INVESTIGATING THE ROLE OF UML IN THE SOFTWARE

MODELING AND MAINTENANCE

A Preliminary Industrial Survey

Giuseppe Scanniello

Dipartimento di Matematica e Informatica, University of Basilicata, Viale Dell'Ateneo, Macchia Romana, Italy

Carmine Gravino, Genny Tortora

Dipartimento di Matematica e Informatica, University of Salerno, Via Ponte Don Melillo Fisciano, (SA), Italy

Keywords: Software Development, Software Maintenance, UML, Survey.

Abstract: In the paper we present the results of an industrial survey conducted with the Italian software companies

that employ a relevant part of the graduate students of the University of Basilicata and of the University of

Salerno. The survey mainly investigates the state of the practice regarding the use of UML (Unified

Modeling Language) in the software development and maintenance The results reveals that the majority of

the companies use UML for modeling software systems (in the analysis and design phases) and for

performing maintenance operations. Moreover, maintenance operations are mainly performed by low

experienced practitioners.

1 INTRODUCTION

In the software engineering community, there is a

growing interest in the use of empirical studies

(Basili, 1993) (Kitchenham et al., 2002) such as

controlled experiments, qualitative studies, surveys,

and archival analyses to investigate processes,

methods, and tools for software development and

maintenance. Among these kinds of studies, a survey

is an investigation performed in retrospect (Wohlin

et al., 2000) to gather qualitative and quantitative

data very often using questionnaires that are

designed through taking a sample which is

representative of the population to be studied. The

survey results can be analyzed to derive descriptive

and conclusions that are only applicable to the

population where the sample has been taken.

In order to investigate the state of the practice

regarding the use of UML (Unified Modeling

Language) in the software development and

maintenance we have conducted an explorative

survey (Wohlin et al., 2000) among the Italian

software companies that employ the greater part of

the people with a Master’s or a Bachelor’s degree

from either the University of Basilicata or University

of Salerno. The companies that host students for

external stages have been involved as well. All the

invited companies (i.e., the target population of the

study) represent our industrial contact network.

The research questions the presented study aims

at answering for the selected population are:

[RQ1] What is the relevance of UML?

[RQ2] What is the used development process

model?

[RQ3] Is UML used in the early phases of the

software development?

[RQ4] Is UML used to support software engineers in

the execution of maintenance operations?

[RQ5] Which software models (i.e., analysis and/or

design) are provided to the software

engineers to perform maintenance

operations?

[RQ6] Which is the more common kind of software

maintenance operation?

[RQ7] Which is the mean effort required to perform

maintenance operations?

In order to answer the above research questions

the following subsequent steps have been

accomplished:

141

Scanniello G., Gravino C. and Tortora G. (2010).

INVESTIGATING THE ROLE OF UML IN THE SOFTWARE MODELING AND MAINTENANCE - A Preliminary Industrial Survey.

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

141-148

DOI: 10.5220/0002870601410148

 SciTePress

1. design of a questionnaire to investigate the

research questions;

2. conduction of the survey, leveraging the relative

industrial contact network;

3. analysis of the results.

The design and the results of the survey are

present and discuss in this paper. In particular, the

remainder of the paper is organized as follows.

Section 2 provides the notions and concepts we used

in the paper and presents the design of the survey as

well. Section 3 describes the results achieved, while

general findings and threats to validity are reported

in Section 4. Final remarks and future work

conclude the paper.

2 DEFINITION AND DESIGN

The survey goals can be summarized as follows:

- Primary Goal: comprehending the relevance of

UML for our industrial contact network

(employing the greater part of the students from

the Universities of the authors).

- Secondary Goal: identifying the core business

activity of the interviewed industries between

systems design and development and software

maintenance.

2.1 Conceptual Model

The conceptual model clarifies the meaning of some

terms and describes all the entities of interest for the

conducted survey.

- Project: a completed software development

project.

- Development Process Model: a structure

imposed on the development of a software

product (Pressman, 2005).

- Software Artefact: a tangible product created

during the software development process. It can

be created in the early and final stages of

software development (Pressman, 2005).

- Requirements Analysis: a subdiscipline of

systems engineering and software engineering

that is concerned with determining the goals,

functions, and constraints of hardware and

software systems (Ciolkowski et al., 2003,

Laplante, 2009).

- System Design: a process to define the

architecture, components, modules, interfaces,

and data for a software system to satisfy

specified requirements.

- Software Maintenance: the modification of a

software product after delivery to correct faults,

to improve performance or other attributes, or to

adapt the product to a modified environment

(ISO/IEC, 2006) .

