MODELING WORK PROCESSES AND SOFTWARE

DEVELOPMENT

Notation and Tool

Andre L. N Campos and Toacy C. Oliveira

Instituto Alberto Luiz Coimbra de Pós-Graduação em Engenharia (COPPE), Federal University of Rio de Janeiro (UFRJ)

Caixa Postal 68.511, CEP 21941-972, Rio de Janeiro, RJ, Brazil

Keywords: Process modelling, BMPN, ARIS, SPEM, IDEF0, Modeling notations, Modeling tools.

Abstract: Process modeling are becoming more essential to the activities of acquisition and development of systems.

There are a number of possible notations and tools for process modeling, and sometimes it isn’t an easy

choice. This papers try, through research in a real environment, identify selection criteria and recommend

the most appropriate notations and tools for process modeling.

1 INTRODUCTION

The process modeling, over the last decade, has

become an important mechanism for understanding

the dynamic behaviour of organizations. This

mechanism appears to be superior to techniques

previously used by professionals in information

systems, and can increase the success of technology

implementation projects in dynamic environments

(Green & Rosemann, 2000). Experts in the process

engineering field and in the area of information

technology, benefit from a better understanding of

organization’s processes. The conceptual modeling

of these processes facilitates the development of

appropriate software to support that process (Agular-

Savén, 2003).

In fact, the modeling process is essential to

represent complex systems, being of great

importance in defining the rules of integrated

information systems to business processes. Actually,

the process modeling is the key element to achieve

high quality models that can support the creation of

effective information systems (Ortiz-Hernández,

Nieto-Ariza, Estrada-Esquivel, Rodrígues-Ortiz, &

Montes-Rendon, 2007). Bargis (Bargis, 2008) agrees

that the current software is more than mere artifacts.

According to him, it is a complex phenomenon

techno-social, and that complexity just can be

addressed through the modelling process, as the

basis of a model-centric software development

processes.

Thus, even before the research in modelling and

software development processes themselves, this

paper seeks to identify a notation for the wide

process modeling, which can be easily understood

by humans, and yet, which has the potential to

generate information processable by machines.

2 MODELING NOTATIONS

There are several notations or methods available for

modeling processes, and many other tools that

support these notations (Gartner Group, 2008).

Some of the tools support more than one notation.

This work focuses on four notations of process

modeling, considering its relevance to the topic or its

consolidation in the market and in academia. The

notations are: ARIS BPMN, SPEM and IDEF0.

The presence of the ARIS notation in this work

is justified by its wide use, having been the object of

several studies (Araujo, et al., 2004) (Scheer, 2003)

(Santos Jr, Almeida, & Pianissolla, 2008) (Souza, et

al., 2009). In addition, consecutive surveys by the

Gartner Group indicated that notation as the market

leader (Gartner Group, 2008), which can be seen in

Figure 1.

The notation ARIS (Architecture for integrated

Information Systems) is a framework for enterprise

modeling. This notation can describe your

organization's information architecture and

integration through its processes, and different

organizational views can be described. To

337

L. N Campos A. and C. Oliveira T..

MODELING WORK PROCESSES AND SOFTWARE DEVELOPMENT - Notation and Tool.

DOI: 10.5220/0003581403370343

In Proceedings of the 13th International Conference on Enterprise Information Systems (ICEIS-2011), pages 337-343

ISBN: 978-989-8425-55-3

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

Figure 1: Mains tools of process mapping (Norton,

Blechar, & Jones, 2010).

visualize the workflow diagrams are used Event-

Driven Process Chain, or simply EPCs, diagrams are

used. Besides the description of activities and its

flow, the notation defines roles or profiles, and

information produced and used in the process

(Santos Jr, Almeida, & Guizzardi, 2010). The main

elements of an EPC are seen in Table 1.

Table 1: Elements of the EPC.

