Mitigating Difficulties in Use-Case Modeling

Cristiana Pereira Bispo

, Ana Patricia Magalhães

1,2

, Sergio Fernandes

and Ivan Machado

Post Graduated Program in Computing and Systems, Salvador University, Salvador, Brazil

State University of Bahia, Department of Exact Sciences and Earth, Salvador, Brazil

Federal University of Bahia, Computer Science Department, Salvador, Brazil

Keywords: Information Systems, Strategy for Use Case Modeling, Use Case, Software Modeling, Requirement.

Abstract: The specification of software requirements in an enterprise system is crucial for software quality. The use-

case (UC) approach is often used to describe software requirements because, among other benefits, its

simplicity and ability to convey detailed information favors communication between business analysts,

requirements analysts and, crucially, end users, who can easily understand and validate requirements.

However, UC models are easier to understand than to specify, and difficulties in use case modeling (UCM)

may negatively affect the quality of UC models, and so its usefulness. UC models quality can be enhanced by

several modeling strategies mapped in the literature. However, no studies were found to show which of these

strategies can be used to mitigate specific difficulties. There is a gap between UCM difficulties and UCM

strategies. This paper presents a difficulty-strategy correlation proposal based on quality attributes of the UC

model. This correlation was initially evaluated in a controlled experiment with students of an undergraduate

program in computer science.

1 INTRODUCTION

Information Systems (IS) are crucial for business

corporations, as they provide information that

supports decision making, helping to reduce costs and

improve the quality and efficiency of services or

products. ISs may even enable new business models

that would not be viable without them, and provide

business with a competitive advantage. For all these

benefits to materialize though, a precise requirements

specification of an IS is of paramount importance.

Properly defining the requirements of these systems

is crucial for meeting the real needs of stakeholders

and avoiding significant costs if requirements

specification failures are identified late in the

software development lifecycle.

Furthermore, even before IS deployment, during

its development, requirements specification is a

critical source of information for analysis and design,

implementation, testing and project management.

Failures in requirements specification could be easily

propagated to the artifacts which use them as input,

decisively influencing the success of the project.

The Use Case (UC) approach is often used to

describe software requirements because, among other

benefits, it favors communication between business

analysts, requirements analysts and crucially end

users, who can easily understand and validate

requirements (Nascimento et al., 2017).

Use case modeling (UCM) has gained wide

acceptance from software analysts, designers and

testers (Tiwari and Gupta, 2015). The rules for

creating UC models are relatively simple to use and

follow. However, whether they are misapplied, it is

likely that low quality UC models (Anda et al., 2001)

with potentially significant impacts on the generated

product would be created. The poor quality of UC

models has been attributed to the inability of

requirements specifiers in creating UC models.

Typically, they face difficulties in both understanding

and representing the requirement (Anda et al., 2009);

understanding the domain of the problem

(Nascimento, 2017); specifying information

unambiguously (Bolloju, 2006), among others.

There are several strategies to support UCM, as

discussed in (Kitchenham and Charters, 2007) as

follows: i) virtualization technique for creating a

conceptual mental model that represents the user's

thinking of how the system works (Beimel and

Kedmi-Shahar, 2018). The authors emphasize that

this strategy can reduce the difficulties that affect the

accuracy, completeness and redundancy of the UC

Bispo, C., Magalhães, A., Fernandes, S. and Machado, I.

Mitigating Difﬁculties in Use-Case Modeling.

DOI: 10.5220/0009338100430052

In Proceedings of the 22nd International Conference on Enterprise Information Systems (ICEIS 2020) - Volume 2, pages 43-52

ISBN: 978-989-758-423-7

model; ii) Scenario Patterns (Ko et al., 2018) that help

requirements specifiers identify possible missing

requirements in the UC models they have created; and

iii) other strategies that use concepts of Business

Process Notation (Bouzidi et al., 2017), Role-Playing

(Nkamaura and Tachikawa, 2016), Antipattern (El-

Attar and Miller

, 2010).

