A DEVELOPMENT PROCESS FOR WEB GEOGRAPHIC

INFORMATION SYSTEM

A Case of Study

María José Escalona, Arturo Torres-Zenteno, Javier Gutierrez

Department of Computer Languages and Systems, University of Seville, Seville, Spain

Eliane Martins, Ricardo da S. Torres, M. Cecilia C. Baranauskas

Department of Information System, Institute of Computing, State University of Campinas, Campinas, Sao Paulo, Brazil

Keywords: Software engineering, Geographic information systems, Software requirements and specifications.

Abstract: This paper introduces a process for developing Web GIS (Geographic Information Systems) applications.

This process integrates the NDT (Navigational Development Techniques) approach with some of the

Organizational Semiotic models. The use of the proposed development process is illustrated for a real

application: the construction of the WebMaps system. WebMaps is a Web GIS system whose main goal is

to support harvest planning in Brazil.

1 INTRODUCTION

A geographic information system (called GIS from

now) is a software system that manages geo-

referenced information. GIS systems are an

automated system used for storing, analyzing and

manipulating geographical information.

Geographical information represents objects and

actions where geographical location is indispensable

information (Aronof, 1989; Bull, 1994). In this

context, a Web GIS system offers different GIS

services for analysis and visualization of

geographical information on the Web (Kim, 2002).

The purpose of this paper is to propose a process

for developing Web GIS systems. This process

integrates models from Navigational Development

Techniques (NDT) methodology (Escalona et al.,

2004) with models from the Organizational Semiotic

technique (Liu, 2000).

The process proposed in this paper has been

applied for developing the WebMaps system.

WebMaps is a real Web GIS system whose main

goal is to support harvest planning in Brazil

(Macário et al., 2007).

This paper is organized as follows. Section 2

introduces preliminary concepts needed to

understand the process. Section 3 describes the

techniques used in the process. Then, section 4

presents the case study based on WebMaps. Finally,

section 5 outlines conclusions and ongoing works.

2 PRELIMINARY CONCEPTS

2.1 Navigational Development

Techniques (NDT)

NDT (Escalona et al., 2004) is a methodological

process based on the navigation of web and

hypermedia systems. NDT defines the Requirements

Engineering and Analysis phases of a software

development process. In the Requirements

Engineering phase, NDT defines four models:

information storage requirement model, functional

requirements model, actors model, and interaction

requirements model. The information storage

requirements model specifies the information needs

of the system under development. This model

answers the following questions: what information

must the system store? What information does the

system need?

Functional requirements model specifies the

operation of the system, this model answers the

following question: what can the system do? The

112

José Escalona M., Torres-Zenteno A., Gutierrez J., Martins E., da S. Torres R. and Cecilia C. Baranauskas M. (2008).

A DEVELOPMENT PROCESS FOR WEB GEOGRAPHIC INFORMATION SYSTEM - A Case of Study.

In Proceedings of the Tenth International Conference on Enterprise Information Systems - HCI, pages 112-117

DOI: 10.5220/0001668101120117

 SciTePress

actors model specifies the roles of actors that

interact with the system, their incompatibilities and

generalization among them.

The interaction requirements model is a relevant

model for a navigational system. The interaction

requirement model provides the information and the

functionality asked by the user.

In the Analysis phase, NDT defines two models.

The conceptual model is a set of conceptual classes.

These classes represent the static structure of the

system. The navigational model is a class diagram

with special classes that offers a view of the

conceptual model and shows how to navigate

through the information managed by the system

The main documents obtained in NDT are: the

System Requirement Document (SRD) in the

Requirements Engineering phase and the System

Analysis Document (SAD) in the Analysis phase.

2.2 Organizational Semiotics

As a “doctrine of signs”, Semiotics covers several

disciplines (Liu, 2000). The main goal of the

Semiotic is to facilitate the understanding of the way

signs are used by people for all purposes. Every

organized behavior is affected by the

communication and interpretation of the signs by

individual or by groups of people. Thus, the

Organizational Semiotics studies organizations, by

taking advantage of concepts and methods from

Semiotics.