Element Symbol

Event

Function

Conectors

Flow

Interface

Another important notation is the IDEF0. It was

prepared at the request of the Air Force in the 1970s,

and built from a well-defined graphical language

known as SADT (Structured Analysis and Design

Technique). Widely used in the United States and

some European countries, the IDEF0 allows us to

represent activities, flows, support resources, and

controls that take action on the activity. The method

is basically composed of boxes and arrows, and the

location of the box where the arrow connects

indicative of its kind: input, output, or resource

control (Soung-Hie & Ki-Jin, Desiging performance

analysis and IDEF0 for enterprise modelling BPR,

2000), as can be seen in Figure 2.

Figure 2: Meaning of the elements of the IDEF0.

An IDEF0 model consists of a hierarchical series

of diagrams that show progressive levels of detail,

expanding knowledge about the process and

describing their functions and interfaces within a

system. The contribution of the modeler is

important. He must have an abstract view of the

model, which can be difficult in large systems

(IEEE, 1998) (Soung-Hie & Ki-Jin, Designing

performance analysis and IDEF0 for enterprise

modelling in BPR, 2002).

The SPEM (Software and Systems Process

Engineering Metamodel), in turn, was considered in

this study because it is a notation created for the

domain of software development field. It was

adopted by the OMG (Object Management Group)

in 2002 and since then it has become as an official

standard of the organization. It is a meta-model that

defines the standard stereotypes UML (Unified

Modeling Language) for modeling software

processes. In fact, SPEM is a network of concepts

and relationships that are expressed by existing

stereotypes in UML, but they could also be

described by any sufficiently expressive language,

perhaps even English, Spanish or Portuguese

(Gonzales-Perez & Henderson-Sellers, 2007). Some

of the stereotypes can be seen in Table 2 (Abdala,

Lahoz, & Sant'Anna, 2003).

Another important notation is the Business

Process Model and Notation (BPMN). Business

Process Management Initiative (BPMI) published

specification for version 1.0 in May 2004, joining

the OMG in June 2005 with the goal of building a

standard process modeling to become the main

resource in the field. Currently, BPMN is in its

version 2.0 (OMG, 2010).

The model BPMN defines a process diagram that

is based on flow charts (flowchart) for the

description of work operations. This is a graphical

network composed mainly of activities (work) and

arrows (flow). The purpose of the notation is to be

easily understood, but also able to generate

information for the execution environments of

ICEIS 2011 - 13th International Conference on Enterprise Information Systems

338

Table 2: Main stereotypes of SPEM.

Element Symbol

WorkProduct

WorkDefinition

Guidance

Activity

ProcessPerformer

ProcessRoler

Documento

processes, for example, through Business Process

Execution Languate (BPEL). In fact, it is part of the

design of the model to create a bridge between the

work processes and the implementation of these

processes (White, 2004).

While BPMN has a wide range of representation

elements, a few elements are sufficient to represent

most of the work processes, which includes the

processes of software development. These elements

are presented in Table 3.

Previous works are related to this research. For

example, Macedo and Schmitz (Macedo & Schmitz,

2001) sought to find the best tool for modeling

software processes, but with the goal to tie in the

software development and the business processes.

They evaluated the tools Aris Toolset and Provision

Workbench. They considered aspects such as the

functional architecture, process modeling method,

level of detail of procedures, ease of use of the

editor, semantic checker, among others. However,

there was no score for these aspects and their work

does not recommend a tool. Still, the notations were

not considered further.

Similar work (Benedictis, Amaral, & Rozendfeld,

2003) conducted a survey of notations Aris and

IDEF0. Tools that would support these notations

were also evaluated. The purpose of this research

would identify the appropriate notation and tool for

modeling work processes with emphasis on product

development. In terms of tool, they analyzed the

Aris Toolset, Microsoft Visio and Microsoft

PowerPoint. After reviewing the two notations and

three tools, the paper concludes by making only

general considerations. We must also consider that

Microsoft Visio and Microsoft Poewerpoint are

general purpose tools and not specifically for

process modeling.

Table 3: BPMN object model.

Element Symbol

Event

Activity

Gateway

Sequence flow

Message flow

Association

Pool

Lane

Data

Group

Anotation

3 EVALUATION OF NOTATIONS

The research objective was to identify a notation that

is understandable by humans and executable by

machines. An experiment was conducted in a real

environment, the Oswaldo Cruz Foundation, an

important public health research institution in Brazil.

