DESIGN AND IMPLEMENTATION OF ADAPTIVE WEB

APPLICATION

Raoudha Ben Djemaa, Ikram Amous, Abdelmajid Ben Hamadou

ISIMS, LARIM, Route Mharza Km 1, BP 1030 – Sfax, Tunisie

Keywords:

Adaptive web application, MCSD-case, user profile, Java Servlets.

Abstract: Web applications are now widely used for diffusing and processing information over the network.

Methodological guidelines which assist web application developers in their task must take into account

users’ preferences to generate adaptable web interfaces. This paper presents MCSD-case, a generator of

adaptive web application, which addresses the issue of designing and automatically deploying such

application. MCSD-case‘s target is to facilitate the automatic execution of the design and the automatic

generation of adaptable web interface. Using MCSD-case, designers can specify, at a conceptual level, the

features of the web application to be generated. These features are made operational by instantiating

different models dedicated to the description of the expected functionalities, the users and some features

concerning adaptation. The implementation of MCSD-case is based on HTML and Java Servlets to

instantiate the models which are translated in XML files then XSL files to generate the HTML page

corresponding to the user aims.

1 INTRODUCTION

From its introduction in the early 90's the World

Wide Web (WWW) is in a continuous development.

Its rapid expansion results in an increasing number

of Web applications being developed, especially

with sources that contain frequently changing

information such as databases (relational or object-

oriented databases, XML repositories etc.)(Bieber et

al., 1998). This leads to the fact that there is a higher

need to automate, at least partially, the design

process of web application. For this reason, there are

many methods like RMM (Isakowitz et al, 1995),

OOHDM (Schwabe and Rossi, 1998), WSDM (De

Troyer and Leune, 1998), etc. These methods have

been originally developed for a manual hypermedia

design process; they are not particularly well-suited

in the context of automated adaptive web

application. In this context, efforts in the design of

today's hypermedia system have to be pursued and,

especially, to be concentrated on the adaptation

issue.

Adaptation can be defined as the ability a system

has to dynamically adapt itself to any user by giving

her/him the feeling that the system has been

designed especially for her/him. The need for

adaptation arises from different aspects of the

interaction between users and hypermedia systems.

Users’ categories which deal with these systems are

increasingly heterogeneous due to their different

interests and preferences. The web can be accessed

through a number of different devices (PC, WebTV,

WAP phone, etc.). Finally taking into account the

user preferences (desired layout, navigation patterns,

etc.) can lead to a more effective interaction. For this

reason, in previous work (Djemaa et al., 2004), we

have presented a user driven method for modeling

adaptive Web applications, called MCSD. The

different models of this method are: the

functionalities model, the audience model, the

adaptation model, the analysis model, the conceptual

model and the navigational model.

In this paper, we will present MCSD-case, a

generator of adaptive web application, which

addresses the issue of designing and automatically

deploying such application. The design approach

supported by MCSD-case is presented in section 2.

The prototype architecture is detailed in section 3.

Then, we compare our approach and prototype with

some related works in section 4. Finally, section 5

concludes the paper and suggests future research

directions.

151

Ben Djemaa R., Amous I. and Ben Hamadou A. (2006).

DESIGN AND IMPLEMENTATION OF ADAPTIVE WEB APPLICATION.

In Proceedings of the Eighth International Conference on Enterprise Information Systems - HCI, pages 151-156

DOI: 10.5220/0002452101510156

 SciTePress

Table 1: Design steps in MCSD-case.

N° Step Model Figure

Analysis of the functional

requirements of the web

application

Functionalities model based on three

concepts: functional class, functional

role and Functionalities.

Modeling of users Audience model expressed through a

UML class diagram generated by an

algorithm of generation of Functional

Spaces.

A1 A2 A4 A5

Internal actor External actorAdministrator

Actor

Specification of the user’s profile Adaptation Model expressed as a

specific UML class diagram.

graphic charter

back ground

color

style

Personal

informations

pi-code

name

first-name

age

sex

langage

hardware

hard-code

type -hard

memory

resolution

device

capabi liti es

Software

soft-code

soft-name

soft-version

Profil

prof-code

ceate()

delete()

consult()

modify()

0..1

navigation

nav-code

type -nav

Preferenc es

pref-code

0..1

Content s

code-fonct

typ-fonct

Presentation

pres-c ode

1..*

charter

chart-code

1..*

composition

charter

name

charter-code

Region

rg-code

rg-name

1..*

Description of the presentation Presentation model composed of

Composition charter and

Graphical charter

2 WEB APPLICATION DESIGN

WITH MCSD-CASE

MCSD-case is an environment dedicated to the

automatic generation of web application given some

conceptual specifications. Its target is to facilitate

the automatic execution of the design and the

automatic generation of adaptive web interfaces. It

should be possible to program the web applications

in such a way that it can automatically execute the

process specified by the designer. The design

approach of MCSD-case is resumed in Table 1. Each

of these steps is presented in the following section.