Organizational Semiotics offers methods to

understand the organizational context. The Methods

for Eliciting, Analyzing and Specifying Users’

Requirements (MEASUR) (Stamper et al., 1997).

MEASUR is a set of methods based on modeling

rules and requirements specification that can be used

for system development

Some of the methods included in MEASUR are

the PAM (Problem Articulation Method), the SAM

(Semantic Analysis Method) and the NAM (Norm

Analysis Method). The PAM is one of the methods

used in this paper. It may be applied in early stages

of the development process, when a vague and

undefined problem is found. The techniques

included in PAM are:

• Stakeholder analysis: This technique performs

an analysis of the actors, customers, providers,

partners, competitors, users and government.

• Evaluation table: this technique verifies for

each stakeholder the main problems at the time of

applying those techniques.

• Semiotic framework: This technique specifies

characteristics in the level of physic world, empiric

world, syntactical world, semantic world, pragmatic

world, and social world.

• Collateral analysis: This technique helps

during analysis and building of relations among

unitary systems, which involve complex systems.

This work also used the SAM method. This

method is used for capturing and representing the

requirements using a formal and precise model

• Ontology diagram: The ontology diagram

defines concepts and terms used in the domain of a

particular problem.

Finally, this paper refers to a case study in which

we applied an adaptation of the mentioned methods

at the Institute of Computing, UNICAMP

(Baranauskas et al., 2005). This adaptation is

included:

• a description of the context of the problem.

• a description of the functional and non

functional requirements.

• an organization of the content: it specifies the

structure and organization of the content into the

system.

3 DEVELOPMENT PROCESS

FOR WEB GEOGRAPHICAL

INFORMATION SYSTEMS

In the first activity, the process begins with a

Requirement Engineering phase. Next, an Analysis

phase is performed. Later, the Abstract Interface

model is built. Finally, the target system is

implemented.

Figure 1 shows the development process. Shows

how models from NDT and from Organizational

Semantic are built following this process. The

OOHDM (Oriented-Object Hypermedia Design

Method) (Rossi, 1996) models are presented in the

Fig. in order to reference the classic models.

OOHDM is one of the most studied approaches.

This approach for web developing is based on

Hypermedia Design Model (HDM) (Garzotto et al.,

1993) a structured approach to model hypermedia

system

Next sections describe in depth the development

process for Web GIS systems.

3.1 Requirements Engineering

The main goal of the requirement engineering is to

find the real needs of the users. These needs will

drive the implementation of the system under

development. Recent researches have proved that

A DEVELOPMENT PROCESS FOR WEB GEOGRAPHIC INFORMATION SYSTEM - A Case of Study

113

many software projects have failed due to errors in

requirements elicitation (Boehm). This means that

requirements are often incompletes or

misunderstood.

Figure 1: Development process for Web GIS applications

(extended).

This process introduces the use of the

Organizational Semiotics during the requirement

elicitation to solve that problem (Liu, 2000).

Organizational Semiotics uses a mature set of

techniques present in MEASURE (Stamper, 1997),

for the definition of requirements. This phase

involves a set of workshops among all interested

people. The main objective of the workshops is to

find the actors of the system and the functional and

non-functional requirements. Social, politic, cultural,

and ethics issues of the problem are also considered

(Baranauskas et al., 2005).

Once the requirements have been obtained using

Organizational Semiotics, NDT methodology

(Escalona et al., 2004) is applied for improving

requirements. Actor model, information model and

interaction model are added to the requirements.

3.2 Analysis

The objective of the analysis phase is to detail,

define and validate the requirements. In the NDT

methodology, the analysis phase involves the

definition of two main models: conceptual model

and navigational model. A relational model might be

generated from the conceptual model. The relational

model is useful for the implementation of the

databases of the system under development.

