Specification of Visual Instructional Design Languages Dedicated

to Learning Management Systems

Aymen Abedmouleh, Pierre Laforcade and Lahcen Oubahssi

LIUM, Avenue Olivier Messiaen, 72085, Le Mans Cedex 9, France

Keywords: Visual Instructional Design Language, Domain Specific Modeling, Learning Management System.

Abstract: Despite of the growing development of learning technologies into education, designing learning scenarios

and exploiting them for setting up a learning situation is still a complex task. Visual Instructional Design

Languages (VIDL) and their dedicated graphical editors have been identified as important conceptual tools

for achieving more creative design solutions within a design process. In this article we propose the

application of Domain-Specific Modeling tools for specifying and developing VIDLs and editors dedicated

to specific Learning Management Systems. An experimentation concerning the Moodle LMS is discussed.

1 INTRODUCTION

The Technology Enhanced Learning (TEL) research

domain has provided many solutions to support

distant instructional design: Educational Modeling

Languages (EML) facilitating the specification of

learning situations as formal learning scenarios for

delivering and exchanges purposes (Koper and

Manderveld, 2004), Visual Instructional Design

Languages (VIDL) (Botturi and Stubbs, 2007)

focusing on the support of imagination, thinking,

communication for practitioners communities,

Learning Management Systems (LMS) providing an

operational TEL environment for delivering online

learning situations.

Despite of these tools and facilities, the learning

design is still a complex task. The development of

LMS systems has not decreased the complexity of

design and learning processes in these systems

(Martinez-Ortiz et al., 2009). Standards de facto like

IMS-LD (Koper, 2006) have not succeeded in being

integrated to the existent LMSs widely spread

(Burgos et al. 2007). The VIDLs do not really allow

to exploit the scenarios they ease to specify for

automatizing the delivery and setting up of LMSs

(Laforcade, 2010). Nowadays, teacher-designers

within academic contexts are still designing their

learning situations by directly using the parameters'

forms-oriented screens of LMSs they have at their

disposal.

In our research works we are interested in

helping such practitioners to design distant or mixed

(in relation to their face-to-face sessions) learning

situations. We aim at providing them with specific

VIDLs related both to the LMS they usually use and

to their practices and needs. Such VIDLs have to

meet the VIDLs added-values (visual notation

improving the instructional design reflexion), the

EMLs ones (formalization and binding), and those

from LMSs (configurations and delivering).

Our proposition is based on the idea that every

LMS embeds an implicit instructional design

language. Our approach originally proposes a two

steps approach: (1) identifying and formalizing this

LMS language, and (2) exploiting this language for

the specification of VIDLs and graphical editors.

This LMS-centered approach is strongly relying on a

Model Driven Engineering (MDE) and Domain

Specific Modeling context (DSM) framework. This

article mainly focuses on the second step. We then

assume that the LMS language has been identified

and formalized. Nevertheless, readers can also

consider our proposition within a wide scope dealing

with the use of a DSM approach for specifying

visual languages on top of an existent XSD file.

2 BACKGROUND

2.1 Instructional Design

Instructional Design is the entire process of analysis

of learning needs and goals and the development of

a delivery system to meet those needs (Berger and

199

Abedmouleh A., Laforcade P. and Oubahssi L..

Speciﬁcation of Visual Instructional Design Languages Dedicated to Learning Management Systems.

DOI: 10.5220/0004079501990204

In Proceedings of the 7th International Conference on Software Paradigm Trends (ICSOFT-2012), pages 199-204

ISBN: 978-989-8565-19-8

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

Kam, 1996).The specification of a learning scenario

can be realized by means of an Educational

Modeling Language (EML), which provides a

framework of elements used for its formal

specification (Koper and Manderveld, 2004). EMLs

can be considered as authoring languages focusing

on exchange, binding and delivery objectives.

Other EMLs propose a visual notation and

focuses on specific instructional design theories or

methodologies. These Visual Instructional Design

Languages (VIDL) offer a support encouraging

reflexion, communication, etc. (Botturi and Stubbs,

2007). However they do not systematically provide a

binding support for formalizing models. Some of

them provide then some partial translations towards

some more abstract EMLs like SCORM or the de

facto standard IMS-LD.