2.1 Requirements Analysis

Met with the increasing needs of users of Web

applications, we propose a functionalities model,

gathering users’ informational and functional needs.

Level1: Functional

class

(FC)

Level2: Functional

Role

(R)

Level3:Functionalities

(F)

Level4:Functional

space

(FS)

Administrative Informational Professional

R1 R2 Rk

FS-adm

FS-

2 FS-

FS-

F11

F12

F1i

F1n

F21

F22

F2i

Fk1

Fk2

Fki

Fm1

Fm2

Fmi

Fmr

Figure 1: the functionalities model.

This model is to be looked in figure 1.

At the highest level, functionalities are classified

into three classes, called Functional Classes,

summarising the behaviour and the functioning

mode of Web applications. These classes are

abstracted by Administrative, Informational and

Professional ones.

• The Administrative functional class gathers all

the actions and the functionalities which take part in

the configuration of the system management,

relative to the system administrator.

• The Informational functional class gathers all

the functionalities enabling the users to have access

to the system to acquire specific information. This

class translates the set of the static and dynamic

informational users’ needs.

• The Professional functional class is devoted to

the representation of the actors’ functional needs.

Indeed, Web applications are based on technologies

which make their contents dynamic, enabling, thus,

the user to modify the applicative state of the server

by carrying out a set of functionalities.

Each of these functional classes, represented

above, will be subdivided into sub parts called

Functional Roles,fixed by the designer.

At the following level, the "Functional Role"

concept will be decomposed by the designer into a

set of elementary functionalities representing both

informational and functional needs. The choice of

the decomposition parameters is left to the designer.

ICEIS 2006 - HUMAN-COMPUTER INTERACTION

152

At the lowest level of the Functionalities Model,

the concept "Functional Space” provides each actor

with a list of authorized functionalities.

In our approach, the Functionalities Model

implementation is carried out in two stages. During

the first stage, the designer is invited to define the

first three levels of the model. The fourth level,

relating to the functional space of the actors, is

approached only after having defined the list of the

actors involved in the application. However, this list

is defined in the following section.

2.2 Modeling of Users

Web applications are intended to be used by a

variety of individual users, not necessarily

presenting the same needs, or having the same rights

of access to data. Users, in fact, show varied

knowledge and experience, and express various

preferences, etc. Therefore, the identification of the

groups of actors, called in our approach Audience

class (De Troyer and Leune, 1998),

is a means of

reducing complexity subjacent with such

heterogeneity.

By definition, an audience class is a potential

user group which belongs to the application target

audience and which has the same informational and

functional needs. These classes are not necessarily

disjoined (a user can belong to several classes of

audience).Nevertheless, despite belonging to the

same audience class, the actors show individual

characteristics sometimes varying to a subset of

individual class or to a particular individual of this

class. Thus, the system must be able to take into

account various audience classes and their derived

sub-classes.

In our approach (Djemaa et al., 2004), we have

proposed an algorithm which generates the list of the

actors of the application starting from the concept of

Functionality. This algorithm makes it possible to

specify all the classes of audience, to represent them

in a hierarchy and to determine the concept

"Functional Spaces" relating to each class. We have

presented in (Djemaa et al., 2004), the process of

operation of this algorithm.

2.3 Adaptation Model

A web application can be accessed through a

multitude of devices and by different users. Each

device has its own capabilities (display size,

memory size, network speed, etc...). Every user has

his specific preferences (desired layout, navigation

patterns, etc...).

An intelligent web application needs to take into

account these constraints (abilities) coming from

both devices and users, and adapt the presentation

preferences.

In fact, two types of adaptation or

personalization are distinguished: adaptability

(Atezni et al., 1997) and adaptativity (Frasincar and

Houben, 2002). Adaptability (or static adaptation)

means that the generation process of web interfaces

is based on available information that describe the

information about device capability and user

preferences in which the user will use the generated

presentation. Adaptativity (or dynamic adaptation) is

the kind of adaptation included in the generated

adaptive hypermedia presentation that itself changes

while being browsed. In this paper, we focus on

generating web intrfaces by using only static

adaptation.

In order to specify adaptation in MCSD, we have

extended our method by an adaptation model. In

(Djemaa et al., 2006) we have presented different

dimensions treated for this proposition. This model

takes into account the user’s devices capabilities

(hardware and software), the user preferences

presentation (desired layout, navigation patterns,

etc.) and personal informations (e.g. Age, language,

etc).

2.4 Presentation Model

The presentation model is intended to create the

models of the Web pages. Two dimensions are

considered:

1. the organization of the Web pages through

the definition of composition charter (Djemaa et

al., 2006).The composition charter describes the

various regions of a page and their composition

thus allowing homogenizing the site pages in

terms of contents.