- Ordinary Maintenance: reactive modifications

of a software product performed after delivery

to keep a software product usable, to improve

performance or maintainability, and to detect

and correct latent faults.

- Extraordinary Maintenance: modifications of

a software product performed after the delivery

phase to correct discovered problems.

- UML: a formalism to specify, visualize, and

document models of software systems,

including their structure and design (UML,

2009).

We identified three areas of interest where we

wanted to collect data:

- Demographic Information: we collected

demographic information about the respondents

and their organizations.

- Relevance and Typology: we collected

information about the relevance (i.e., number of

organizations that use UML with respect to the

complete sample) and quantity (i.e., the

number) of projects where UML is used to

develop and maintain software products. Great

care was taken to collect information about the

UML notations mainly used in the software

development and maintenance.

Demographic information includes the respondent’s

organization. In particular we collected: business

domain, size of the organization and respondent’s

group/business unit, typical duration and kind of

projects, average experience and skill of group

members. In addition, we collected information

about the respondents themselves, such as: age,

gender, educational qualification, role in the

organization (e.g., IT manager, Project manager).

To evaluate the company relevance the

respondents were also asked to specify whether their

organizations use UML for software development

and maintenance. Accordingly, we divided the

respondents’ organizations in the following groups,

which depend on whether the organizations have:

ICEIS 2010 - 12th International Conference on Enterprise Information Systems

142

- (Group 1) used UML for modelling software

systems in the software development and for

performing maintenance operations;

- (Group 2) used UML only in the early phases

of the software development (i.e., analysis and

design);

- (Group 3) used UML for performing

maintenance operations;

- (Group 4) used UML neither in software

development nor maintenance.

For the first and second group we asked the

typical size of a developed software product and the

adopted development process model. For the second

and the third group, questions were also asked on the

employees involved in the software development

and maintenance, respectively. These questions were

also asked to the respondents of group 1. For the

organizations of the fourth group the questionnaire

finished after collecting the demographic

information.

2.2 Identification of the Target

Population of the Sample

The target population consisted of decision makers

in software developing organizations. Indeed, we

considered ICT organizations that develop, sell, and

maintain software as a main part of their business

(e.g., software house) or develop software as an

integral part of their products or services (e.g.,

commerce in the healthcare domain).

The selection of the organizations (sampling) has

been conducted using the network contacts of the

groups of the researchers who conducted the survey.

Within this network there were the companies that

host students for external stages and employ the

majority of the persons with the Master or the

Bachelor degree in Computer Science from the

University of Salerno and the University of

Basilicata.

2.3 Questionnaire Design and Data

Collection

We have developed the questionnaire following the

standard schema proposed by Ciolkowski et al.

(2003). Figure 1 shows the design of the defined

questionnaire, which according to the conceptual

model consists of four different paths. The

questionnaire was divided in 3 sections and was

structured such that the total number of questions

depended on whether the respondent’s organization

uses UML to develop and maintain software

systems. The first part (Section 1), common to all

paths, was used to get information on both the

respondent and his/her company (i.e., demographic

information). The questions of Section 1 were 12,

while Section 2 included 8 questions. Finally,

Section 3 included 9 questions. The questionnaire

mainly contained closed questions. Only a few

questions required filling in numbers or text.

Organization

activity

Respondent and

organization data

Maintenance

Information on the

Projects

Development

Information on the

Projects

Only Maintenance

Development

Other

Section 1

Maintenance as well

yes

Section 3

Section 2

Figure 1: Designed questionnaire.

The questionnaire was introduced with a brief

motivation sketching the general problem of the

investigation. The importance of this study and our

objectives were inserted in an accompanying letter

attached to the questionnaire and sent to the

respondents. Furthermore, we clarified that all the

information was considered confidential and that the

data were used only for research purposes and

revealed only in aggregated form. Great care was

also taken to ensure ethical requirements and privacy

rules imposed by the Italian regulations (these were

inserted at the end of the questionnaire).

We collected the questionnaire via e-mail. The

main drawback of collecting the data in such a way

is that information concerning the time to fill in the

questionnaire is not available. However, this

information may consider useless for the study

presented here.

2.4 Survey Preparation, Execution and

Analysis Procedure

The survey has been performed according to the

INVESTIGATING THE ROLE OF UML IN THE SOFTWARE MODELING AND MAINTENANCE - A Preliminary

Industrial Survey

143

following four activities:

1. Preparation and Design of the Questionnaire.

We used similar questionnaires to identify a set

of questions.

