TOWARDS EXTENDING IMS LD WITH SERVICES

AND CONTEXT AWARENESS

Application to a Navigation and Fishing Simulator

Valérie Monfort

1,2

, Slimane Hammoudi

and Maha Khemaja

Université de Sfax, MIRACL, Tunisie

Université de Paris 1, Panthéon Sorbonne 90 rue de Tolbiac, 75634 Paris cedex 13, France

ESEO 4, Rue Merlet de la Boulaye, B.P. 9249 009, Angers Cedex 01, France

PRINCE Research Group, ISITC, GP1 Hammam Sousse 4011, University of Sousse, Tunisia

Keywords: e-Learning, LD Standard, Web Services, Context Aware Adaptation, Meta Modeling, MDD.

Abstract: A few e-Learning platforms propose a solution for ubiquity and context aware adaptability. Current

standards, as Learning Design (LD), require an extension to propose context awareness. Based on previous

related works, we define a fully interoperable and learner (ambient) context adaptable platform, by using

metamodeling based approach mixing MDD, parameterized transformations, and models composition. The

scope of this paper is to extend LD metamodel as a first step. We use a concrete software engineering

industrial product that was promoted by French Government.

1 INTRODUCTION

E-learning aims the delivery of a learning, training

or education program by electronic and it involves

the use of a computer or electronic device (e.g. a

mobile phone), in some way, to provide training,

educational or learning material. Concerning the

architecture point of view, e-learning platforms

gather two separated and distributed parts as:

authoring tools (for pedagogical contents definition)

and execution platforms. So, e-learning may: i) use

several media and devices, ii) promote specific

training according to learner skills, iii) send specific

events to increase complexity of lessons and to

assess learner reactions, …

Previous works allowed us to use Web services to

get interoperability and flexibility to changes. But,

we noticed the lack of adaptability, so, we extended

Web services to introduce adaptability with aspects

(Kiczales, 1997) (Tomaz, 2006). We noticed this

very efficient and pragmatic solution was very

technical. Recently, we have investigated a model

driven approach and context awareness to provide

developers mechanisms that allow them representing

an application in abstract way (in a model) and, then,

automatically generating the corresponding code

(Monfort, 2010), (Monfort, 2009). We aimed to

explore adaptability and flexibility on a service

platform using context with the benefits of an MDD

(Model Driven Development) (OMG, 2001)

development strategy. These benefits are related to

productivity, quality, adaptability and maintenance.

Moreover, e-learning standards as Learning

Design (LD) tend to extend their semantic to Web

services standards (Dietz, 2004). We studied e-

Learning standards meta models, but, we noticed no

semantics concerning context aware adaptability.

We aim to propose a fully interoperable and

learner (ambient) context adaptable platform, firstly

by using meta modeling based approach mixing

MDD, parameterized transformations, and models

composition. This paper is based on a concrete

industrial project and we present here a part to

illustrate our research work. This project aims to

develop a concrete navigation and fishing simulator.

This e-learning application allows different learners

to discover navigation and fishing business and to

become fishery captain, boat mechanics, sailor, …

We shall process as followed. The second section

presents a part of the fishing and navigation

simulator project. The third section shows the

diversity of e-learning standards, the efforts of

connection between e-learning and Web services

standards, and focuses on LD. The fourth section

Monfort V., Hammoudi S. and Khemaja M. (2010).

TOWARDS EXTENDING IMS LD WITH SERVICES AND CONTEXT AWARENESS - Application to a Navigation and Fishing Simulator.

In Proceedings of the 12th International Conference on Enterprise Information Systems - Information Systems Analysis and Speciﬁcation, pages 80-91

 SciTePress

introduces context aware meta model. The fifth

section presents parameterized transformations and

our global approach. The sixth shows a composition

example between LD and context aware meta model

as an extension of LD meta model based on our

industrial project. Last sections present some related

works and conclusion. Let us present our project.

2 THE FISHING SIMULATOR

PROJECT

We aim to design and implement a genuine fishing

simulator. This software was intended to navigation

and fishing schools or fishing ship owner

companies. Following partial UML model (fig.1)

shows the different services proposed by the

navigation and fishing e-learning system. This

platform provides diploma for students (learners) in

fishing schools and certificate for companies. The

Teacher may be human or not. The system can take

decision and can send specific events to complicate

lesson if the student has a good level.