Difficulties in UCM and strategies to assist UCM

requirements specifiers are already mapped in the

literature. However, no studies were found to show

which strategies can be used to mitigate specific

difficulties. Such evidence highlights a gap between

difficulties and strategies, characterizing a problem

that could be addressed by establishing a connection

between them. The underlying hypothesis is that, if

for each pointed out UCM difficulty a well-defined

and tested strategy is indicated that seeks to address

the lack of understanding of the requirements

specifiers to model UCs, such the difficulty would be

mitigated and consequently the quality of the UC

models will be improved.

In this effect, the main goal of this paper is to

present a proposal of correlation between difficulties

in UCM and strategies for mitigating these

difficulties.

Quality attributes of UC models were defined as

the link between UC modeling difficulties and UC

modeling strategies. In this study, we used them to

create the difficulty-strategy correlation. This link

emerged from the consideration that difficulties

negatively affect the quality attributes of UC models,

while modeling strategies positively affect the quality

attributes of these models. Once a correlation

between difficulties and modeling strategies was

made, an assessment was performed with one of the

modeling strategies to assess its effectiveness in

mitigating the modeling difficulties to which it was

correlated. The result shows the pertinence of the

performed correlation, demonstrating that the

proposed correlation can be a promising way to

mitigate the difficulties that affect the requirements

specifiers in the UCM.

The remainder of this paper is organized as

follows: section 2 introduces the underlying concepts

of this work and section 3 briefly discusses related

work. Section 4 describes the proposed correlation

between UCM difficulties and UCM strategies as

well as provides detailed information on the

methodology. Section 5 presents the correlation

evaluation. Finally, section 6 draws concluding

remarks and opportunities for future work.

2 BACKGROUND

Use cases are used to describe and document software

requirements. They are mainly used, among other

purposes, as a facilitator in the communication

between project team members and others involved

with the system and the use environment (Cockburn,

2000). UC modeling is the activity of designing a use-

case model, which describes in detail the functional

requirements of the software. They make use of

graphic and textual notation to, respectively

(Jacobson, 2004), i) create the UC diagram that

provides a visual summary of the system services and

their interaction with the environment and users

(called actors); and ii) describe the interactions

between the system and its actors. Difficulties

regarding the syntax and semantics of graphic and

textual elements in the elaboration of the UC model

compromise the quality attributes (Anda et al., 2009)

of this model, such as completeness, ambiguity and

inconsistency.

In this paper, a difficulty is considered to be any

lack of knowledge of requirement specifiers that

prevent them from modeling UCs meeting specified

quality requirements. The term strategy refers to any

UCM resources (guidelines, procedures, or activities)

proposed by researchers to improve the quality of UC

models. Making a connection between difficulties

and modeling strategies implies defining a

relationship that either links or associates a difficulty

to a strategy. However, correlating involves finding a

rationale for a relationship to be conceived. The

rationale investigated and identified as an effective

basis for correlating UC difficulties with UC

strategies are the quality attributes of the UC models

defined in (Anda et al., 2009).

3 RELATED WORK

The studies deemed as related to the purpose of this

research focus on the same aspect: the difficulties of

UCM requirements specifiers that prevent them from

building UC models meeting the defined quality

requirements.

Nascimento et al. (2017) sought to explore and

understand the difficulties in UCM by conducting

four experimental studies. As a result, they presented

a model of difficulties. Anda et al. (2006), Bolloju

(2006) and Siau and Loo (2006) also investigated and

reported difficulties in UCM. These works do not

present any strategy to mitigate these difficulties.

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems

To mitigate the difficulties that requirements

specifiers face when modeling UCs, several authors

propose the application of resources already used in

other domains to verify their effectiveness in UCM.

Bouzidi (2017) employed business process models to

derive UCs because these models are often available

in a company in the form of work instructions or

administrative manuals in a clear and structured

manner. Conversely, El-Atar and Miller (2012)

presented an antipattern-based strategy for UCM, in

which bad practices are identified to be replaced by

recommended solutions. In a preceding investigation,

we identified other strategies, and their respective

contributions to UCM (Bispo et al., 2019). However,

these studies do not indicate which strategies could

actually mitigate the UCM difficulties.

The difficulty-strategy correlation proposed in

this paper guides the requirements specifier in

selecting the most appropriate strategy to mitigate a

given difficulty. It avoids the adoption of ineffective

practices, and presents various alternatives for

applying tested and evaluated procedures to assist

UCM.