Learning scenarios are then deployed or

executed in a Learning Management System (LMS).

Nevertheless LMSs failed in providing EMLs

support for automatic deployment. Past attempts

proposed to deeply modify LMSs by adding them an

execution engine dedicated to the considered EML

(Berggren et al. 2005). Some other research works

tried to bridge EMLs and LMSs by the means of

web services (Dodero et al., 2010) or models

transformation (Abdallah et al., 2008). Nevertheless

they met some semantics losses inherent to the

translation mapping. Finally, the most spread way to

deliver learning scenarios consists in using them as a

formal guide to directly set up “by hand” the

equivalent course within the LMS.

2.2 Domain Specific Modeling

The Domain-Specific Modeling (DSM) (Kelly and

Tolvanen, 2008) is a software engineering

methodology for designing and developing systems.

The DSM approach is an application of principles

and techniques from the Model Driven Engineering

domain. DSM involves the systematic use of

graphical languages to represent the various facets

of a system. They are specific to a domain and can

be defined as the set of concepts and their relations

within a specialized problem field. They offer

primitives whose semantics are familiar to all

practitioners in that domain.

Thanks to some past experimentations and

studies about using DSM for specifying VIDLs we

concluded in (Laforcade, 2010) that such DSM

approach can help the emergence of communities of

interests or practices sharing the same domain-

vocabulary and formalisms. DSM tools can be used

to support the specification of VIDLs and the

development of specific editors.

3 LMS-CENTERED VIDLS WITH

A DSM APPROACH

3.1 Context

Issues from existing approaches lead us to propose

an original approach focusing on existent LMSs. We

assume that LMSs are widely spread into academics

institutions and that it is relevant to focus on helping

teacher-designers in using them whereas propose yet

another design solution that do not deal with binding

or automatic deployment facilities compliant with

these LMSs.

Our approach then follows two objectives: to

facilitate the design of learning scenarios in

accordance with the LMS abilities (hiding low level

and technical details required by the form-oriented

LMSs screens), and to propose a solution exploiting

these scenarios as productive models for pre-

configuring the corresponding courses within the

targeted LMS. We then consider our instructional

design proposition as LMS-centered.

In order to overcome the limits and constraints

inherent to the technological and technical choices

related to the development of the LMS, we propose

to focus on an LMS external solution. However, a

communication bridge is necessary between these

external LMS-centered design tools and the LMS.

3.2 Formalization of the LMS

Instructional Design Language

Our approach is based on the idea that every LMS is

not pedagogically neutral but embeds an implicit

instructional design language relying on specific

paradigms and educative theories followed by the

LMS providers.

We propose to identify and formalize them in

order to exploit them as new specific formats for

import/export exchanges between LMSs and

external instructional design tools (similarly to the

more specific formats some ones sometimes provide

like the self-Moodle-format for importing quizzes).

In our mind, such self-labels can be considered

as equivalent to the standard ones (SCORM, IMS-

LD) because of their focus on instructional design

but they have to exclude the managing of resources

in order to be deployed as a self-contained XML

file. From an MDE viewpoint, this LMS language

could be considered as an abstract syntax (the

ICSOFT 2012 - 7th International Conference on Software Paradigm Trends

200

instructional design entities, properties and

relations), its related semantics, formalized with an

XML-oriented concrete syntax. Our proposal

requires an LMS modification: a specific import

add-on has to be developed and added to the LMS.

Nevertheless such extensions are generally allowed

by most spread LMSs.

3.3 LMS-Centered Vidls following a

DSM Approach

The explicit formalization of LMSs‘way of

designing’ will allow tools providers to propose

different design tools communicating with LMSs.

Some ones could focus on delivering or

implementation issues for LMS-compliant scenarios

by means of transformations or direct binding

facilities. Other tools could focus on LMSs

interoperability by translating some source LMS-

centered scenarios to a specific targeted LMS

language.

In our research work our interest is to help

teacher-designers that use to directly design their

learning situations from the LMS interfaces. We then

aim at developing specific external LMS-centered

design tools helping them to focus on design aspects

at a sufficient level of abstraction from a considered

LMS (e.g. hiding some low-levels configurations

required by LMSs). On one hand future-authoring

tools could deal with some instructional design

aspects in a first external design-time but, on the

other hand, some low-level aspects will still require

to be set up in a second design-time on the LMS.