Figure 1: e-learning Services modeling.

Fig. 2 shows a process where the Teacher is

preparing training and is sending it to the student

who identifies himself by invoking identification

and authentification services, linked to rules

manager. While training learner may receive specific

events during current lesson to assess learner’s

skills. The learner is evaluated at any time. This

platform provides diploma for students (learners) in

fishing schools and certificate for companies. The

Teacher may be human or not. The system can take

Figure 2: Training services according to BPMN.

decision and can send specific events to complicate

lesson if the student has a good level.

Fig. 2 shows a BPMN process where the Teacher

is preparing training and is sending it to the student

who identifies himself by invoking identification

and authentification services, linked to rules

manager. While training learner may receive specific

events during current lesson to assess learner’s

skills. The learner is evaluated at any time.

The following e-learning architecture (fig.3)

shows services proposed by providers. Actors as

learner and teacher can work anywhere and use

different media. Services are available according to

SaaS model and managed by providers according to

Cloud Computing (Reese, 2009) principles. Contents

are defined by the teacher with authoring tools

according to standards (see previous section). He

defines e-learning tasks sequences, and so, e-

learning services orchestrations. He informs the

services providers by loading training content and

the concerned learner. The learner may use genuine

navigation equipments and/or simulation. The

genuine navigation equipments are linked to a

middleware able to interpret signals coming from

equipments and to send them to Execution platform.

The learner may also use PDA or mobile phone.

Learner receives the training scenario and while

training sends information and invokes services. The

e-learning middleware (ESB for Enterprise Service

Bus) manages: routing messages, transporting

messages and transforming exchanged data.

The providers repositories manage at least these

basic following services as: business, training

TOWARDS EXTENDING IMS LD WITH SERVICES AND CONTEXT AWARENESS - Application to a Navigation and

Fishing Simulator

supervisor, course virtual management, planning

management, collaborative management,

subscribing management, time/tracking

management, …

The fishing and navigation e-learning platform

architecture is fully flexible, it is able to take into

account any functional or technical changes. E-

learning platforms are generally based on specific

standards. Let us see how these standards contribute

to converge on Web services standards that support

interoperability between systems.

Figure 3: General architecture.

3 E-LEARNING STANDARDS

3.1 Convergence on Web Services

Standards

Many definitions may be found, we can sum up

them with this definition: e-learning aims the

delivery of a learning, training or education program

by electronic means. E-learning involves the use of a

computer or electronic device (e.g. a mobile phone)

in some way to provide training, educational or

learning material. E-learning systems gather two

separated and distributed parts as: authoring tools

(for pedagogical contents definition) and execution

platforms.

Authoring tools are based on several standards

as: LOM (Learning Object Metadata) to define a

structure for elements and meta data useful to

describe pedagogical resources. Content Packaging

proposes to describe a pedagogical data package

referencing resources in a set of interoperable

packages.

Learning Design defines a pedagogical scenario

with its components (roles, activities, environment,

results). SCORM (Sharable Content Object

Reference Model) is a set of technical standards

allowing to find, import, share, reuse, export,…

normalized teaching contents.

On the other hand, IMS-GWS (General Web

Services Standards) allows interoperability between

any Web services. IMS-GWS may be compared to

WS-IBP (WS-I Basic Profile) to promote

interoperability with Web services and profile. They

use the same mechanisms and semantics based on

SOAP, WS-Security and Addressing and allow

referencing a document into SOAP message.

IMS-TI (Tools Interoperability) and WS-

Federation also promote interoperability between e-

learning platforms and/or components. They aim to

build a trustable architecture by using WS-Security

tokens. IMS-TI uses IBAT tools based on UML

meta modeling and models transformations from one

platform to another. BPEL and IMS-SS (Simple

Sequencing) may be compared as far as concerned

training tasks orchestration. IMS-SS provides a

simple description and BPEL semantic is wider.

Other e-learning standards are useful to manage

pedagogical contents. They are all based on XML so

they are Web services compliant.

In following section we focus on Learning

Design (LD) that helps to define a pedagogical

scenario with its components (roles, activities,

environment, and outcomes).