4 CORRELATION BETWEEN

UCM DIFFICULTIES AND UCM

STRATEGIES

The strategies identified in the literature were

proposed to improve the quality of UC models,

instead of indicating which specific difficulties are

mitigated by each strategy. In order to address such a

concern, we defined a two-procedures methodology,

as detailed next.

4.1 Correlation Methodology

The procedures adopted to make the correlation

possible were two-fold: (1) obtaining a precise

definition of the meaning of each difficulty, and then

grouping them into categories; and (2) obtaining a

precise definition of each quality attribute - in order

to gain a deeper understanding of each attribute, in

such a way that it would be possible to identify, in a

UC model, which quality attributes were either met or

not.

4.1.1 Categorizing UCM Difficulties

To categorize the difficulties of UCM, the studies that

present these strategies, earlier presented in Bispo et

al. (2019), were analyzed with the support of

Grounded Theory (GT) (Corbin and Strauss, 2008)

which helps in the construction of data-based

theories.

According to Corbin and Strauss (2008), GT can

be used when there is a need to understand a certain

situation from a volume of information about the

observed phenomenon; how and why the participants

act in a certain way; and how or why a particular

phenomenon or situation unfolds this way or that. An

example, illustrated in Table 1, is an excerpt from one

of the analyzed studies.

Table 1: A piece of text examined using GT.

“The main research question posed by this case study

is whether the proposed strategy can improve the

overall quality of UC models. This is achieved on two

fronts: (a) by restructuring the UC diagrams to adhere

to the notational syntax rules and semantics set by

OMG (OMG, 2010); and (b) by changing UC

descriptions to comply with recommended guidelines

and widely accepted practices (Sect. 2). Therefore, the

effectiveness of using our proposed approach will be

assessed by comparing the resulting UC model with

the original UC model, with respect to the aspects

mentioned in (a) and (b)...” (El-Attar and Miller,

2010).

By using GT procedures, as Figure 1 illustrates,

the highlighted phrase (taken from the example

citation in Table 1) "... restructuring UC diagrams to

adhere to syntax and semantic rules..." was

interpreted as: difficulties that prevent UC diagrams

from being modeled in accordance with syntax and

semantic rules. This interpretation is supported by the

information that the proposed strategy affects those

aspects (syntactic and semantic rules) so that there is

a general improvement in the quality of the UC

models also taken from the example in Table 1.

Another set of studies were examined following

the same approach, and similar interpretations were

made to precisely define a set of difficulties. To this

set the difficulties of UCM reported by Nascimento et

al. (2017), Anda et al. (2009), Bolloju (2006) and Siau

and Loo (2006) were added.

After defining the set of difficulties, they were

grouped into categories, as Figure 2 shows. This is

supported by our finding that a strategy which

supports the identification of an UC, also supports the

identification of actors and relationships. In other

words, the same strategy supports the identification

of the different elements (UC, actor, etc.) in the UC

diagram. Therefore, it was possible to group all these

difficulties in: Difficulty in identifying UCs, actors or

relationships. Furthermore, considering the same

example, for the word identify, some authors used the

Mitigating Difﬁculties in Use-Case Modeling

Figure 1: Using the GT method to precisely define difficulties in UCM.

word extract, while others used discover, all with the

same sense of finding the UC diagram element among

the requirements. The categorization considered the

synonyms as well. For the difficulties illustrated in

the example in Figure 2, the following category was

obtained: Difficulty in identifying/extracting/

discovering UCs, actors or relationships. The same

interpretation was adopted to obtain the other

categories.

Figure 2: UCM Difficulties Grouping.

The following categories of difficulties have been

established, with their respective meanings succinctly

expressed:

 Difficulty in identifying/extracting/discovering

UCs, actors or relationships.

o The requirements specifier face

difficulties in finding any functionality or

actor; or actor and actor; actor and UC; and

UC and UC relationships.

 Difficulty in representing/expressing elements

of UC model.

o This difficulty is related to the

representation of relationships. For

example, the requirements specifier

identifies that there is a relationship (UC-

actor, UC-UC, ...) but she is unsure about

how to represent it. Whenever an UC

extends another UC, for instance, the

requirements specifier may mix up the

direction of the arrow, the base case, and /

or the extended case.