2. Invitation to Participate. The organizations of

the industrial contact network were invited by

phone (to ask for their availability) and then an

e-mail was sent to provide detailed instructions

on the state of the practice survey and its aim.

The questionnaire and the information related to

it were attached to this e-mail.

3. Collecting. One of the authors collected the

filled in questionnaire that the respondents

returned by e-mail.

4. Analysing. The answers of the collected

questionnaire were analyzed with respect to the

goals of the industrial survey.

We collected 22 questionnaire correctly filled in.

Note that the invited company were 53, thus

obtaining 42% as response rate.

3 ANALYSIS

Due to the nature of our study the analysis is mainly

based on basic descriptive statistics. In the following

we summarize the obtained results according to the

design of the developed survey questionnaire (see

Section 2.3).

3.1 Respondents’ Background and

Companies’ Characteristics

The respondents’ age ranges from 28 to 56 years old

with an average of 41 years. The female are only 5

and the majority of respondents work for a company

located in the South of Italy (15). The companies are

geographically distributed as follows: 16 are from

the south of Italy, 4 are from the central, and 2 are

from the north. Note that southern Italy is better

represented, also because the universities that

conducted the survey are located in southern Italy.

On the other hand, further information on the

respondents and their companies are summarized in

Table 1.

Regarding the role of the respondents, 15 of

them stated that they are project managers, while the

others declared to be developers (2) or architects (4).

Only one respondent involved in the study is a

company administrator. Concerning the education, 3

respondents declared to have a Bachelor degree,

while 18 respondents have a Master degree, and 1

only one respondent has a first level Master’s

degree. None has a PhD. As expected, the majority

of the respondents (17) has a specific IT degree. 14

companies are independent, while 8 are controlled

by a bigger and more powerful company. Among

those companies, 13 are S.p.A, (i.e., companies

quoted on the Italian stock exchange), while 9 are

private companies. With respect to the size of the

companies and according to the recommendation

2003/361/EC (micro companies have less than 10

employees; small companies have from 10 to 50

employees; medium companies have from 50 to 250

employees; large companies have more than 250

employees), 13 of the survey responses come from

small and medium-sized companies and 8 come

from large ones. Only one is a micro company.

Regarding the business units, we have 4 micro, 8

small, and 9 medium sized units. Only 1 can be

considered a large business unit.

Respondents also provided information about the

experience and the education of the employees. The

analysis reveals that on average 11% of the

employees have less than one year of experience,

52% of the employees have less than five years of

experience, and 37% of the employees have more

than five years of experience.

Regarding the education, on average 29% of the

employees have a Bachelor degree, 44% of the

employees have a Master degree, and 27% of the

employees have a High School diploma.

Information on the employees hired during the

last year was also specified by the respondents. In

particular, on average the percentage of employees

with an ICT Bachelor degree hired every year is

40%, while 45 is the percentage of employees with

an ICT Master degree. 15% is the percentage of

employees without an ICT degree on average hired.

Moreover, on average the 68% of the employees

hired during last year had no previous experience in

the ICT field. The remaining (32%) have on average

3 years of experience in the ICT field.

Regarding the industrial domains, about 10 work

in the area of software consultancy. The same

number of companies works as software

house/vendors, while only 2 companies declared to

be Web agency.

Concerning the main activities of the involved

companies, 20 focus on the development and

maintenance of software systems. A company only

focuses on software development.

ICEIS 2010 - 12th International Conference on Enterprise Information Systems

144

Table 1: Section 1 answers (Respondent and Organization Data).

QUESTION

ANSWER

Respondent information

[1.1.1] Role in the business

unit

IT manager Project manager Software architect Software developer Other (role)

0 15 4 2

1 (company

administrator)

[1.1.2] Education level

Bachelor degree Master degree

First level Master’s

degree

Second level

Master’s degree

PhD/Doctorate

3 18 1 0 0

[1.1.3] Qualification in ICT

field

Yes No

17 5

Company information

[1.2.1] Type of Company

Self-supporting Subsidiary

14 8

[1.2.2] Type of Company

Stock company Company controlled by a private entity Governmental organization

13 9 0

[1.2.3] Number of employee

< 10 [10-50] [50-250] > 250

1 3 10 8

[1.2.3] Number of employee

of the business unit

< 10 [10-50] [50-250] > 250

4 8 9 1

[1.2.4] Employee experience

[0 – 1] years [1 – 5] years > 5 years

11% 52% 37%

[1.2.5] Employee Degree

Bachelor degree Master degree High Scholl Diploma Other

29% 44% 27% 0%

[1.2.6] Education level of

employee hired in the last

year

ICT Bachelor degree ICT Master degree

High Scholl

Diploma

Other

40% 45% 15% 0%

[1.2.7] Experience of

employee hired in the last

year

No experience in ICT field Experience in ICT field (average number of year)