3.2 LD Model

LD is a description of a method enabling learners to

achieve intended learning objectives and outcomes

by performing predefined learning activities. More

specifically, a learning design is a means allowing

the Instructional designer to describe a learning

scenario in terms of a set of activities that learners

should perform according to the different roles that

they may play within environments (e.i. Run-time

environment). Environments are described in terms

of Learning Objects and Services that should assist

learners during the-learning process. IMS-LD

(Instructional Management Systems-Learning

Design) (Boticario, 2007) (Davinia, 2004)

specification provides for previously described

concepts a meta-model (fig.4.) that was and is still

used by LD authoring tools developers. According

to IMS-LD specification, LD concepts must meet

ICEIS 2010 - 12th International Conference on Enterprise Information Systems

height requirements. We name the third one because

it deals with personalization that is relevant to our

work. The LD specification states that: “The content

and activities within a unit of learning can be

adapted based on the preferences, portfolio, pre-

knowledge, educational needs, and situational

circumstances of users. In addition, the control over

the adaptation process must be given, as desired, to

the student, a staff member, the computer, and/or the

designer”. However, IMS-LD provides neither

means nor modeling solutions to take into account

contextual data of mobile users for instance. We

should stress that the contextual data, unlike those

already defined for personalization, are dynamic and

may depend on the user’s external environment.

We notice LD proposes an entity for context in its

meta model but, according to us, this approach is too

semantically poor. Moreover, this entity does not

take into account ambient and/or context adaptation.

It is the reason why we want to propose to extend

this model with model showed in fig.2 with

composition mechanisms according to (Klein, 2006)

(Lundesgaard, 2007). But, first of all, let us define

what we mean by Model Driven Development and

“context aware”.

Figure 4: Semantic aggregation in LD specification

(http://tecfa.unige.ch/guides/tie/pdf/files/pedago-ld.pdf).

4 CONTEXT AWARENESS

MODELING

4.1 Model Driven Development

At the beginning of this century, software

engineering needs to handle software systems that

are becoming larger and more complex than before.

Object-oriented and component technology seem

insufficient to provide satisfactory solutions to

support the development and maintenance of these

systems. To adapt to this new context, software

engineering has applied an old paradigm, i.e.

models, but with a new approach, i.e. Model Driven

Development (MDD). In this new global trend,

Model Driven Architecture (MDA) is a particular

variant. MDA is based on standards from the Object

Management Group (OMG) (Tomaz, 2006); it

proposes an architecture with four layers (Dey,

2001): meta metamodel, metamodel, model and

information (i.e. an implementation of its model).

MOF (Meta Object Facility) is a standard from

OMG for metamodels specification. The

development is based on the separation of concerns

(e.g. business and technical concerns), which are

afterwards transformed between them. So, business

concerns are represented using Platform-

Independent Model (PIM), and technical concerns

are represented using Platform-Specific Model

(PSM). Finally, it is well recognized nowadays that

model transformation is one of the most important

operations in MDA. In the context of the basic four

levels Metamodeling architecture of MDA, various

scenarios of model-to-model transformation have

been identified. Fig.5 presents the most common

scenario of these transformations, which is

compatible with the MOF2.0/QVT standard (Tomaz,

2006). Each element presented in Fig. 5 plays an

important role in MDA. Transformation rules

specify how to generate a target model (i.e. PSM)

from a source model (i.e. PIM). To transform a

given model into another model, the transformation

rules map the source into the target metamodel. The

transformation rules are based on a transformation

language, such as the standard QVT. The

transformation engine takes the source model,

executes the transformation rules, and produces the

target model as output.

Adaptable Service platforms have been proposed for

the development of mobile context-aware

applications. The development of such platforms

involves a number of challenges from which we

consider two main issues in the context of our

approach of model driven development: i)the

definition of a metamodel to describe the contextual

domain in which a given application or service is

defined, ii)a mechanism to integrate the context into

the business application using a model driven

approach. These two main issues are described in the

sequel.

TOWARDS EXTENDING IMS LD WITH SERVICES AND CONTEXT AWARENESS - Application to a Navigation and

Fishing Simulator

Figure 5: Architecture of Transformation Process.