 Difficulty in describing/detailing the semantics

of a UC model.

o The requirements specifier is not sure

about how to precisely define the meaning

of any model element. For example, given

the diagram, the specifier may find it

difficult to define and clarify scenarios or

behaviors of UCs, flows, and interactions.

 Difficulty in understanding/interpreting the

problem domain.

o The requirements specifier may find it

complex to model the system considering

the particularities of its actual

environment. Therefore, they may leave.

o aside important information in defining

requirements.

 Difficulty in understanding implicit

requirements.

o The requirements specifier face

difficulties in accurately specifying a

requirement that is not explicitly defined

by the domain stakeholders.

 Difficulty in synthesizing use cases.

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems

o The requirements specifier may not be

sure about the granularity of an UC. She

does not understand that she must combine

correlated actions into a single function

that adds value to the actor.

 Difficulty in precisely organizing the various

pieces of information into the UC model.

o The requirements specifier fails to be

concise and model only what is necessary

and sufficient.

UCM difficulties compromise the UC model

because they affect quality attributes (Nascimento et

al., 2017). Thus, it is necessary to know what a quality

attribute of the UC model is and how to identify it. If

it is compromised, a strategy for UCM may be

indicated.

4.1.2 Refining Quality Attributes Definitions

The usefulness of the UC model is a function of its

quality. There are many recommendations in the

literature about what quality means in a use case

model. A list of the attributes (and their definition)

from which the quality of UC models is evaluated can

be found in (Tiwari and Gupta, 2015), such as

completeness, consistency and ambiguity.

However, the definition of each attribute is

insufficient to identify its occurrence in the UC

model. This is because some definitions are not

sufficiently clear and detailed, and different authors

attribute different meanings to the same quality

attribute. For example, the definition of the

consistency attribute states that “the structure of the

UCs and the use of language and grammar must be

consistent across all UCs” (Anda et al., 2009). Here,

the word consistent was used to define consistency,

which does not elucidate the definition of the

attribute.

A clear understanding of how the difficulties of

requirements specifiers affect the quality of the UC

model is required to enable the correlation proposed

in this paper. To achieve this, it was necessary to

refine the definition of quality attributes mentioned in

the strategies for UCM. Thus, the definition of each

quality attribute was extended from the definition

found in the literature. This extension was based on

the approach of Zayan et al. (2018) which

systematically uses examples for model

understanding and domain knowledge transfer.

According to this approach, suitable examples help to

understand subjective or abstract definitions.

Example-based Understanding Acquisition. Table

2 illustrates an example of the approach taken to

clarify the definition of the consistency quality

attribute.

Table 2: Example that highlights inconsistent and

consistent diagrammatic structure.

The first line of Table 2 aggregates and

synthesizes definitions scattered in the literature for

quality attribute consistency. In the next line a

keyword has been associated with the attribute to

clarify its meaning. The following sentence is

designed to assist the requirements specifier in

judging some part of the UC model with respect to the

attribute. Then we hypothesized a scenario and two

simulated structures to help the requirements

specifier in understanding the quality attribute.

Similarly, the same procedure was adopted for

other quality attributes. Thus, a clearer and more

detailed definition was constructed and is

summarized as follows:

• Accuracy or Completeness or Integrity -

There should be no missing information nor

elements in the UC diagram and in the

corresponding textual descriptions;

• Consistency - The UC model information

should have the expected semantics. There

should not be any conflicting elements in the

diagrams and in their textual descriptions;

• Correctness - The UC diagram and its

descriptions must correctly represent the

requirements;

• Understandability - The information and

rules contained in the UC diagrams and

textual descriptions must be accurate and

clearly defined;

Mitigating Difﬁculties in Use-Case Modeling

• Ambiguity - There should be no information

in the UC diagram and textual descriptions

that can have more than one meaning;

• Redundancy - There should be no excessive,

repetitive or superfluous information in the

UC diagram and descriptions;

• Abstraction Level - The UC diagram and

descriptions should present only what the

software should do, at an appropriate level of

granularity. That is, the UC should not be

broken down into parts that have no value in

themselves.

The attributes Accuracy, Completeness and

Integrity of a use case model usually have the same

meaning.