NDT is helpful in the generation of navigational

model for the pages that include information

processing. However, NDT is not helpful for

organizing static information. The Organizational

Content approach is, then, used to complement the

content and models generated by NDT. The

Organization content used has been proposed and

adapted from the ontology diagram by the research

team at Institute of Computing, Unicamp

(Baranauskas et al., 2005).

3.3 Implementation of the Abstract

Interface Model

During the system design phase or abstract interface

generation phase, the design of the pages is needed.

In this phase, the web designer may also use the

artifacts from the analysis phase. This means that the

web designer should use the final navigational

model for implementing user prototypes.

4 CASE STUDY – WEBMAPS

WebMaps is an interdisciplinary project still in

development at the Institute of Computing,

Unicamp. The main goal of this project is the

monitorization and planning of harvests in Brazil.

Several experts of different areas of knowledge

participate in WebMaps, including: image

processing, software engineering, database, geo-

processing, eco-environmental, and computer-

machine interaction. WebMaps integrates

heterogeneous data from several sources. WebMaps

also offers several services to manage all those

information. This system allows defining a base

platform for the designing, implementation and

evaluation of integrated policies of harvests

planning. The system also allows the execution of

queries about the evolution of the harvests for a long

time line.

The WebMaps system is able to generate

Normalized Difference Vegetation Index (NDVI)

curves (CPETC, 2007). A NDVI curve describes the

vegetation indexes for the normalized difference for

a given region. Those indexes show the amount of

vegetation in a concrete time period. Those indexes

are obtained by processing satellite images -

Moderate Resolution Imaging Spectroradiometer

(MODIS) images.

ICEIS 2008 - International Conference on Enterprise Information Systems

114

The images are obtained periodically, therefore

indexes are always updated. The images cover all

the Brazilian territory during a four years period.

The images have been obtained from NASA

(MODIS, 2007). Every image has a size of 109’6

MB and a new image is obtained every 15 days.

The processing of the satellite images for the

generation of NDVI curves is a slow process. Due

this reason, a set of masks were generated and

applied to the images to reduce the processing time.

The architecture is composed of a client layer and a

server layer. The server is divided into three

modules. All modules have been implemented using

Java and C languages. The first module is in charge

of supporting user and property registration. The

second module supports the queries. Finally, the

third module stores the managed data using

PostgreSQL database and hard disk files.

JSP pages and C modules are stored in the server

side. The goal of C modules is to process the

satellite images and to generate the masks used to

build the NVDI curves, as mentioned earlier. The

curves are generated by taking into account the

query defined by the user. The interaction between C

and Java code has been implemented using the Java

Native Interface (JNI). Benefits of using native code

interfaces in the developing of system are exposed in

(Czajkowski, 2001).

Next paragraphs describe all the phases of the

process introduced in the previous section and show

their application in the WebMaps development.

4.1 Requirements Engineering

This phase was initiated with the application of the

Organizational Semiotics (Liu, 2000) adapted

methods. Several workshops were organized for

eliciting requirements. The workshops were

conducted by requirements engineers and

stakeholders in a cooperative way. Seventeen

participants from several knowledge areas worked

together in the workshops. The knowledge areas of

the participants were: image processing, software

engineering, databases, geo-processing, agro-

ecologic studies, human-computer interaction,

among others. The artifacts of the Organizational

Semiotics were used as communication tools among

workshops. Reference (Baranauskas et al, 2005)

presents in detail the use of the Organizational

Semiotic during the workshops.

The requirement engineering phase uses the

patterns and models developed by NDT (Escalona et

al., 2004) – see Section 2.1. This section only

exposes the application of the process and it does not

show all patterns

First, NDT patterns drive the identification of the

system objectives. The goal of this task is to know

the limits of the problem. Some WebMaps

objectives are: showing data about a concrete

geographical region, working with geo-referential

data, and adapting the system according to user

profiles.

Once the objectives are identified, NDT proposes

to identify the system requirements. One of the

system requirements are the information

requirements. Those requirements define the

information that the system must store. Once

information requirements are specified, the actors

who will interact with the system must be identified.

