TOWARDS A PROCESS TO INFORMATION SYSTEM

DEVELOPMENT WITH DISTRIBUTED TEAMS

Gislaine Camila Lapasini Leal

Production Engineering Department, State University of Maring´a, Maring´a, Brazil

C´esar Alberto da Silva, Elisa Hatsue Moriya Huzita, Tania Fatima Calvi Tait

Computer Department, State University of Maring´a, Maring´a, Brazil

Keywords:

Distributed software development, Development process.

Abstract:

The software engineering area offers support to manage the information systems development process. Soft-

ware engineering processes have artifacts, roles and activities well deﬁned, allowing adjustments in speciﬁc

situations. However, these processes are generally oriented to development of coallocated projects. They are

not taken into account peculiarities in regard to coordination, control and communication, when the develop-

ment is with distributed teams. The purpose of this paper is to contribute to the Software Engineering area

presenting a comparative analysis of some development processes from the perspective of development with

distributed teams. Additionally, it points to the need to process deﬁnition that includes the peculiarities of this

approach to software development.

1 INTRODUCTION

Nowadays, we live in a time that is increasingly gro-

wing need for high quality software. There it is neces-

sary to seek ways to improve the development pro-

cess, whether through the use of tools or artifacts

generated. An example would be the use of formal

methods to describe the speciﬁcation, design and test

(Abdurazik et al., 2000).

The software process is deﬁned as an ordered set

of activities for the management, development and

maintenance of software, and should be aligned with

the organizational conditions (Fuggetta, 2000). The

software as part of an information systems brings or-

ganizational elements besides technical such as: hu-

man resources or teams, organizational culture among

others. The processes of software development are, as

noted in the current literature, focused on the develop-

ment of projects co-located.

The Distributed Software Development (DSD) is

a recent approach to software development that meets

the demands of globalization. This new strategy

added to the software development challenges related

to cultural differences, geographic dispersion, coor-

dination and control, communication and team spirit.

The distributed teams face to challenges related

to technical and non-technical factors. The technical

factors relate to the problems of network connecti-

vity and the differences between the environments of

development and testing. The non-technical factors,

have issues of trust, communication, conﬂict and cul-

tural differences (Damian, 2002). So the challenges

of communication, coordination and control genera-

ted by this approach, arising from the physical dis-

tances and time, demand for studies that present the

guidelines for deﬁning a process of software develop-

ment that is suitable for distributed teams.

This paper presents an analysis of some software

development process from the perspective of dis-

tributed teams, and raises some important points that

should be considered in this development approach.

Thus, some guidelines are established for a dis-

tributed software development process.

2 RESEARCH METHODOLOGY

The research conducted in this work is characterized

as being qualitative basic and exploratory and was

prepared in two steps. In the ﬁrst step was made

an initial review of the literature, aiming to obtain a

consistent theoretical basis for continued study and to

493

Camila Lapasini Leal G., Alberto da Silva C., Hatsue Moriya Huzita E. and Fatima Calvi Tait T. (2010).

TOWARDS A PROCESS TO INFORMATION SYSTEM DEVELOPMENT WITH DISTRIBUTED TEAMS.

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

493-496

DOI: 10.5220/0002909804930496

 SciTePress

have view the state of the art. The second step con-

sisted in the systematic review following the standard

of protocol deﬁned in (Biolchini et al., 2005).

A systematic review was conducted to expand

the coverage of the initial review of the literature.

The main objective was the identiﬁcation of works

that deal with software processes used by distributed

teams. In addition, the review also aimed to identify

the needs of a development process for this develop-

ment approach.

A systematic review allowed to identify the soft-

ware development process which has been used by

distributed teams. Most studies selected were related

to problems (cultural, geographic dispersion, coordi-

nation and communication) found in the use of dis-

tributed development, emphasizing the need to deﬁne

a process that incorporates the features of this new

development approach.

The results provide valuable information for

deﬁning a process that meets the needs of distributed

software development.

3 CHALLENGES IN

DISTRIBUTED SOFTWARE

DEVELOPMENT

This section describes the criteria and analysis of: Ra-

tional Uniﬁed Process (RUP) (Kruchten, 1999), Ex-

treme Programming (XP) (Beck and Andres, 2004),

Agile Uniﬁed Process (AUP)(Ambler, 2008) , Ex-

tended Workbench Model (Paulish, 2007), (Avritzer,

Hasli and Paulish, 2007) and Local Agile Game-based

Process (LAGPRO) (Ribeiro, Czekster and Webber,

2006). After then, some guidelines for deﬁning a pro-

cess considering the challenges of communication,

coordination and control of the distributed develop-

ment approach are also presented (Subsection 3.2).

3.1 Criteria and Process Analysis

The RUP and XP processes were chosen because they

are widely used, and is a traditional methodology and

an agile, respectively. Rocha et. al (Rocha et al.,

2008) presents an experience report about the adapta-

tion of RUP for small distributed development teams.

The AUP was chosen because it is a simpliﬁcation

of RUP and uses the concepts of agile methods. The

Extended Workbench Model and LAGPRO are used

in DSD context, according (Paulish, 2007), (Ribeiro,

Czekster and Webber, 2006), and (Urdangarin, 2008).