4.2 MDD and Context for Service

Adaptability

Context awareness is a quite new discipline in e-

learning domain. For instance, in (Strang,

2004)(Mary, 2005), the authors noticed the context

acts like a set of constraints that influence the

behavior of a system (a user or a computer)

embedded in a given task. They discussed the nature

and structure of context but they notice the lack of

representation of context in e-learning domain. The

emergence of new technologies, in particular

wireless communications and the increasing use of

portable devices (smart phones, Personal Digital

Assistants (PDA), laptops…), has stimulated the

emergence of a new computing paradigm called:

pervasive computing. In fact we have moved from

the desktop computing paradigm to the mobile and

ubiquitous computing paradigm. Pervasive

computing refers to the seamless integration of

devices into the user’s everyday life. “Appliances

should disappear into the background to make the

user and his tasks the central focus rather than

computing devices and technical issues.”(Weiser,

1991). Computing applications now operate in a

variety of new settings; for example, embedded in

cars or wearable devices. They use information

about their context to respond and adapt to changes

in the computing environment. They are, in short,

increasingly context aware. The context awareness

of such applications is the subject of a recent field of

studies in pervasive computing called: context-aware

systems. This terminology was discussed in (Shilit,

1994) and presented as “software that adapts

according to its location of use, the collection of

nearby people and objects, as well as changes to

those objects over time”. Since then, there have been

numerous attempts to define context-aware

computing. (Strang, 2004), (Tomaz, 2006) define

context-awareness as the ability of a program or

device to sense or capture various states of its

environment and itself. Referring to these latter

definitions a context-aware application must have

the ability to capture the necessary contextual

entities from its environment, use them to adapt its

behavior (run time environment) and finally present

available services to the user. In this sense and to

describe context-awareness independently from

application, function, or interface, (Pascoe, 1998)

proposes four features of context-aware application :

(1) Contextual sensing which refers to the detection

of environmental states and their presentation to the

user; (2) Contextual adaptation refers to the

adaptation of application behavior to the current

context; (3) Contextual resource discovery is the use

of context data to discover other resources within the

same context; (4) Contextual augmentation in which

the environment is augmented with digital data

associated to a particular context.

In (Dey, 2001), the authors introduce another

definition in which they insist on the use of context

and the relevance of context information. The

authors consider that: “a system is context-aware if

it uses context to provide relevant information

and/or services to the user, where relevance depends

on the user’s task”. They explain how to use context

and propose a classification of the features of

context-aware applications that combine the ideas of

(Shilit, 1994). They consequently define three

categories of features that context-aware

applications may support as (1) presentation of

information and services to the user; (2) automatic

execution of a service; and (3) tagging of context to

information for later retrieval.

All the works converge to a general architecture

composed of five ordered layers presented in fig.6.

The complete description of every layer is given in

(Strang, 2004).

Figure 6: Architecture of Context-Aware System.

“Context Information” in context-aware systems

and what could be the definition of context has been

the topics of many recent works. Various definitions

are given and are summarized in (Mary, 2005).

Other definitions are extremely broad; the most

popular one is given by (Dey, 2001): “Context is any

information that can be used to characterize the

situation of an entity. An entity is a person, place, or

conformsTo

PIM

conformsTo

from

exec

conformsTo

MOF

Transformation

rule

Source

metamodel

Engine

del

Target

metamodel

output

input

Transfo.

Engine

Source

model

Target

model

Transformation Language

Application/Services

Storage/Management

Raw Data Retrieval

Preprocessing

Sensors

ICEIS 2010 - 12th International Conference on Enterprise Information Systems

object that is considered relevant to the interaction

between a user and an application, including the user

and applications themselves”. The authors give a

general definition that can be used in a wide range of

context-aware applications. To refine their definition,

they identify four categories of context that they feel

are more practically important than others. These are

location, identity (user), activity (state) and time

(Dey, 2001). In (Winograd, 2001) the author

approves this definition and claims that it covers all

proposed works in context. However he considers it

as a general definition that does not limit a context.

Thus he proposes his own definition in which he

limits a context in “a set of information, which is

structured and shared. It evolves and is used for

interpretation“. We stress that the notion of hierarchy

(structure) of context introduced by (Winograd,

2001) is important. The definition proposed in (Chen,

2000) also presents the context as hierarchically

organized. In this work the authors differentiate

between environmental information that determines

the behavior of mobile applications and that which is

relevant to the application. They thus define the

context as “the set of environmental states and

settings that either determines an application's

behavior or in which an application event occurs and

is interesting to the user”.

4.3 Maintaining the Integrity

of the Specifications

We consider that the defining context here is a set of

information structured in three dimensions:

- Actor: A person which is a central entity in our

system.