2. their appearance based on the specification

of graphic charters (Djemaa et al., 2006). In fact,

in order to adapt the presentation of the web

application pages, we propose in MCSD-case,

Java Servlets interfaces which require users

information’s concerning their preferred choices

about:

− Visual characteristics: police (color, size,

style), background (color or image).

− Navigation patterns: absolute or relative.

− Type of media used for each web page

(representing functionality) and the corresponding

joint file.

This information serve to instantiate the profile

model which is exported by MCSD-case in XSL

files (cf section 3).

DESIGN AND IMPLEMENTATION OF ADAPTIVE WEB APPLICATION

153

The concept of charter is to be brought closer to

a presentation model. Our objective is to guarantee

coherence between various pages of the web

application. This cohesion of the interface is,

according to us, particularly necessary in the

adaptive web application. The final step, before the

generation of web interfaces, consists in filling the

different models with information. This step is

detailed in the following section.

3 ARCHITECTURE OF

MCSD-CASE

The previous section dealt with the engineering

process of MCSD-case methodology. This section

focuses on the process of generation of adaptive web

applications and describes how the system is

dynamically adjusted to varying audience classes.

The architecture of MCSD-case is shown in figure 3.

For a designer modeling an adaptive web

application using MCSD-case consists firstly, in

instantiating the functionalities model. Once

instantiated, this model is translated into XML files

to be stored in the data repository n and the system

execute the algorithm of generation of audience

classes o which built the audience model. The last

is also translated into XML file p which contains

the functional space for each audience class.

At this stage, the content is adapted to each

audience class. So, to adapt the user’s presentation

preferences, the designer is invited to instantiate the

profile model using three java Servlets interfaces

that treat respectively the following axes:

1. Personal information about the user like name,

age, sex, language.

2. User’s presentation preferences defined in

term of web pages composition and graphical

aspects specified by two charters called composition

and graphical charters.

3. User’s devices capabilities which define some

characteristics of the hardware and software

environment.

After instantiation, the profile model is exported

in XSL files in the data repositoryq. Finally, XML

and XSL fies are extracted from the data repository

r and they are sent to the parser s in order to

publish the HTML page corresponding to the

audience classt. In figure 3 we show the interfaces

of MCSD-case used to adapt the content of the web

application represented by the steps of the

instantiation of the functionalities model, the

execution of the algorithm of the generation of

audience classes and finally by the instantiation of

the audience model to generate us results functional

spaces for each audience class.

In figure 4, we show some results generated by

MCSD-case used to adapt the presentation of the

web application and we present the interface of

administration of the prototype. At this stage,

MCSD-case is able to assure the adaptability of the

the content and the adaptability of the preferences

presentation.

Figure 2: Architecture of MCSD-case.

User ‘s machine

Data repository

XML, DTD, XSL

t generation of

adaptive interface

HTML

XML File

XSL File

HTTP server (apache)

Servlets En

ine

(

TOM CAT 4.0

)

n XML

ortatio n

p XML

ortatio n

FS-ac

Audience

model

Instanciation of the

functionalities

model

For an audience class Cai

Instanciation of a

rofile model

User/Pre ferences

profil

User/P latform

rofil

User/Personal

profil

q XSL

ortatio n

parser

r Extraction

Network

ICEIS 2006 - HUMAN-COMPUTER INTERACTION

154

a. Home page of MCSD-case b. How MCSD-case instantiate the

functionalities model?

c. How MCSD-case instantiate

the audience model?

Figure 3: Adaptability of the content in MCSD-case.

d. XML file of the functionalities

model

e. Result of the algorithm of

generation of audience model

f. Administration of MCSD-

case

Figure 4: Administration and results of MCSD-case.

4 RELATED WORKS

Most of the Web application design approaches

proposed at the end of the 90’s are inspired from

earlier hypermedia design methodologies (RMM

(Isakowitz et al, 1995) or OOHDM (Schwabe and

Rossi, 1998)) which address content, presentation

and navigation issues. These three dimensions are

also taken into account in our approach, with a

special emphasis put on adaptability to user. As a

consequence, we propose an explicit modeling of

users as opposed to Araneus (Atezni et al., 1997).

WSDM (De Troyer and Leune, 1998) also

represents users explicitly, but does not uses

algorithm of generation of functional space for each

user. The last concept allows us to adapt the content

to user which constitutes a particularity of MCSD-

case.

Adaptability also appears through the choice of a

graphical charter, and through the composition of

pages which can be define in MCSD-case for each

audience class. The independence between content

and presentation, notably allowed by the XML/XSL

technology, promotes adaptability since it makes

possible the association of different presentations for

a same content. WebML (Ceri et al., 2000) and

KIWIS (Villanova-Oliver et al., 2002) adopts a

similar approach, but with lesser possibilities for

user for instantiate his profile.