A set of benchmarks based on the characteristics

of distributed software development identiﬁed in (Ur-

dangarin, 2008) (criteria A - G), the basic elements

of a software process (criteria H - J ) and the type of

methodology (criterion K and L) were considered to

a comparison among above mentioned software deve-

lopment process. They are described as follow:

• (A) Diversity: identiﬁes whether the process pro-

vides mechanisms for control, coordination and

communication of differences (cultural and lan-

guage) among the teams involved in the project.

• (B) Requirements: analyzes if the process has a

formal speciﬁcation of requirements, aiming at

minimize the problems of ambiguities in the ar-

tifacts and so reduce the communication.

• (C) Centered: identiﬁes if the process has centra-

lized control of project activities.

• (D) Decentralized: identiﬁes if the process is char-

acterized by decentralized control of project activ-

ities.

• (E) Monitoring: identiﬁes if the process support

an adequate control of project developed by re-

mote times.

• (F) Collaboration: checks if there are tools to su-

pport collaboration among teams as a way of sha-

ring information.

• (G) Trust: identiﬁes if the process offers support

to establish trust among members of different

teams.

• (H) Roles: identiﬁes if the process has criteria to

deﬁne roles.

• (I) Artifacts: identiﬁes if are adequately deﬁned

all artifacts produced.

• (J) Activities: identiﬁes if the process deﬁnes the

activities to be carried out in each phase.

• (K) Traditional methodology: verify if the process

focus is a traditional methodology of software de-

velopment.

• (L) Agile methodology: verify if the process has

characteristics of an agile methodology.

Table 1 shows the comparison of the processes in

relation to the criteria above mentioned. Next is pre-

sented an analysis on them.

There is not a process that includes criteria related

to the DSD (criteria A - G) and present the basic ele-

ments (roles, artifacts and activities). Thus, there is a

lack of a systematic process that incorporates the fea-

tures of DSD. Therefore, it is necessary that deﬁned

activities, artifacts and roles required at all stages of

development using the distributed approach were de-

ﬁned.

ICEIS 2010 - 12th International Conference on Enterprise Information Systems

494

Table 1: Comparative analysis of processes.

Process/

Criteria A B C D E F G H I J K L

RUP x x x x x

XP x x x x x x x x

AUP x x x x x

Extended

Workbench

Model x x x x x x x

LAGPRO x x x x x x x x x

RUP, XP and AUP deﬁne the minimum set of ele-

ments that make up a process. However, do not have

procedures to address issues related with DSD prob-

lems, which were represented by the criteria A to G.

Although the Extended Workbench and LAGPRO

have been developed for use in project developed in

distributed way, they haven’t mechanisms to show the

diversity, they talking in account only requirements

formal speciﬁcation. Extended Wokbench does not

mention the factors related to collaboration and trust.

Furthermore, we didn’t found reports about the arti-

facts generated. In LAGPRO case, the activities are

not identiﬁed. Thus, it is possible to show that is ne-

cessary a process that meets these new demands gene-

rated by the distributed approach and presents a deﬁ-

nition of process elements.

Audy et al. (Audy and Prikladnicki, 2008), point

out that in a distributed software development envi-

ronment is fundamental a common development pro-

cess for the team, because a methodology helps di-

rectly in sync, providing all team members a common

tasks and activities naming also, and a common ex-

pectations for all individual involved in the process.

According to Rocha et al. (Rocha et al., 2008),

when the context is Distributed Development, the sce-

nario changes if compared with to traditional soft-

ware development, because the variables and risks

increase. So if there is not a good methodology

for the development process, the project will have a

good chance to not correspond to the initial planning.

In their ﬁrst insight into the study and use of DSD,

Rocha et al. (Rocha et al., 2008) emphasized the need

of more adequate process, since, based on research

carried out was not easy to identify process models for

the DSD. So, it reinforces the need for a development

process that effectively provide adequate support to

characteristics of DSD.

3.2 Guidelines for Software Distributed

Development Process

From the analysis described in Section 3.1 was de-

ﬁned a set of guidelines for the deﬁnition of a soft-

ware process that meets the needs of distributed deve-

lopment, listed as follow:

• adopt hybrid management (centralized-

decentralized) related with to the activities

control;

• encourage communication between development

and integration test teams offering artifacts with

relevant information for them;

• use formal speciﬁcation of tests to mitigate the

problems of ambiguity and to reduce the need for

communication;

• using a component-based architecture for develop

in a distribution way or perform the phase distri-

bution, such as development and testing, thereby

reducing communication among teams;

• support for awareness, so that the people involved

are aware of place and responsibilities assigned to

each team member;

• adopt language with low learning curve and se-

mantics to provide information to developers, for

example, the use of UML because its is known in

both academia and industry;

• provide continuous integration to reduce casual

problems of integration;

• have iterative development and frequent deliveries

to provide greater visibility of project managers;

• deﬁne an infrastructure that enables collaboration,

documentation control and artifact versions con-

trol; develop and apply test suites; and establish

method to control documentation;

• deﬁne a language to formalize the process and in-

teraction among the teams; and,

• deﬁne of a leader to foster trust and commitment

among members.