4.2 Link between Difficulties and

Strategies for UCM

After the difficulties in UCM were categorized and

quality attribute definitions for the UC model were

refined, it was possible to understand how difficulties

affect attributes and identify which attribute is

affected (as illustrated in Figure 3).

Figure 3: Relationship between UCM Difficulties and

Quality Attributes in UC Models.

An example of what is shown in Figure 3 is: the

difficulty to identify a UC makes the model

incomplete because an expected functionality for the

system will not be found. For this example, using a

strategy that supports the identification of UCs will

avoid model incompleteness, nullifying the effect of

the presented difficulty. Thus, the quality attributes

are a link between difficulties and strategies.

Figure 4 shows the methodology for

implementing the proposal to correlate an UCM

difficulty with an UCM strategy.

Figure 4: Quality Attributes as link between UCM

Difficulties and Strategies.

4.3 Making the Correlation

The correlation, illustrated in Figure 5, is the

proposed solution to the problem of lack of

connection between the difficulties in UCM and the

strategies for UCM that we identified.

At the top of Figure 5 each difficulty is linked by

an arrow to one or more quality attributes that are

affected by it, in addition to the attribute code next to

the difficulty. For example, next to Difficulty

identifying / extracting / discovering UCs, actors or

relationships is Q1, which corresponds to the quality

attribute Accuracy or Completeness or Integrity, plus

the arrow link to these attributes. (in the middle of

Figure 5).

Each quality attribute, in turn, is enhanced by the

use of the strategy to which it is linked to by an arrow.

There is also the attribute code (s) next to the strategy

name.

Example: the strategy of Use Case Fragments

enhances the quality attributes Q2, Q4, Q5 and Q6,

respectively, Consistency, Comprehensibility,

Ambiguity and Redundancy. Attributes Q4 and Q5 are

affected by the same difficulty, while attributes Q2

and Q6 are affected by different difficulties.

Figure 5 also provides a short explanation of the

meaning of each correlation element, as follows:

difficulty, quality attribute, and strategy. It can be

inferred from the correlation that:

• a difficulty may affect more than one quality

attribute;

• when a quality attribute is affected by a

difficulty, other attributes may be, as a side

effect, also affected;

• a strategy can leverage more than one

attribute which can be affected by the same

difficulty or more than one distinct difficulty.

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems

Figure 5: Correlation between UCM Difficulties and UCM Strategies.

Mitigating Difﬁculties in Use-Case Modeling

5 CORRELATION EVALUATION

This section describes the correlation assessment, by

reporting the goal, the research questions, the

hypotheses, and the analysis and interpretation of the

data.

The Antipattern-based strategy was selected to be

validated because it is one of the most mentioned in

the literature for UCM. As we can see in Figure 5,

Antipattern enhances the consistency and ambiguity

attributes. These are affected by the difficulties in

describing / detailing semantics in the UC model and

in understanding implicit requirements.

The purpose of the evaluation was defined

according to Goal Question Metric (GQM) (Basili et

al., 1994), as described next: To analyze the

Antipattern-based strategy for the purpose of

assessing its effectiveness in mitigating the difficulties

for describing / detailing semantics in the UC model

and to understand implicit requirements regarding

consistency and ambiguity, from the point of view of

requirement specifiers.

Based on the overall goal, the following research

questions were defined (RQ):

RQ1: Is the Diagram Produced with the

Support of the Strategy Free of Defects that

Would Make It Inconsistent And Ambiguous?

This question aimed to evaluate whether the use of

the strategy corrected defects or prevented the

emergence of new ones. According to Kalinowski

(2012), a defect in the UC model is the failure to

comply with any UC good writing rules or guidelines

whose effect is to compromise some quality attribute

RQ2: Does using the Antipattern-based

Strategy Mitigate the Difficulties the

Requirements Specifiers Encounter that Affect

the Consistency and Ambiguity of the Use Case

Diagram? This question aimed to verify whether the

difficulties of requirements specifiers affecting

consistency and ambiguity had disappeared or

reduced.

In order to conduct the evaluation a set of

hypotheses were formulated: null (H0) and

alternative (HA) hypotheses, illustrated in Figure 6,

corresponding respectively to the existence of defects

in a set of diagrams modeled without the strategy

(called this set of UCD_Controlled) and another set

of diagrams modeled with Antipattern-based strategy

(called this set of UCD_Antipattern).