Every actor is related to an interaction role. In the

WebMaps system, there are for different user

categories: non registered, registered, coordinator,

and member of equipment.

After the specification of the actors, the

incompatibilities among those actors are analyzed.

The definition of incompatibility is basic, due every

incompatibility may define different interface of the

other interfaces generated for the other actors. The

data that will be stored into the system and the users

that may manage those data have been defined in the

previous activities. However, it is also needed to

know which operations may be performed with the

stored information and which functionalities and

services may be offer to the users.

NDT proposes two diagrams for capturing and

defining the functional needs of the system under

test. The first diagram is the use cases diagram

(Baranauskas et al, 2005). The second one is the

textual information diagram. This diagram is

specified using patterns to define the meanings of

the information.

The use case model of WebMaps was generated

from the artefact obtained in the activities performed

in the workshops and, mainly, from the application

of the Organizational Semiotic. Use cases also show

the interaction of the different actors with the

system.

At this time, the information that the system will

store and the actors that will interact with the system

and the functionality of the system are known and

formally defined. However, this is not enough for

navigational systems. The system must also offer

information and functionality at the right moment.

Furthermore, it defines the most adequate

presentation strategy. The order of information is a

relevant aspect. Those requirements are stored in

interaction requirements. NDT defines the

A DEVELOPMENT PROCESS FOR WEB GEOGRAPHIC INFORMATION SYSTEM - A Case of Study

115

interaction requirements with two patterns: recover

criteria (or phrases) and visualization prototypes.

Visualization prototypes are defined after the

definition of the phrases. Visualization prototypes

allow specifying expected navigation process into

the system.

Once requirements are identified, they must be

validated. The construction of a traceability matrix is

a value technique which allows validating the results

obtained from the requirement phase.

The requirement engineering phase ends at this

time. The document of system requirement (SRD) is

the final artefact obtained.

4.2 Analysis

The goal of the analysis phase is the definition of the

analysis models: the conceptual and the navigational

model. The conceptual model represents the static

structure of the system, while the navigational model

represents its navigational structure. NDT describes

a systematic process that allows generating both

models from the artefacts obtained in requirement

engineering phase.

The conceptual model is composed of a single

diagram. This diagram is represented by two

elements: the conceptual classes diagram and the

information dictionary. The conceptual classes

diagram uses the notation proposed by UML. The

information dictionary describes, in a textual way,

the meaning of classes and associations from the

conceptual class diagram.

Once the static structure of the system has been

created, the navigational model is defined. The

navigational model represents how to navigate

through the conceptual information. This model is a

view of the conceptual model. It represents the

elements of the navigation and their relations. The

navigational model may be quite different depending

on the roles of the user logged into the system. As

mentioned before, NDT proposes a set of process to

generate this model systematically.

NDT describes the elements of a navigational

model using patterns, in the same way that models

used in the previous phases. After the definition of

the basic navigation model, a revision of the model

is needed. In the case study, we found out that the

navigation model generated involves the dynamic

information navigation as well as information that

need to be processed by the server. However, it was

detected the lack of a navigational model for the

static content. The Organizational Semiotics offers

an artefact to solve this limitation. Figure 2 shows

the relation between the ontology model with the

system content. Those models were generated by the

human-machine interface engineers of WebMaps

(Baranauskas et al., 2005).

Figure 2: Ontology model (Baranauskas et al, 2005).

The ontology model presents a context that

involves concepts and terms used in the problem

domain. A more complete description about the

generation of those models can be found in

Baranauskas et al, 2005. Main menus of the system

are generated from the product obtained after the

application of Organizational Semiotics adapted

methods. Now, it is possible to build the final

navigational mode, shown in Figure 3. The basic

navigational model is associated with the main

menus of the system.

4.3 Abstract Interface Model and

Implementation

In this phase, a web designer collaborated with the

WebMaps team. The web designer built the abstract

interfaces. The language used was Java.

Figure 3: Final navigational model.