- Environment: in which the person evolves and

- Computational devices which are used by a person

to invoke services and capture the different states of

the environment. All the information related to the

three dimensions can also be shared by other mobile

applications. According to (Monfort, 2010),

(Monfort, 2009), fig.7 shows our context

metamodel. Our metamodel identifies and adds the

most relevant and generic contextual entities that

will be held in account in modeling any mobile and

context aware application. This context meta model

consists of six generic contextual entities

(represented in dashed colour), and four deduced

entities specific to a category of mobile applications.

The class “ContextView” groups all contextual

entities involved in a given application. It is

identified by name attribute and has two types of

relation: the aggregation “involves” and the

association “belongsTo”. The first relation expresses

that a given “ContextView” is composed of many

“ContextEntity” that are involved in a context-aware

application. The second relation “belongsTo”

expresses the use of historical context information.

A given context entity may have participated in

different context views. This information can be

helpful in the design of future context views. The

second generic entity of the metamodel is the

“ContextEntity”. As we see on the figure bellow, it

is specialized in three generic entities: Actor,

ComputationalEntity and Environnement. Actor may

be a person or another object that has a state and

profile. It evolves in an environment and uses

computational devices to invoke services. With the

ComputationalEntity, the computational device is

used by the actor to access the services and to

capture contextual information from the

environment. Usually, a mobile device is used in

context aware mobile applications, and can obtain

information concerning the type of device it is

(PDA, laptop, cellular phone…), the application, the

network, etc. The environment is constituted of all

the information surrounding the actor and its

computational device that can be relevant for the

application. It includes different categories of

information as:

i) Spatial context information can be location,

city, building,

ii) Temporal context information comprises time,

date, season,

iii) Climate can be temperature, type of

weather… The last entity is a profile. We are

convinced this entity is important in any user centred

context aware application. In fact, profile is strongly

attached to the actor and contains the information

that describes it. An actor can have a dynamic and/or

a static profile, and as ContextEntity class, he owns

a status. The static profile gathers information

relevant for any mobile context-aware application. It

can be the “date of birth”, “name” or “sex”. On the

opposite, dynamic profile includes customized

information depending on the specific type of

application and/or the actor. It can be goals,

preferences, intentions, desires, constraints, etc.

TOWARDS EXTENDING IMS LD WITH SERVICES AND CONTEXT AWARENESS - Application to a Navigation and

Fishing Simulator

5 PARAMETERIZED

TRANSFORMATIONS

FOR CONTEXT BINDING:

GLOBAL APPROACH

5.1 Parameterized Transformations

The separation of concerns (business and context) is

emphasized at a model level of our approach where

PIM and context models are defined independently,

and then merged by suitable transformation

techniques. Two types of transformations are

involved in our proposal. The first type of

transformation called “Parameterized

transformation” allows merging context information

with business logic at model level.

Figure 7: A Context Meta Model.

We have investigated (Vale, 2008) this type of

transformation which is not explored and there is not

a standard transformation language implementing it.

We will discuss shortly this type of transformation.

A CPIM model (Contextual Platform Independent

Model) is then obtained and fits together business

requirements with contextual data. The second type

of transformation is the traditional transformation

technique using a language such as QVT, which

operates in two steps.

Based on parameter paradigm, parameterized

transformation is a transformation technique which

is not sufficiently explored nowadays and there is

not a standard transformation language

implementing it. OMG (OMG, 2005) has defined the

concept of parameter as follows:

“A parameter specifies how arguments are

passed into or out of an invocation of a behavioural

feature like an operation. The type and multiplicity

of a parameter restrict what values can be passed,

how many, and whether the values are ordered”.

In (Frankel, 2003) David Frankel mentions the

importance of parameterization in model operations

using the association of tagged values with PIM and

PSM models. Tagging model elements allows the

model language to easily filter out some specific

elements. Transformation by parameter could be

used to improve new functionalities (values,

properties, operations) or to change the application

behaviour (activities).

In our approach, we are convinced that

parameterized transformation focusing on PIM to

PIM transformations is the fitted solution. The

designer must specify the parameters to be inserted

at the transformation phase. In our proposition these

parameters are context or context-aware and after

the transformation the application will join the

context information specified into the parameters as

illustrated in Fig. 8. A PIM model can be developed

without contextual details. User name, profiles,

device type, location can be added as parameters in

transformations. The same PIM can be re-

transformed and refined many times adding, deleting

or updating context information. The designer has to

specify into the application model the elements that

will receive the context information. A mark,

identified by the symbol #, is given for these

elements to be recognized by the transformation

engine. The marked elements represent context-

aware elements, in others words, the model elements

that can be contextualized.