5.1 Data Analysis and Interpretation

The assessment encompassed the modeling of a

similar scenario by sixteen computer science

undergraduate students, which acted in the role of

requirements specifiers. Each participant built two

UC diagrams: one without using the strategy based on

Antipatterns and another using the strategy.

Inspection of the diagrams provided the results

illustrated in Figure 7. As the objective of the

experiment was not to evaluate the performance of

the strategy, the time spent to execute the experiment

was not considered. However, the effect of the

strategy on the number of defects in the UCs diagram

was considered.

It can be seen from Figure 7 that, with regard to

ambiguity, defects were reduced from 62 to 14, and

with regard to consistency, defects were reduced from

64 to 18 when using the Antipattern-based strategy.

The hypotheses were assessed through the Shapiro-

Wilk (1965) test, and it was possible to answer the

research questions.

Figure 6: Hypotheses formulated for the assessment.

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems

Figure 7: Effect of Antipattern-based strategy on reducing

ambiguity and inconsistency of a UC diagram.

For RQ1, the use of the Antipattern-based strategy

in UC diagram modeling has considerably reduced

the defects that make the diagram ambiguous and

inconsistent, although not all defects have been

corrected.

For RQ2, in this paper we assumed that defects

are generated by the UC modeling difficulties found

by the requirements specifiers. Based on this

assumption, if the defects that affect the consistency

and ambiguity of the diagram were reduced in

Antipattern modeling, then the difficulties that

generated these defects have also been mitigated.

However, there are limitations in the results, which

are considered indicative and not conclusive.

5.2 Threats to Validity

To prevent bias in the validation we now discuss

some threats to the validity of this empirical study.

Regarding internal validity as the level of knowledge

and experience of the participant in use case

specification may influence the results of the study,

we provided a training for every participant in

modelling use cases. Besides, we only selected

participants who were enrolled in the software

engineering discipline.

Concerning external validity, to minimize the risk

of sample representativeness, the diagrams produced

by the participants were also inspected by people who

did not participate in the experiment. The

representativeness of the chosen domain as well as

the size and complexity of the scenario are other

threats to the experiment. As there was no possibility

of using a real case, a scenario widely used in

software modeling was chosen. However,

experiments using real scenarios are necessary to

better validate our proposal. Related to construction

validity, we performed two pilot studies in order to

validate the material used in the experiment.

Distortions in understanding the anti-pattern strategy

were minimized through a summary of examples of

its use.

Finally, concerning conclusion validity, the

statistic method used may influence on the

conclusion. Therefore, we consulted a specialist to

define which method adopt.

6 CONCLUSIONS

To mitigate the difficulties in use case modeling, this

paper presented a proposal to correlate difficulties

and modeling strategies, with the link between both

being quality attributes of the use case model.

The correlation proposes modeling strategies to

improve the quality of the UC model. Therefore, as a

consequence, if a strategy improves quality, it

promotes learning. If the requirements specifier

learns, the difficulty is mitigated.

Thus, a preliminary assessment of a correlation

triad (difficulty-attributes-strategy) was performed.

The strategy tested was the one based on Antipattern

and the results showed that there is clear indication

that it mitigates the difficulties to which it is related:

difficulty to describe / detail semantics in the UC

model and to understand implicit requirements. Other

difficulties may be mitigated by other strategies

indicated in the correlation and which should be

tested in future work.

The strategy-difficulty correlation proposed in

this paper organizes and guides the requirements

specifier in the selection of the most appropriate

strategy to mitigate a given difficulty. This oriented

indication that the correlation provides avoids the

adoption of ineffective practices, as well as making

the requirements specifier aware of several

possibilities of applying tested and evaluated

procedures to assist UCM.

REFERENCES

Anda, B., Dreiem, H., Sjøberg, D. and Jørgensen, M., 2001.

Estimating software development effort based on use

cases – experiences from industry. in: M. Gogolla, C.

Kobryn (Eds.), UML 2001 The Unified Modeling

Language. Modeling Languages, Concepts, and Tools,

Lecture Notes in Computer Science, vol. 2185,

Springer, Berlin Heidelberg, pp. 487–502.