68% 32% (3)

Company activity information

[1.3.1] Company is

Software house/

Software vendor

Software/IT consultancy Other

10 10 2

[1.3.2] The main activity of

the company is

Software development and maintenance Software development

Software

maintenance

Other

20 1 0 1

3.2 Development Activity

The analysis of the answers of Section 2 (see Table

2) reveals that the greater part of the companies (14)

usually develop software systems whose sizes range

from 10,000 to 100,000 LOCs. Software systems

with sizes ranging from 100,000 to 500,000 LOCs

are developed by 5 companies. On the other hand,

the size of the software systems developed by the 2

remaining companies is less than 10,000 LOCs and

greater than 500,000 LOCs, respectively.

Regarding the development process model, 4

respondents indicated that the waterfall model is

typically used, while 4 respondents specified that

agile methodologies are employed. The remaining

13 respondents selected the Unified Process as

widely used development process model used in

their companies.

On average 18% of the employees has less than

one year of experience, 51% has less than five years

of experience. 32% of the new employees have more

than five years of experience. Regarding the

education, on average 42% of the employees has a

Bachelor degree, 42% has a Master degree, and 16%

has a High School diploma.

All the respondents indicated that UML is

typically used during the development phase.

Indeed, on average UML has been employed in the

74% of the completed software projects. In the

remaining 26% of the projects UML was not used.

Concerning the employed UML diagrams,

almost all companies typically employ Use Case

Diagrams (19), Class Diagrams (20), and State

Diagrams (17) to model software systems. The

Component Diagrams and the State Diagrams are

instead used by 14 and 13 companies, respectively.

On the other hand, 8 companies use the Activity

Diagrams.

The respondents were also asked to specify the

UML diagrams adopted in the analysis and the

design phases. The analysis of the answers has

revealed that the diagrams more used in the analysis

phase are the Use Case diagrams (13), the Class

Diagrams (10), and the State Diagrams (13).

However, the Sequence Diagrams (5) and the

Activity Diagrams (8) are used as well. As expected

INVESTIGATING THE ROLE OF UML IN THE SOFTWARE MODELING AND MAINTENANCE - A Preliminary

Industrial Survey

145

Table 2: Section 2 answers (Development Information on the Projects).

QUESTION ANSWER

[2.1] Size of developed software

system

<10,000 LOC [10,000 - 100,000] LOC [100,000 - 500,000] LOC > 500,000 LOC

1 14 5 1

[2.2] Development process

model

Waterfall Unified Process Agile Methodologies Other

4 13 4 0

[2.3] Employee Degree

Bachelor

degree

Master degree High Scholl Diploma Other

42% 42% 16% 0%

[2.4] Employee experience

< 1 year [1 – 5] years > 5 years

18% 51% 32%

[2.5] Use of UML

Yes Percentage of software systems where UML has been used

20 74%

[2.5.1] UML diagram usually

employed

Use Case

diagram

Class

Diagram

Component

Diagram

Sequence

Diagram

Activity

Diagram

Deployment

Diagram

State

Diagram

19 20 14 13 8 0 17

[2.5.2] UML diagram used

during analysis phase

Use Case

diagram

Class

Diagram

Sequence

Diagram

Activity

Diagram

State

Diagram

13 10 5 8 13

[2.5.3] UML diagram used

during design phase

Class

Diagram

Component

Diagram

Sequence

Diagram

Deployment

Diagram

19 14 10 0

for the design phase the widely employed notations

are the Class Diagrams (19), the Component

Diagrams (14), and the Sequence Diagrams (10).

3.3 Maintenance Activity

Table 3 summarizes the information regarding the

maintenance operations performed by the

respondents’ organizations (i.e., the answers of the

questions of Section 3). In particular, the

respondents specified that 64% of systems on which

their companies perform maintenance operations are

Objected-Oriented applications. Furthermore, the

analysis of the answers revealed that 13 companies

usually perform maintenance tasks on software

systems developed by other companies. 11

respondents indicated that the typical maintenance

operations performed by their companies are

ordinary maintenance, while the remaining 9

companies usually carry out also extraordinary

maintenance operations.