Figure 8: Parameterized Transformation Concepts.

The transformation language must support

parameterization techniques. In our case the

parameters can be a Context Property and/or a

Context Data Type. We use templates to specify

which elements in application model are potentially

context-aware as depicted in Fig. 9. The

transformation engine has to navigate into the PIM

ICEIS 2010 - 12th International Conference on Enterprise Information Systems

model verifying the parameters and the elements

marked and then make the transformation which

consists in an update of contextual properties in a

PIM.

Template parameter (Vale, 2008) is an element

used to specify how classifiers, packages and

operations can be parameterized. UML 2.0 presents

that any model element can be templateable. For

independent context-aware models we need to

identify context elements that could be

parameterable. A parameterable element is an

element that can be exposed as a formal template

parameter for a template, or specified as an actual

parameter in a binding of template (Vale, 2008).

Context parameter can be expressed as constraint

and compared with the elements signature in

template parameter. This operation is named

“matching operation”. UML presents a Template

Signature element that defines the signature for

binding the template. Lets us see now related works

concerning this approach.

Figure 9: Context binding.

5.2 Our Global Approach

This section aims to present the two techniques of

context and aspect could be combined to achieve

service adaptability using a model driven approach.

Through Model Driven Development, context

models are built as independent pieces of application

and at different abstraction levels then attached by

suitable transformation techniques called

parameterized transformation. Context model

specify contextual entities that are involved in a

given context aware application. From a context

model, an aspect based services (ABS) model is

derived. This aspect model specifies the behaviors

linked to the context model.

- Fig. 11 illustrates the main models and

transformations techniques involved in our

MDD approach. Five main objectives are

illustrated: A separation between context

information (CM) and business

- logic (PIM) in individual models,

- The derivation of an aspect based service

model (AM) from a context model. A

context model specifies the contextual

entities with their properties (static view),

while the service based aspect model specify

behaviors (dynamic view).

- The integration of the context model into the

business logic using parameterized

transformation techniques. At this stage, the

CPIM model is enriched by contextual data

but the behavior part for adaptability at

execution level is missing.

- The Weaving process adds adaptability

mechanism producing a CPSM model.

- Finally, A CPSM model is mapped into a

service platform for future execution of

context-aware services.

Notice Aspect Based services weaving is a

specific model dynamic composition as related to

(Klein, 2006). In our approach aspect may be purely

business (AM) or technical (CPSM). Faced to these

previous research works and to our observations

concerning e-learning standards we propose now to

extend LD standard with model composition based

on our concrete case study.

6 EXTENSION OF LD

TO CONTEXT AWARENESS

6.1 Model Composition

Let us see now fig. 4 and fig. 7 and focus on

“Component “ and “Context View” classes

respectively. We propose following part model

(fig.10.) that shows Component Class may be

composed of Context View class and, so, compose

the two approaches (Monfort, 2009): (ambient)

context awareness and LD meta models.

Figure 10: LD and Context Meta Models composition.

During course, learner according to his skills

receives a navigation and/or fishing scenario as “ go

TOWARDS EXTENDING IMS LD WITH SERVICES AND CONTEXT AWARENESS - Application to a Navigation and

Fishing Simulator

to 100 miles from Saint Jean de Luz and fish tuna”.

So, the learner has to do obligatory tasks as: to check

weather, to define the road, to check the fitted nets,

to check mechanic,… The generic BPMN (Business

Process Modeling) process shows (fig.2) the

different tasks to do by the learner and the teacher.

The teacher programs a course that will be received

by the learner anywhere he is, via any media, …

after identifying himself. The learner is assessed in

real time and the teacher may send him events. At

the end of the module, the diploma is delivered or

not.

The learner is in front of his laptop and receives

the training. All the navigation tools (radar, sounder,

GPS,…) are simulated. According to his skills, the

teacher (human or system) can send to the learner

desktop specific events as mist, rain,… and the

learner has to react properly. Moreover, the system

provides an estimation of learner skills in real time.