We concretely propose to specify LMS-centered

VIDLs and to develop external dedicated authoring-

tools. According to the DSM approach we follow,

such VIDLs specification can rely on the LMS

abstract syntax previously captured by the XML

schema. This concrete format acts as a base for the

specification of VIDLs metamodels and as a binding

target for the serialization of future produced models

(machine-interpretable models). The automatic

delivering of VIDLs-compliant models will be

achieved by the means of both binding facilities

(from authoring editors) and importation services (to

add to LMSs). The DSM tooling will assist the

specification of the VIDL domain model from the

XML schema and they will guide the definition of

graphical concrete syntaxes and mapping models

linking abstract and concrete syntaxes as well as

capturing other semantics. DSM tools also take

charge of the editor code generation from domain,

graphical and mapping models.

On one hand this LMS-centered approach allows

to overcome the translation losses inherent to the

semantics gap between the VIDL and the targeted

LMS. On the other hand, such approach necessary

limits the VIDL expressiveness and usages (design

close to the LMS semantics in opposition to

conceptual design close to the practitioners needs).

Nevertheless, first practitioners’ feedbacks (from

surveys and interviews conducted within the

teachers-designers community from our academic

institution using a Moodle-based LMS) argue in

favor of our original position. They do not have

well-formalized instructional design background and

practices to support. They ask for very first design

tools allowing an abstraction of the LMS low-level

details.

4 LMS EXPERIMENTATION

4.1 The Moodle LMS

We chose to apply our proposal on the Moodle LMS.

It provides a learning environment to create courses,

define activities, manage and grade students and so

forth. Moodle include many types of activities (as

lessons, assessments, forums, databases, quizzes,

etc.). Moodle has an open source code and has a

modular and extensible architecture allowing the

addition of new modules. It has also a large

community of users and developers. The design of

courses on Moodle is based on the setting up of

many interfaces based on long forms mixing

pedagogical elements with technical ones.

The identification and formalization of the

implicit Moodle instructional design language have

already been performed and discussed into

(Abedmouleh et al., 2012) by combining three

viewpoints: users interfaces analysis (what designer

see), functional analysis (what the LMS can do), and

database and other technical sources analysis (how

the LMS realizes and persists the design

components). The final XML schema we finally

fixed has also been used to develop a dedicated

Moodle module allowing importation of course

contents. This module appears as a block in the

course space for a teacher-designer. So it requires

the context of an empty created course to be used.

The importation/exportation process allows a kind of

round-trip design process ensuring that

configurations directly made using the LMS

designing facilities (including low-level data) will be

preserved and merged according to the changes

realized outside the LMS.

Specification of Visual Instructional Design Languages Dedicated to Learning Management Systems

201

4.2 Practitioners’ Requirements

For a first experimentation we decided to focus on

objectives and practitioners’ needs allowing the

specification/development of a prototypal VIDL and

editor in order to verify our DSM approach and

tooling. These are the requirements:

• to design graphically sections by spatially

arranging them without a definitive ordering;

 to allow the drawing of connecting arrows

between sections to represent their future

ordering within the course;

 to propose in the palette the basics activities and

resources facilities provided by the LMS;

 to allow the addition of these activities and

resources into the sections to define their use

without having to specify all the usual data

required for each of them.

Practitioners concretely would like a

diagrammatic-oriented authoring tool, specific to

their LMS, allowing them to focus on the global

design of their courses.

4.3 DSM Tooling

Since several years we use the open-source unified

set of modeling frameworks, tooling, and standard

implementations from the Eclipse Modeling Projects

(Eclipse, 2012): EMF and GMF. Our experience

proves us that final editors, developed thanks to this

tooling, tackle the need for graphical editors about

learning scenarios (Laforcade, 2010).

Nevertheless, this Eclipse tooling requires some

expertise about DSM and MDE principles (meta-

modeling at first). In order to customize the

generated editor or design more complex user-

friendly editors (e.g. for modeling various views or

perspectives for a same learning scenario),

developers will have to acquire a higher level of

expertise about the frameworks and the underlying

Eclipse RCP principles.