Respondents also provided information about the

experience and the education of the employees

performing maintenance tasks. Regarding the

education, on average 43% of the employees has a

Bachelor degree, 26% has a Master degree, and 31%

has a High School diploma. The analysis revealed

that on average 17% of the employees has less than

one year of experience, 55% has more than 1 year

and less than five years of experience, and 28% has

more than five years of experience.

Moreover, 14 respondents declared that a typical

ordinary maintenance operation requires less than 5

hours, while 5 respondents specified that less than

10 hours are required to perform this kind of

maintenance operation. On the other hand, the

majority of respondents (17) answered that a typical

extraordinary maintenance operation requires from

10 to 50 hours. The remaining companies need less

than 10 hour to perform a typical extraordinary

maintenance operation.

Regarding the documentation exploited to

perform extraordinary maintenance tasks, the

majority of the respondents (14) specified that both

analysis and design documents are usually used. It is

worth to noting that 3 respondents declared that only

source code is usually exploited to perform ordinary

maintenance operations. The remaining respondents

answered that only analysis documentation is used

by the companies to perform ordinary maintenance.

The results for extraordinary maintenance operations

are nearly the same.

4 RESULTS

In the following subsections we summarize the main

findings of the study and highlight the threats that

may affect the obtained results.

4.1 General Findings

The analysis presented in the previous section has

revealed that the majority of the companies (20) use

UML in the analysis and design phases. In

particular, the respondents declared that UML was

employed in 74% of the completed projects in order

to model objected-oriented software systems.

Moreover, UML is also employed to carry out

maintenance tasks since 75% of the respondents

revealed that analysis and design models, produced

during the software development process, are widely

ICEIS 2010 - 12th International Conference on Enterprise Information Systems

146

Table 3: Section 3 answers (Maintenance Information on the Projects).

QUESTION ANSWER

[3.1] Maintenance operations

performed on

O-O System Other

64% 36%

[3.2] Maintenance operations

performed on system developed by

other companies

Yes No

13 7

[3.3] Typical maintenance operations

Ordinary Extraordinary

11 9

[3.4] Employee Degree

Bachelor

degree

Master degree High Scholl Diploma Other

43% 26% 31% 0%

[3.5] Employee experience

[0 – 1] years [1 – 5] years > 5 years

17% 55% 28%

[3.6] Typical ordinary maintenance

operations takes

< 1

person/hours

[1 – 5] person/hours [5 – 10] person/hours > 10 person/hours

0 14 5 0

[3.7] Typical extraordinary

maintenance operations takes

< 10

person/hours

[10 – 50] person/hours [50 – 100] person/hours > 100 person/hours

3 17 0 0

[3.8] Documentation used for ordinary

maintenance

Analysis and

design models

Analysis models Design models

Nothing (only source

code)

14 3 0 3

[3.9] Documentation used for

extraordinary maintenance

Analysis and

design models

Analysis models Design models

Nothing (only source

code)

15 4 0 1

used by the company maintainers. Maintenance

operations are mainly performed by low experienced

maintainers. In fact, the respondents revealed that

72% has less than 5 years of professional

experience. Furthermore, on average 43% has a

Bachelor degree in Computer Science. Employees

with a Master degree in Computer Science are less

frequently used to maintain existing software

systems (on average 26% of maintainers).

Furthermore, the average effort to perform

maintenance operations ranges from 1 to 5 person

hours for an ordinary maintenance operation for

70% of the companies. On average, 85% of the

companies spend from 10 to 50 person hours in case

of extraordinary maintenance operations.

4.2 Threats to Validity

In the following the threats to validity (i.e., internal,

external, and construct) that could affect the

conducted study are presented and discussed in

increasing priority order.

Internal validity threats regard external factors

that may affect the observed results. In industrial

surveys, it is usually impossible to know whether the

respondents answer truthfully, or whether other

effects may bias the results. Also, the respondents’

motivations could affect the answers and then the

survey results. Another factor that may have

influenced this validity is the number of invited

organizations that did not answer the questionnaire

(less than 50%). Another negative factor could be

the difficulty to comprehend the questions (e.g.,

ambiguous, not clear, not well formulated).

However, we designed the questionnaire to

minimize this problem. The reader may also object

to the fact that the organizations within our

industrial network may influence the internal

validity as well. However, one of the goals of this

survey concerned the study of the productive reality

where the greater part of the students graduated at

the University of Basilicata and at the University of

Salerno are employed.