The following partial models show a resulting

composed model: i) training meta model (that could

be later formatted to e-learning standards), ii) a

contextual model that was already composed to

component class from LD. Another composition may

also be done with fishery business metamodel. For

each training module, a link may be done with

specific business data. For instance, a training

module about tuna fishery involves the choice of the

fitted net. A mark is put on the required classes.

We have now to describe these models with Kermeta

tool and to define transformation rules from these

business models to technical models (corresponding

to the implementation platform).

6.2 Model Instantiation

Fig. 13 shows the different layers mentioned

previously. We present a part of fig. 12 as a Meta

model including for instance the following classes:

actor, profile, environment (Weather, location,

mobility settings). M1 classes instantiate M2 meta

classes and M0 instance level includes all the

instances of M1 classes. Let us see related works.

7 RELATED WORKS

Among related works, we can distinguish on one

hand those dedicated to e-learning as CSCW and

CSCL and, on the other hand, i) those that use MDD

for general context aware applications, ii) and those

that apply MDD and context for e-learning

application. We focused on context adaptability and

to our opinion, works as CSCW and CSCL do not

provide concrete solutions. In (Sheng, 2005) the

authors propose a UML based context metamodel for

the development of context-aware mobile

applications implemented on a Web services

platform.

Figure 11: Models, context and aspect for adaptability.

The proposed metamodel does not refine contextual

information and focuses on the association between

basic contextual structures with service invocation

interfaces for both contextual providers and context-

aware applications. In (OU, 2006), they have applied

MDA in context-aware application development.

They focus on the development of context-awareness

based on ontologies. However, neither a context

metamodel is proposed nor are transformation

techniques used.

In (DeFarias, 2007) authors investigate a number of

context models described in literature and propose a

context metamodel based on the main concepts and

strengths found in these models. The metamodel is

formally described using MOF and has been used as

a basis for the development of context-aware

applications and an associated service platform. All

these research works aim to explore adaptability and

ICEIS 2010 - 12th International Conference on Enterprise Information Systems

flexibility on a service platform using context and

models. But, neither of these works proposes an

explicit approach to integrate context into business

logic. For e-learning applications, some approaches

aims to use metamodeling:

i) to define e-learning interoperable and platforms

independent system

ii) and to extend standards as in (Boticario, 2007)

(Dietz, 2004) (Davinia, 2004).

Some researchers introduce adaptability with Multi

Agent System but we do not choose Artificial

Intelligence based approach. Previous works as

(Monfort, 2010), (Monfort, 2009), propose solutions

to model context. We did not find any concrete and

relevant related works concerning such an approach

in e-Learning domain, but we are convinced our

approach is pertinent because we got good results

with fishing simulator and in other Web based

application domains.

8 CONCLUSIONS

This paper proposes a metamodel approach to

introduce (ambient) context awareness in LD model.

It is based on our previous works about adaptability

and models composition based MDD. We propose

examples coming from a concrete industrial project.

We aim:

i) to define an independent platform model based

on services,

ii) to implement models transformations to link

these models to implementation platform,

iii) to promote automatic code generation…

We have to present transformation rules via a

technical platform based on services and supporting

context awareness and to enrich our example with

these rules.

Future works will contribute to propose other extra

formalism to describe dynamicity (as sequence

diagrams and Petri nets) and aspect based approach

supporting context awareness.

REFERENCES

Boticario, J. Olga. C Santos: An open IMS-based user

modeling approach for developing adaptive-learning

managements systems. Retrieved from www-

jime.open.ac.uk/2007/02/.

Chen, G., and Kotz, D. (2000). A survey of context-aware

mobile computing research. Tech. rep., Dept. of

Computer Science, Dartmouth College.

Davinia, H., Juan I. Asensio Perez, Yannis A. Dimitriadis,

"IMS Learning Design Support for the Formalization

of Collaborative-learning Patterns" Advanced

Learning Technologies, IEEE International

Conference on Advanced Learning Technologies

(ICALT'04), 2004. , pp. 350-354.

De Farias, C. R. G., Pires, L. F., and van Sinderen, M.

(2007). A MOF Metamodel for the Development of

Context-Aware Mobile Applications. In Proceeding of

the 22nd ACM Symposium on Applied Computing

(SAC'07) pages: 947 - 952.

Dey, A. K. (2001). Understanding and Using Context.

Personal and Ubiquitous Computing 5, 1, 4-7.