As WebMaps is a Web GIS system, it needs

large data processing. Reference (Torres-Zenteno et

ICEIS 2008 - International Conference on Enterprise Information Systems

116

al., 2006) presents the functional testing of

WebMaps. The goal of the test process was to

determine if the behaviour of the system is similar to

that one indicated by the functional requirements.

5 CONCLUSIONS

This paper has presented a development process

model focused on Web GIS applications. The model

integrated artifacts from NDT and artifacts from the

Organizational Semiotic. This process was applied

to a real system called WebMaps. This system is a

Web GIS application, which main goal is to support

harvest planning in Brazil.

After the application of the process, a functional

prototype of WebMaps was obtained.

The system requirements document and analysis

document were created following the indications of

the process. Both documents were not included in

this paper due space limitations. However, this paper

has shown examples of the most relevant models to

improve the understanding of the process.

We are currently defining the conceptual model

of the geographic data. We plan to add this model in

the proposed development process.

ACKNOWLEDGEMENTS

This work was supported by CNPq, CNPq

WebMaps project, FAPESP, and CAPES.

REFERENCES

Aronoff, S., 1989. Geographic Information Systems. WDL

Publications, Canada.

Baranauskas, M. Cecilia C., Schimiguel, Juliano, Simoni,

Carlos A., Medeiros, Cláudia B., 2005. Guiding the

Process of Requirements Elicitation with a Semiotic-

based Approach – A Case Study, HCI – International

USA.

Boehm, B., 1981. Software Engineering Economics,

Prentice-Hall.

Bull, G., 1994. Ecosystem Modelling with GIS.

Environmental Management. 18(3):345-349.–438.

Czajkowski, G., Daynes, L. and Wolczko, M.: Automated

and portable native code isolation. Software

Reliability Engineering, 2001. ISSRE 2001.

Proceedings. 12th International Symposium on 27-30

Nov. (2001) Page(s):298 – 307.

CPTEC, http://satelite.cptec.inpe.br/ndvi/, última visita:

septiembre de 2007.

Escalona, M. J., 2004. Modelos y técnicas para la

especificación y el análisis de la navegación en

sistemas software. PHD Thesis. Escuela Técnica

Superior de Ingeniería Informática. Universidad de

Sevilla.

Garzotto, F., Schwabe, D. and Paolini P., 1993. HDM-A

Model Based Approach to Hypermedia Application

Design. ACM Transactions on Information System, 11

(1), pp 1-26.

Kim, Do-Hyun, Kim, Min-Soo, 2002. Web GIS service

component based on open environment. Geoscience

and Remote Sensing Symposium, IGARSS '02. IEEE

International Volume 6, Page(s):3346 - 3348 vol.6.

Liu, K., 2000. Semiotics in Information Systems

Engineering. Cambridge University Press, Cambridge.

Macário, Carla, Medeiros, Claudia, Senra, Rodrigo, 2007.

IX Brazilian Symposium on GeoInformatics - Geoinfo

2007 p.239-250.

MODIS, Moderate Resolution Imaging Spectroradiometer,

http://modis-atmos.gsfc.nasa.gov/, última visita:

septiembre de 2007.

Stamper, R. K., 1997. Organizational Semiotics. In

Mingers J. & Stowell F. (eds.) Information Systems:

An Emerging Discipline. Mc Graw Hill, 267-283.

Rossi, G., 1996. An Object Oriented Method for Designing

Hipermedia Applications. PHD Thesis, Departamento

de Informática, PUC-Rio, Brazil.

Torres-Zenteno, A., Martins, E., Escalona Cuaresma M. J.,

Torres da S, R., 2006. Teste de Desempenho em

Aplicações SIG Web. IX Workshop Iberoamericano

de Ingeniería de Requisitos y Ambientes de Software

IDEAS’06, 449-462. Argentina.

UML, Unified Modeling Language. http://www.uml.org/,

última visita: septiembre de 2007.

A DEVELOPMENT PROCESS FOR WEB GEOGRAPHIC INFORMATION SYSTEM - A Case of Study

117