This was done in two steps: 1) an analysis performed

by process modelling professionals, 2) an analysis

by regular users, from several areas of the

organization.

With the goal of using the same framework for

analysis both by professionals and regular users, we

chose a generic process as an example. The

hypothetical process used was the issue airline

MODELING WORK PROCESSES AND SOFTWARE DEVELOPMENT - Notation and Tool

339

tickets, originally described in Portuguese.

The notations considered in this study were

evaluated by four process modelling professionals.

For the evaluation we established a set of criteria,

which consolidated proposals from other studies

(Macedo & Schmitz, 2001) (Benedictis, Amaral, &

Rozendfeld, 2003), and added criteria in order to

broaden the scope of the research. The criteria can

be seen in Table 4. Each modeler built a model from

the reference process, using a different notation,

namely ARIS, BPMN, SPEM and IDEF0.

Table 4: Criteria for analysis of models by the team of

modelers.

Criterion Detailing

It has well-defined syntax

The objects are related through rules that

must be obeyed, then, the method sets the

proper connections between the objects.

It has well-defined semantics

There are specific objects that have

different meanings, and cannot be used

for a different representation than that for

which it was created.

It is widespread

The method is widely used in academia

and / or marketing.

It is easy to learn

The method is easy to learn and can be

understood without further knowledge

and skills by the user.

It is easy to read and

understand

The processes modeled based on the

method can be easily understood by most

people, even with little or no prior

training about the method.

Represents activities

Activities performed by human beings

are representable with the method.

Represents information

Information, data or documents can be

represented with the method.

Represents resources

Resources such as technology, systems,

and others, can be represented with the

method.

Represents integration between

processes

The coupling between processes, ie the

integration between them, can be

represented with the method.

Represents the actors who

perform the activities

People, positions and areas (departments)

can be represented with the method, and

related to the activity they perform.

Represents the execution of an

activity for more than an actor

The same activity may be related to two

performers at the same time.

Represents system

requirements

System requirements that would support

an activity can be represented, or at least

indicated with the notation ..

It is extensible

The notation provides its own extension

by adding new objects.

After the process modeling an interview was

conducted with modelers where each one presented

the model built, spoke of the difficulties and ease of

use, and responded to questions raised by the other

modelers. Then, each modeler has a complete

evaluation of the four models considered by

completing the evaluation sheet.

This evaluation was performed using a Likert

scale, in which the responses for the items vary

according to degree of intensity, and the categories

are ordered and equally spaced with the same

number of items in all categories (Alexandre, 2003).

The evaluation was conducted at five levels, namely:

1 - Totally disagree, 2 – Mostly disagree mostly, 3 -

I'm not sure, 4 - Mostly agree, 5 - Totally agree.

Table 5: Evaluation of models according to the criteria.

In Table 5 is possible to make two observations -

how each criterion is met in overall by the models,

and how close the models are to meet all criteria. In

the analysis the averages are calculated per criterion,

horizontally, the average per modeler, vertically, and

the overall averages for the tool, which is the overall

average for that particular notation.

At this point the work indicated a tie between

ARIS and BPMN notations, with a score very close

to the maximum possible. The IDEF0 and SPEM

models were lower the previous, and team of

modelers were not inclined towards either of the two

best placed notations.

Aiming to broaden the search and check the level

of understanding of the models that were tied in first

place, another survey was conducted. This time,

involving 50 people from different areas of the

organization without any formal training in process

modeling. The participants secretaries, people in the

area of HR, Purchasing and other general

administrative professionals.

The research was composed of referral process in

both modeled in BPMN notation as ARIS, the actual

process of reference in Portuguese, and a

questionnaire to indicate which of the two models

ICEIS 2011 - 13th International Conference on Enterprise Information Systems

340

would be easier to understand. There was an

advantage on the ARIS over BPMN, as in Figure 4.

Figure 3: Research models of understanding with people

from different areas.