External validity threats concerns the

generalization of the results. These threats are

always present in case of industrial survey. For our

survey, although the companies belonged to a

variety of domains and covered small, medium and

large company from north, centre, and south of Italy,

we cannot be sure that our sample is representative

of ICT Italian industry, in general.

The metrics used in the study as well as the

designed questionnaire may threaten the construct

validity. In our case, the questionnaire was designed

using standard ways and scales (Ciolkowski et al.,

2003). Furthermore, the questions were formulated

to minimize possible ambiguities.

Note that due to the nature of the conducted

study, it is only through replications that we can gain

more confidence in the survey results. Accordingly,

we have planned to extend our investigation

involving industries that do not belong to our

industrial contact network.

INVESTIGATING THE ROLE OF UML IN THE SOFTWARE MODELING AND MAINTENANCE - A Preliminary

Industrial Survey

147

5 FINAL REMARKS AND

FUTURE WORK

This study revealed a set of interesting outcomes

that can be summarized as follows. The greater part

of the interviewed companies uses UML for

modelling software systems (in the analysis and

design phases) and for performing maintenance

operations. Maintenance operations are mainly

performed by low experienced practitioners. In fact,

the companies generally use professionals, who have

a Bachelor degree in Computer Science and have

less than 5 years of experience. Professionals with a

Master degree in Computer Science and with less

than five years of experience are also used.

However, they are less frequently used for

maintaining existing software systems with respect

to persons with a Bachelor degree.

Another interesting point concerns the average

effort to perform maintenance operations. It ranges

from 1 to 5 person hours for an ordinary

maintenance operation (e.g., corrective changes),

while it is from 10 to 50 person hours in case of an

extraordinary maintenance operation (e.g., perfective

or adaptive changes).

The results of this study represent an interesting

starting point for further empirical investigations. In

fact, we have been designing and conducting

controlled experiments to assess the effectiveness of

UML in the requirements engineering process and in

the software maintenance as well.

We are also going to encourage replications to

increase the confidence in the achieved results. This

will increase our body of knowledge on the role of

UML in the software development and maintenance.

In fact, at the best of our knowledge the presented

study is the first investigation on the UML usage in

the Italian software industry regarding the use of

UML in the software maintenance. Differently,

Dobin and Parsons (2006) conducted a survey on

how the UML diagrams are used in the analysis

phase. The authors in the survey did not consider

both the develop process used within the interviewed

companies and the role of UML in the execution of

maintenance operations.

Concluding the survey results also positively

answer an interesting question from the pedagogical

point of view: “Are the students that took a Bachelor

or a Master degree either at the University of

Basilicata or at the University of Salerno ready to be

employed in the industries of our contact network?”

ACKNOWLEDGEMENTS

We want to thank all the companies that participated

to the survey.

REFERENCES

Basili V. R., “The Experimental Paradigm in Software

Engineering”, Proc. International Workshop,

Experimental Software Eng. Issues: Critical

Assessment and Future Directions, LNCS 706,

Springer-Verlag, 1993.

Ciolkowski, M. Laitenberger, O., Biffl, S.,. “Software

reviews: The state of the practice”. In IEEE Software,

vol. 20, no 6, 2003, pp. 46 51.

Dobing B., Parsons J., “How UML is used”; In

Communications of the ACM, vol. 49, no 5, 2006, pp.

109-113.

Kitchenham B. A., Pfleeger S.L., Pickard L.M., Jones

P.W., Hoaglin D. C., El Emam K., Rosenberg J.,

“Preliminary Guidelines for Empirical Research in

Software Engineering”, IEEE Transactions on

Software Engineering, vol. 28, no. 8, pp. 721–734.

UML: Introduction to OMG's Unified Modeling

Language™ available, 2009 http://www.omg.org/

gettingstarted/what_is_uml.htm.

ISO/IEC 14764:2006 available at http://www.iso.org/

Laplante, P., Requirements Engineering for Software and

Systems (1st ed.). Redmond, 2009: CRC Press. ISBN

1-42006-467-3.

Oppenheim, N., Questionnaire Design, Interviewing and

Attitude Measurement, Pinter Publishers, 1992

Pressman R. S., Software Engineering: A Practitioner's

Approach, Sixth Edition, United States, McGraw Hill,

2005.

Wohlin C., Runeson P., Host M., Ohlsson M. C., Regnell

B., Wesslen A., Experimentation in Software

Engineering - An Introduction, Kluwer, 2000.

ICEIS 2010 - 12th International Conference on Enterprise Information Systems

148