Anda, B., Hansen, K. and Sand, G., 2009. An investigation

of use case quality in a large safety-critical software

development project. In Information Software

Technology, vol. 51, n.12, pp. 1699–1711.

Basili, V., Caldiera, G. and Rombach, D., 1994. Goal

question metric paradigm. In: Marciniak, J. (ed.)

Mitigating Difﬁculties in Use-Case Modeling

Encyclopedia of Software Engineering, vol. 1, pp. 528–

532. John Wiley & Sons, Inc., New York.

Beimel, D. and Kedmi-Shahar. E., 2018. Improving the

identification of functional system requirements when

novice analysts create use case diagram: the benefits of

applying conceptual mental models. Requirements

Engineering (2018): 1-20.

Bispo, C., Fernandes, S. and Magalhães, A. P., 2019.

Strategies for Use Case Modeling: A Systematic

Literature Review. In Proceedings of the XXXIII

Brazilian Symposium on Software Engineering (SBES

2019). ACM, New York, NY, USA, 254-263. DOI:

https://doi.org/10.1145/3350768.3351795.

Bolloju, N., 2006. Exploring Quality Dependencies among

UML Artifacts Developed by Novice Systems

Analysts. In 12th Americas Conference on Information

Systems (AMCIS 2006) pp. 472.

Bouzidi, A., Haddar, N., Abdallah, M. B. and Haddar, K.,

2017. Deriving Use Case Models from BPMN Models.

IEEE/ACS 14th International Conference on Computer

Systems and Applications (AICCSA), Hammamet, pp.

238-243.

Cockburn, A., 2000. Writing Effective Use Cases. Reading,

Addison Wesley: Massachusetts.

Corbin, J. M. and Strauss, A., 2008. Basics of Qualitative

Research: Techniques and Procedures for Developing

Grounded Theory. 3rd Edition. SAGE Publications.

El-Attar, M. and Miller, J., 2010. Improving the quality of

use case models using antipatterns. Software &

Systems Modeling, Volume 9, Issue 2, pp 141–160.

Jacobson, I., 2004. Use cases - Yesterday, today, and

tomorrow. Software and System Modeling, vol. 3,

No.3, pp. 210-220.

Kalinowski, M., Card, D. N. and Travassos, G.H., 2012.

Evidence-Based Guidelines to Defect Causal Analysis.

IEEE Software, v. 29, p. 16-18.

Kitchenham, B. and Charters, S., 2007. Guidelines for

performing Systematic Literature Reviews in Software

Engineering. version 2.3, Keele/Staffs-UK and

Durham-UK.

Ko, D., Kim, S. and Park, S., 2018. Automatic

recommendation to omitted steps in use case

specification. Requirements Engineering.

Nascimento, E. S., Silva, W., Franca, B. B. N., Gadelha, B.

and Conte, T., 2017. A Model on the Difficulties to

Specify Use Cases. In Conference Ibero-American on

Software Engineering (CIBSE), Argentina.

Nkamaura, T. and Tachikawa, Y., 2016. "Requirements

engineering education using role-play training," 2016

IEEE International Conference on Teaching,

Assessment, and Learning for Engineering (TALE),

Bangkok, pp. 231-238.

OMG, 2010. Unified Modelling Language Superstructure -

version 2.3. http://www.omg.org/spec/UML/2.3/.

Shapiro, S. S. and Wilk, M. B., 1965. An analysis of

variancetest fornormality (complete samples),

Biometrika 52, 591–611.

Siau, K. and Loo, P., 2006. Identifying difficulties in

learning UML. Information Systems Management, vol.

23, n. 3, pp. 43-51.

Tiwari, S. and Gupta, A., 2015. A systematic literature

review of use case specifications research. In

Information and Software Technology, vol. 67, pp.

128–158.

Zayan, D., Sarkar, A., Antkiewicz, M., Maciel, R. S. P. and

Czarnecki, K., 2018. Example-driven modeling: on

effects of using examples on structural model

comprehension, what makes them useful, and how to

create them. Software & Systems Modeling, 18(3),

2213–2239. doi:10.1007/s10270-017-0652-3.

ICEIS 2020 - 22nd International Conference on Enterprise Information Systems