After this stage the BPMN notation is now

considered the most appropriate in terms of human

understanding, and the next steps of research

focused on this notation, since the aspect of human

understanding was a working premise. Moreover,

unlike the ARIS notation, BPMN notation is an open

standard from the OMG and has the support of

major players in this market, including IBM, Oracle,

SAP, Unisys and even IDS Sheer, owner of the

notation and tool ARIS.

4 EVALUATION OF TOOLS

Considering the advantage gained by BMPN rating,

the work continued by evaluating tools available in

the Brazilian market that would support this

notation. The tools analysis phase had the support

from a group of six modeling processes

professionals, including the original group of four

professionals and two modelers who joined the work

in this phase.

Initially we defined the desirable characteristics,

ie the criteria to be considered in evaluating the

tools. Criteria used in other similar works, such as

(Benedictis, Amaral, & Rozendfeld, 2003), and

(Recker, 2010), formed the basis of the criteria of

this work. The criteria are presented in Table 6.

Table 6: Criteria for analysis tools.

1 Tool Requirements

1.2 Language in native language (Portuguese - BR)

1.3 Enables cooperative work

1.4 Integrated repository for models

1.5 Versioning allows

1.6 Allows WEB Publishing

1.7 Allows export to other tools via XPDL.

1.8 Allows integration with tools via BPEL Workflow

1.9 Allows for method validation

1.10 Allows to query the database for ontological entity

Table 6: Criteria for analysis tools (Cont.).

1.11 Navigation between models of different levels

1.12 Additional attributes for model objects

1.13 Allows the use of more than one modeling notation

1.14 Allows to create new custom objects

1.15 Allows to add hyperlinks to other documentation

1.16

Allows to create filter with the set of objects you

wish to use

2 Additional Support

2.1 There are training courses

2.2 There are virtual communities of support

2.3 There are publications available to help and support

The work used the concept of relevance, in order

to equalize the degree of importance among different

criteria. With the exception of the criterion for

integration with workflow tools via BPEL, which

received a score of 4, the remaining criteria classified

as requirements of the tool received a score of 2. The

criteria for receiving additional support scored a 1,

because they are less relevant.

The tools considered were those found in the

Brazilian market: Mega, Tibco, QPR, QPR Express;

BizAgi Free; BizAgi Express; Free Aris, and Aris

complete. For those that had free options, these have

been independently assessed.

Modelers were divided into two groups, each

group responsible for examining a set of tools. Then,

the groups met to present and discuss the results,

generating a consensus evaluation. The evaluation

result can be seen in Table 7.

Table 7: Analysis of tools according to the criteria.

MODELING WORK PROCESSES AND SOFTWARE DEVELOPMENT - Notation and Tool

341

According to the analysis, and taking into

account specific criteria requirements tool, the tool

ARIS Full (complete and paid) received the highest

average score. Considering only the criteria for

additional support, tools QPR, BizAgi Free, and

BizAgi Express were tied in the lead. In the final

average, the tool ARIS Full received the highest

score.

By observing the average of fulfilling the criteria

for the tools it is possible understand what criteria

the manufacturers are more concerned with

delivering. The three criteria are best placed: 1 -

additional attributes for objects in the model, 2 -

WEB publication of the model in 3 - validation

mechanisms for the model.

Another observation is that the average final

score of the paid tools is not so far off compared to

the average final score of the free tools. This result

may indicate a strengthening of the free versions.

There are other factors that may be considered on

tools. For example, the tool QPR was only a tenth

behind the complete ARIS, which may recommend

its use. Still, cost issues may recommend using a

free tool such as BizAgi free, which scored even

higher than some paid tools.

Figure 4: Average final comparative tools.

5 CONCLUSIONS

After the performance modeling and analysis by a

group of experts, followed for an analysis by regular

users, the research points to the BPMN notation as a

solution more understandable by humans and has

potential to be translated into machines. In addition,

regarding tools, the results indicated the complete

ARIS, the QPR Express and BizAgi, the latter free.

An important contribution of this work was to

consider notations and tools whose relationship has

yet been little explored in previous works. Still, we

considered only specific tools of process modeling.