Dietze S., Alessio Gugliotta, and John Marko Rosić

Vlado Glavinić Branko Žitko: Intelligent authoring

shell based on Web services :

http://www.pmfst.hr/~bzitko/radovi/files/INES2004.p

Frankel S. David. (2003). Model Driven Architecture:

Applying MDA to Enterprise Computing, Wiley

Publishing, Inc.

Klein, J, Hélouet, and J. M. Jézéquel. -- Semantic-based

weaving of scenarios. -- In Proceedings of the 5th

International Conference on Aspect-Oriented Software

Development (AOSD'06), Bonn, Germany, March

2006.

Kiczales, G., Lamping, J., Maeda, C., and Lopes, C.

(1997). Aspect-oriented programming. In Proceedings

European Conference on Object-Oriented

Programming, volume 1241, pages 220–242.

Springer- Verlag, Berlin, Heidelberg, and New York.

Lundesgaard, S., Arnor Solberg, Jon Oldevik, Robert

France, Jan Oyvind Aagedal, Frank Eliassen,

Construction and Execution of Adaptable Applications

Using an Aspect- Oriented and Model Driven

Approach, IFIP DAIS 2007, LNCS 4531, 76-89, 2007

Matthias, B., Dustdar, S., and Rosenberg, F. (2007). A

survey on context-aware systems. International Journal

of ad Hoc and ubiquitous Computing 2 007.

Mary, B., and Patrick, B. (2005). Understanding context

before to use it. In 5th International and

Interdisciplinary Conference on Modeling and Using

Context vol. 3554 of Lectures Notes in Artificial

Intelligence, Springer-Verlag, pp. 29-40.

Monfort, V., Hammoudi, S. When Parameterized MDD

Supports Aspect Based SOA , IJEBR 2010,

International Journal of E-Business Research (To

appear).

Monfort, V., Hammoudi, S. ICSOC, Towards Adaptable

SOA: Model Driven Development, Context and

Aspect The 7th International Conference on Service

Oriented Computing, November 23-27 2009,

Stockholm, Sweden.

OMG (Object Management Group). (2001). Model Driven

Architecture (MDA), OMG document number

ormsc/2001-07-01.

ICEIS 2010 - 12th International Conference on Enterprise Information Systems

Ou S., Georgalas N., Azmoodeh M., Yang K. and Sun

X., (2006). A Model Driven Integration Architecture

for Ontology-Based Context Modelling and Context-

Aware Application Development, A. Rensink and J.

Warmer (Eds.) (2006). ECMDA-FA, LNCS 4066,

pp.188–197. Springer-Verlag Berlin Heidelberg.

Pascoe, J. (1998). Adding Generic Contextual

Capabilities to Wearable Computers. In 2nd

International Symposium on Wearable Computers

(Los Alamitos, California), Press, Ed., pp. 92-99.

Reese, G. Cloud Application Architectures: Building

Applications and Infrastructure in the Cloud, O’Reilly

Publisher (2009)

Sheng, Q. Z., and Benatallah, B. (2005). ContextUML: a

UML-based Modeling Language for Model-Driven

Development of Context-Aware Web Services. In

Proceedings of the International Conference on Mobile

Business (ICMB'05).

Schilit, B. N., and Theimer, M. (1994). Disseminating

active map information to mobile hosts. IEEE

Network 8 22-32.

Schneider, L. “LD” Retreived from

http://tecfa.unige.ch/guides/tie/pdf/files/pedago-ld.pdf

Strang, T., and Linnhoff-Popien, C. (2004). A Context

Modeling Survey. In First International Workshop on

Advanced Context Modelling, Reasoning and

Management, UbiComp.

Tomaz, R. F., Hmida, M. B., and Monfort, V. (2006).

Concrete solutions for web services adaptability using

policies and aspects. The International Journal of

Cooperative Information Systems (IJCIS), 15(3):415–

438.

Vale, S. Hammoudi, S. (2008) Context-aware Model

Driven Development by Parameterized

Transformation. Proceedings of MDISIS 2008

Workshop, joint to CAISE’2008. pages 121-133

Weiser, M. (1991). The computer for the 21st century.

Scientific American. PP 94-104.

Winograd, T. (2001). Architectures for context. Human-

Computer Interaction (HCI) Journal 16.

TOWARDS EXTENDING IMS LD WITH SERVICES AND CONTEXT AWARENESS - Application to a Navigation and

Fishing Simulator