4.4 Using EMF

First of all, the domain model defining the abstract

syntax of the VIDL to build has been directly

specified from the LMS XML schema. Indeed, EMF

provides such a facility. EMF also keeps a trace of

the mapping in order to drive, when the model code

will be automatically generated from the meta-

model, the persistence of future models. This tackles

our need for a binding facility.

The figure 1 illustrates an extract of the domain

model (as a class-diagram representation). This

model specifies that a Course is composed of one

Sections, itself composed of ordered Section (the

ordering is a propriety of the ‘Ereference’ between

Sections and Section; it can be checked in the

Eclipse ‘property’ view). Section can include many

Activities (forum, workshop, chat quiz, etc.). All the

Eclasses (Moodle, Course, and so on) are defined in

the model in order to map to the corresponding

ComplexTypes from the initial XML schema.

Figure 1: Extract of the domain model.

4.5 Using GMF

According to the GMF guidance, we had then

specified the notation model in conformance with

the practitioners needs. It specifies some inter-

related drawing primitives (line, rectangle,

compartment, etc.).

Next step concerns the specification of the

tooling model: what users will have at their disposal

in the palette, menus, toolbars, etc.

Figure 2: Extract of the GMF mapping model.

Finally, the mapping model is specified (Figure

2). It aims to link the three previous models. For

instance, it specifies that the canvas (drawing space)

ICSOFT 2012 - 7th International Conference on Software Paradigm Trends

202

maps to the ‘Moodle’Eclass from the domain model

(itself in relation with the top tag “Moodle” defined

within the XSD file), that the compartment within

every ‘Section’/’Rectangle’ is related to the

‘Activities’ Eclass, and so on.

4.6 Resolving Some Meta-modeling

and Binding Issues

We on purpose propose to relate a specific obstacle

we met when dealing with the ordering of sections

within a course. Their ordering is defined within the

XSD file as the one of the “Section” tags from the

future XML files. When getting the meta-model

from the XSD file by the EMF importation service,

this information has been translated as an Ereference

between the EclassesSections and Section (both

relating to the Sections and Section

tags/complexTypes) with the property ordered fixed

at ‘true’. Unfortunately it is not able to map an

arrowed drawing line notation (from the graphical

model) to this property. The GMF logic consists in

mapping this information according to the creation

sequence of visual Section(s) within the

compartment of a course. This is concretely an issue

because practitioners do not know in advance the

concrete order of the sections they are designing. In

order to use the authoring-tool for sketching the

global design of the course, we have to consider

separately the section instantiation order from their

pedagogical one. To solve this issue we made the

following modifications:

• addition of a self nextSection Ereference on the

Section eClass with lower/upper bounds to “0..1”

and a transient attribute to ‘true’ in order to inform

the EMF persistence mechanism to not deal with it

(in red color within figure 1).

• definition of the corresponding notation

(arrowed line) into the graphical model, and of the

corresponding tool for the palette;

• mapping specified into the related model for

bridging together these new elements;

• addition of OCL constraints, to the mapping

model, in order to disable self next relation on a

section and to detect cycles.

• modification of the model code (generated at

first by EMF) in order to redefine the behavior of the

save/load methods: the save method have to re-order

the Section instances from their Sections parent

according to the nextSection relations specified;

similarly, the load method have to set the

nextSection according to the Section tags order

parsed from the XML file.

4.7 The Final Authoring-Tool

From all the previous models (domain, graphical,

tooling, mapping), GMF provides a generator model

to give access to implementation customizations.

Then this last model drives the GMF generative

component to generate an editor code directly usable

as a plugin for Eclipse (a Rich Client Platform

standalone application can be further configured).

This code use the one generated by EMF from the

domain model that we have customized to solve the

meta-modeling issues we met.

Figure 3: Screenshot of a scenario specified thanks to the

generated authoring-tool.

The final editor (figure 3) can be used for two

purposes: (1) to draw and then design learning

scenarios as graphical models and (2) to visualize a

learning scenario from another tool, which depends

on the condition that this file/model is compliant to

the schema used by the persistence facility. The

models are both visualized by a diagram-oriented

view and synchronously serialized as machine-

interpretable XML file in conformance with the

XML schema we started from.

Figure 4: Screenshot of the Moodle course space after

importing the previous model.