Similarly to MCSD-case, tools for Web

application development support the steps of their

associated methods, and generally perform the Web

application generation. A particularity of MCSD-

case is the ability of its use through the network.

5 CONCLUSION

In this paper, we have presented MCSD-case, a

generator for adaptive web application. We have

demonstrated how this generator can facilitate the

automatic execution of the design and the automatic

generation of adaptive web interfaces.

During the different activities that compose the

engineering process of MCSD-case there are a lot of

ontologies involved: semantic level (Functionalities

model and audience model), adaptation level (profile

DESIGN AND IMPLEMENTATION OF ADAPTIVE WEB APPLICATION

155

model) and presentation level (presentation details in

terms of composition and graphic).

The implementation of MCSD-case is based on

java Servlets interface to instantiate the models

which are translated in XML files. MCSD-case uses

then XSL files to generate the HTML page

corresponding to the user. As future work, we

mention:

− The extension of the MCSD-case

implementation to take into account users’ devices

capabilities (hardware and software).

− The construction of a graphical tool that

supports the presentation of the other models of

MCSD and outputs XML files.

− The inclusion of the mecanism of the dynamic

adaptation (adaptativity) in MCSD-case in order to

dynamically elaborate and modify both the

functional space and the navigation patterns,

learning from the user’s behavior.

REFERENCES

Bieber, M., Isakowitz, T., Vitali, F., July 1998. Web

Information Systems. In Communications of the ACM,

41(7), pp. 78-80.

Isakowitz T., Stohr A. & Balasubramanian E., August

1995. RMM: A methodology for structured

hypermedia design. In Communications of the ACM

38(8), pp. 34-44.

Schwabe D. & Rossi G., 1998. Developing hypermedia

application using OOHDM. In Proceedings of

Workshop on Hypermedia development Process,

Methods and Models, pp. 85-94.

De Troyer O., Leune C., April 1998. WSDM: a User-

Centered Design Method for Web Sites, In

Proceedings of the 7th International World Wide Web

Conference (WWW7), Computer Networks, 30(1-7).

Ben Djemaa R. Jmaiel M., Ben Hamadou A., Mai 2004.

Vers une méthode centrée-acteur pour la conception

des applications Web. Eighth Maghrebian Conference

On Software Engineering and Artificial Intelligence.

Mcseai’04.

Frasincar F. & Houben G-J., 2002. Hypermedia

Presentation Adaptation on the Semantic Web. In

Proceedings of the 2nd International Conference on

Adaptive Hypermedia and Adaptive Web-Based

Systems (AH 2002), Malaga, Spain, May 29 - 31, ,

LNCS 2347, pp. 133-142.

Ben Djemaa R. Amous I., Ben Hamadou A., February

2006. GIWA: A generator for adaptive Web

applications. In ICIW’06, IEEE International

conference on Internet and Web Application and

Services.

P.Atezni, G. Mecca and P. Merialdo, August 1997. To

Weave the Web, 23rd Conf. On Very Large Databases

(VLDB’97), pp 206-215, Athens, Greece.

Ceri, S., Fraternali, P., Bongio, A., May 15-19, 2000. Web

Modeling Language (WebML): a modeling language

for designing Web sites. 9th WWW Conference,

Amsterdam.

Villanova-Oliver M., Gensel J., Martin H., Erb C., April

8-10, 2002. Design and Generation of Adaptable Web

Information Systems with KIWIS. 3th IEEE

Conference on Information Technology ITCC-2002,

Las Vegas, USA.

R. Vdovjak, F. Frasincar, G. J. Houben, and P. Barna.

Engineering semantic web information systems in

hera. Journal of Web Engineering, Rinton Press,

2(1&2): pp 003--026, 2003.

Fiala Z., Hinz M., Houben G.J., and Frasincar F., March

2004. Design and Implementation of Component-

based Adaptive Web Presentations. Sac’04, Nicosia,

Cyprus.

Brusilovsky P., 2001. User Modeling and User-Adapted

Interaction, Ten Year Anniversary Issue 11 (Alfred

Kobsa, ed.), pp 87-110.

Fiala Z., Hinz M., Meissner K., and Wehneer F.,

September 2003. A Component-based Component

architecture for adaptive dynamic Web documents.

Journal of Web Ingeineering, Rinton Press, 2(1&2):

pp 058-073.

Henze, N., and Nejdl, W., AH 2002. Knowledge modeling

for open adaptive hypermedia. In Proccedings of the

2nd International Conference on Adaptive

Hypermedia and Adaptive Web-Based Systems.

Rossi G., Schwabe D., and Guimaraes R., Designed

personalized Web application. In WWW10, The Tenth

International Conference on the world Wide Web,

Hong Kong, 2001

ICEIS 2006 - HUMAN-COMPUTER INTERACTION

156