4 CONCLUSIONS

The physical distribution of teams increases the prob-

lems of development management. Cultural differ-

ences, language, time zone among other things, in-

crease the complexity of communication, coordina-

tion and control during software development.

Using a standardized process for planning, offers

support to project managers by allowing you to set

plans in accordance with the standards and quality

procedures of the organization (Berger, 2003). More-

over, the adoption an appropriate process to the char-

acteristics of organizations, provides a consistent and

predictable development (Kruchten, 1999).

Thus, we analyzed the technical, methods and

software processes found in literature that contributed

TOWARDS A PROCESS TO INFORMATION SYSTEM DEVELOPMENT WITH DISTRIBUTED TEAMS

495

to the understanding of its main activities, actors and

artifacts. Thus, the deﬁnition of a speciﬁc process for

DSD provides a common language to establish roles

and activities, enabling a better understanding of the

business domain terms and goal of the project, de-

spite cultural differences and organizational structure.

Therefore, a process must support distributed deve-

lopment improving communication, reducing misun-

derstanding and uncertainty arising from cultural dif-

ferences and language, reducing redundant activities

and excessive work.

The main contribution of this work is the set of

guidelines presented in Section 3.2, which aims to

reduce the problems of communication, coordination

and control in software development with distributed

teams. As future work, we have: the deﬁnition of a

development process that meets the criteria presented

in Section 3.1, deﬁning the phases, roles, activities

and artifacts. This will reduce the problems of am-

biguity and understanding of the artifacts, setting a

single representation for all involved teams, therefore

allowing them to manage and use collaborative tools,

and thus encourage the establishment of trust among

team members.

REFERENCES

Abdurazik A.; Ammann P.; Ding W.; Offutt, J. Evaluation

of Three Speciﬁcation-Based Testing Criteria. In 6th

IEEE International Conference on Complex Computer

Systems (ICECCS’00), Tokyo, Japan, pp. 179-187,

(2000).

Ambler S. The agile uniﬁed process (AUP), Capturado em

http://www.ambysoft.com/uniﬁedprocess/agileUP.

html, August of (2008).

Audy J.; Prikladnicki R. Desenvolvimento Distribu´ıdo de

Software: Desenvolvimento de software com equipes

distribu´ıdas. Rio de Janeiro: Elsevier, (2008).

Avritzer A.; Hasli NG W.; Paulish D. Process Investigations

for the Global Studio Project Version 3.0. In: Inter-

national Conference on Global Software Engineering

(ICGSE 2007), IEEE Computer Society, Washington,

DC, pp. 247-251, (2007).

Beck, K.; Andres, C. Extreme Programming Explained:

Embrace Change, Cambridge: Addison Wesley Pro-

fessional, 224p, (2004).

Berger P. Instanciac¸˜ao de Processos de Software em Am-

bientes Conﬁgurados na Estac¸˜ao TABA. Dissertac¸˜ao

(Mestrado) - COPPE/UFRJ, Rio de Janeiro, RJ,

(2003).

Biolchini, J. and Mian, P. and Natali, A. and Travassos, G.

Systematic review in software engineering: Relevance

and utility. Technical Report. COPPE/UFRJ, (2005).

Fuggetta A. Software Process: A Roadmap. In: Interna-

tional Conference on Software Engineering, Proceed-

ings of the Conference on The Future of Software En-

gineering, (2000).

Damian, D. Workshop on Global Software Development.

In: International Conference on Software Engineering

(ICSE’02), IEEE Computer Society, USA, pp. 19-25,

(2002).

Kruchten, P. The Rational Uniﬁed Process - An Introduc-

tion. Addison-Wesley, Boston, USA, (1999).

Paulish, D. J. Scalable Distributed Organizations for Ultra-

Large-Scale Software. In: International Workshop on

Software Technologies for Ultra-Large-Scale Systems

(ULS’07), pp. 4-4, (2007).

Ribeiro, M.B.; Czekster, R. M.; Webber, T. Improving Pro-

ductivity of Local Software Development Teams in a

Global Software Development Environment. In: In-

ternational Conference on Global Software Engineer-

ing, IEEE Computer Society, pp. 253-254, (2006).

Rocha, R.; Arcoverde, D.; Brito, R.; Arˆoxa, B.; Costa,

C.; Silva, F. Q. B.; Albuquerque, J.; Meira, S. R.

L. Uma Experiˆencia na Adaptac¸˜ao do RUP em Pe-

quenas Equipes de Desenvolvimento Distribu´ıdo. In:

II Workshop de Desenvolvimento Distribu´ıdo de Soft-

ware, Campinas, SP, pp. 81-90, (2008).

Urdangarin, R. G. Uma Investigac¸˜ao sobre o Uso de

Pr´aticas Extreme Programming no Desenvolvimento

Global de Software. Dissertac¸˜ao (Mestrado) - Fac-

uldade de Inform´atica, PUCRS, Porto Alegre, RS,

(2008).

ICEIS 2010 - 12th International Conference on Enterprise Information Systems

496