Another contribution was the structuring of criteria

from previous work, and even the addition of

specific criteria to the objectives covered by the

study. Finally, the recommendation of a method and

a specific tool, and provides more concrete basis for

the advancement of research in this area.

REFERENCES

Abdala, M., Lahoz, C., & Sant'Anna, N. (2003).

Utilizando o SPEM para a modelagem dos processos

da qualidade e do gerenciamento da configuração em

um ambiente integrado. V Simpósio Internacial de

Melhora de Processo de Software .

Agular-Savén, R. S. (2003). Business process modelling:

review and framework. International Journal of

Production Economics .

Alexandre, J. W. (2003). Análise do número de categorias

da escala Likert aplicada a gestão pela qualidade total

através da teoria da resposta ao item. Encontro

Nacional de Engehnaria de PRodução .

Araujo, R., Cappelli, C., Gomes Jr, A., Pereira, M.,

Iendrike, H. d., Ielpo, D., et al. (2004). A definição de

processos de software sob o ponto de vista da gestão

de processos de negócio. IV Simpósio Internacional de

Melhoria de Processos de Software .

Bargis, J. (2008). The importance of business process

modeling in software system design. Sicence of

Computer Programming .

Benedictis, C., Amaral, D. C., & Rozendfeld, H. (2003).

Avaliação dos principais métodos e ferramentas

disponíveis para a modelagem do processo de

desenvolvimento de produto. IV Congresso Brasileiro

de Gestão de Desenvolvimento de Produtos .

Gartner Group. (2008). Magic Quadrant for Business

Process Analysis Tools.

Gonzales-Perez, C., & Henderson-Sellers, B. (2007).

Modelling software development methodologies: a

conceptual foundation. The journal of Systems and

Software .

Green, P., & Rosemann, M. (2000). Integrated process

modeling: an ontological evaluation. Information

Systems , 25 (2).

IEEE. (1998). Standard for Functional Modeling

Language Syntax and Semantics for IDEF0 . IEEE

Explore .

Macedo, R. d., & Schmitz, E. A. (2001). Ferramentas de

modelagem de processo: uma avaliação. XXXIII

Simpósio Brasileiro de Pesquisa Operacional (SBPO).

Norton, D., Blechar, M., & Jones, T. (2010). Magic

quadrant for business process analysis tools. Gartner

RAS Core Research.

OMG. (2010). Business Process Model and Notation

(BPMN).

Ortiz-Hernández, J., Nieto-Ariza, M. E., Estrada-Esquivel,

H., Rodrígues-Ortiz, G., & Montes-Rendon, A.

(2007). SOQUA'07. Dubrovnik, Coratia.

Recker, C. J. (2010). Opportuniites and constraints: the

current struggle witdh BPMN. Business Process

Management Journal , pp. 181-201.

Santos Jr, P. S., Almeida, J. P., & Guizzardi, G. (2010).

ICEIS 2011 - 13th International Conference on Enterprise Information Systems

342

An ontology-based semantic foundation for ARIS

EPCs. ACM Symposium on applied computing .

Santos Jr, P. S., Almeida, J. P., & Pianissolla, T. L.

(2008). Construção de um metamodelo para ARIS

Method: escavação, refatoração e análise. Third

Workshop on ontologies and metamodeling in

software and data engeneering .

Scheer, A.-W. (2003). ARIS Toolset: A Software product

is born. ScienceDirect .

Soung-Hie, K., & Ki-Jin, J. (2000). Desiging performance

analysis and IDEF0 for enterprise modelling BPR.

Elsevier International Journalof Production

Economics .

Soung-Hie, K., & Ki-Jin, J. (2002). Designing

performance analysis and IDEF0 for enterprise

modelling in BPR. International journal of production

economics .

Souza, J., Andrade, T., Pereira, M., Lopes, M., Siqueira,

S., Baião, F., et al. (2009). Integração semântica de

modelos. Rio de Janeiro: Universidade Federal do

Estado do Rio de Janeiro.

White, S. (2004). Introduction do BPMN. BPTrends .

MODELING WORK PROCESSES AND SOFTWARE DEVELOPMENT - Notation and Tool

343