This graphical editor meets the practitioners’

Specification of Visual Instructional Design Languages Dedicated to Learning Management Systems

203

requirements. Figure 3 shows a caption-screen of a

learning scenario within the editor whereas Figure 4

shows the equivalent result after importing the

learning model into Moodle. Teacher-designers can

complete their design directly on Moodle by

focusing on the low-level details not dealt with by

the external design tool.

5 CONCLUSIONS

This article has presented and discussed a Domain

Specific Modeling approach for the specification of

Visual Instructional Design Languages centered on

the Learning Management System semantics. The

DSM theories and practices provide a very

challenging trend for supporting the specification of

VIDLs as well as the generation of dedicated visual

editors. The main practical advantage and added

value is to synchronize human-interpretable visual

models with machine-readable persistence. The

EMF/GMF-based editor delivers learning scenarios

in a machine readable format compliant to a specific

schema like the one we propose to identify and

formalize from a specific analysis of the internal and

implicit instructional design principles embedded

within the LMS.

Nevertheless, the work conducted for now has

also depicted the difficulty to adapt the resulting

meta-model from the XSD file in order to capture

the practitioners semantics and allowing a mapping

in conformance to the notation targeted by end-

users. We have just started a French ANR funded

project in order to study the specification of more

complex LMS-centered VIDLs that will capture

practices and requirements closer to teachers-

designers than the experimentation we discussed

within this submission.

REFERENCES

Abdallah, F., Toffolon, C., and Warin, B., 2008. Models

transformation to implement a Project-Based

Collaborative Learning (PBCL) Scenario: Moodle

case study. In ICALT 2008. Santander (Spain) pp. 639-

643.

Abedmouleh, A., Oubahssi, L., Laforcade, P., and

Choquet, C., 2012. Expressing the implicit

instructional design language embedded in an LMS:

motivations and process. In the 15th IASTED

International Conference on Computers and Advanced

Technology in Education (CATE 2012) Napoli, Italy.

(To be published).

Berggren, D., Burgos, J. M., Fontana, D., Hinkelman, V.,

Hung, A., Hursh, and Tielemans, G., 2005. Practical

and Pedagogical Issues for Teacher Adoption of IMS

Learning Design Standards in Moodle LMS. Journal

of Interactive Media in Education.

Berger, C., and Kam, R., 1996. Definitions of instructional

design. In http://www.umich.edu/ed626/define.html.

Retrieved fromMarch 2012 pp. 16.

Botturi, L., and Todd, S., (Eds.). 2007. Handbook of

Visual Languages for Instructional Design: Theories

and Practices. Information Science Reference.

Burgos, D., Tattersall, C., Dougiamas, M., Vogten, H., and

Koper., R., 2007. A First Step, Mapping IMS Learning

Design and Moodle. In Journal of Universal

Computer Science 13 (7).pp. 924-931.

Dodero, J. M., Martìnez del Val, A., and Torres, J., 2010.

An extensible approach to visually editing adaptive

learning activities and designs based on services.In

Journal of Visual Languages and Computing 21

(2010) pp.332-346.

Eclipse, 2012. The Eclipse Modeling Project. http://www.

eclipse.org/modeling/. Retrieved from March 2012.

Kelly, S., and Tolvanen, J. P., 2008.Domain-Specific

Modeling. ISBN: 978-0-470-03666-2. Wiley-IEEE

Computer Society Press.

Koper, R., and Manderveld, J., 2004. Educational

modelling language: modelling reusable,

interoperable, rich and personalised units of learning.

In British Journal of Educational Technology Vol 35

N°5.pp 537-551

Koper, R., 2006.Current Research in Learning Design. In

Educational Technology & Society, 9 (1), pp.13-22.

Laforcade, P., 2010. A Domain-Specific Modeling

approach for supporting the specification of Visual

instructional design languages and the building of

dedicated editors. In Journal of Visual Languages and

comuting. pp. 347-358.

Martinez-Ortiz, I., Sierra, J. L., and Fernández-Manjón,

B., 2009. Enhancing IMS LD Units of Learning

Comprehension”, In the 4th International Conference

on Internet and Web Applications and Services.

Venice, Italy.

ICSOFT 2012 - 7th International Conference on Software Paradigm Trends

204