C. (1998) Extensible Markup Language (XML)
1.0. World Wide Web consortium (W3C) Available from: http://www.w3c.org/TR/REC-
23 Koper, E.J.R. (2001) ‘XML-documentschema’s en e-learning (XML-documentschema’s and
e-learning)’, Informatie, Vol. 6, pp.30–33.
24 Hodgins, H.W. (2002) ‘The future of learning objects’, in D.A. Wiley (Ed.) The Instructional
Use of Learning Objects, Agency for Instructional Technology, Bloomington, pp.281–298.
25 Dahl, O.J. and Nygaard, K. (1996) ‘SIMULA-an Algol based simulation language’,
Communications of the ALM, Vol. 9, No. 9, pp.671–678.
26 Wiley, D.A. (2002) ‘Connecting learning objects to instructional design theory: a definition, a
metaphor, and a taxonomy’, in D.A. Wiley (Ed.) The Instructional Use of Learning Objects,
Agency for Instructional Technology, Bloomington, pp.3–25.
27 LTSC (2000) Learning Technology Standards Committee, Available from: http://ltsc.ieee.org
28 Available from: http://www.coe.usu.edu/it/id2/ddctoc.htm
29 Van Merriënboer, J.J.G. (1997) Training Complex Cognitive Skills: a Four Component
Instructional Design Model for Technical Training, Educational Technology Publications,
Englewood Cliffs, NJ.
30 Spiro, R.J., Feltovich, P.J., Jacobson, M.J. and Coulson, R.L. (1987) ‘Cognitive flexibility,
constructivism, and hypertext: random access instruction for advanced knowledge acquisition
in ill-structured domains’, in T.M. Duffy and D.H. Jonassen (Eds.) Constructivism and
the Technology of Instruction: a Conversation, Lawrence Erlbaum Associates, Englewood
31 Jonassen, D.H. (1991) ‘Objectivism versus constructivism: do we need a new philosophical
paradigm?’, Educational Technology Research and Development, Vol. 39, No. 3, pp.5–14.
32 Stacey, P. (2003) ‘People to people, not just people to content’, Presentation at the IMS Open
Technical Forum, Vancouver, 20 February, Available from:
33 IMS (2003) Learning Design v 1.0 (Final Specification), Available from:
34 Martinez, M. (2002) ‘Designing learning objects to personalize learning’, in D.A. Wiley (Ed.)
The Instructional Use of Learning Objects, Agency for Instructional Technology,
35 Singh, H. (2000) Achieving Interoperability in e-Learning, Available from:
36 Koper, E.J.R. (2001) Modelling Units of Study from a Pedagogical Perspective: the
Pedagogical Metamodel Behind EML, Available from: http://eml.ou.nl/introduction/docs/ped-