Systematic Mapping Study of Model Transformations for Concrete

Problems

Edouard Batot, Houari Sahraoui, Eugene Syriani, Paul Molins and Wael Sboui

DIRO, Universit

e de Montr

eal, Montr

eal, Canada

Keywords:

Model Transformation, Systematic Mapping, Software Engineering.

Abstract:

As a contribution to the adoption of the Model-Driven Engineering (MDE) paradigm, the research community

has proposed concrete model transformation solutions for the MDE infrastructure and for domain-speciﬁc

problems. However, as the adoption increases and with the advent of the new initiatives for the creation

of repositories, it is legitimate to question whether proposals for concrete transformation problems can be

still considered as research contributions or if they respond to a practical/technical work. In this paper, we

report on a systematic mapping study that aims at understanding the trends and characteristics of concrete

model transformations published in the past decade. Our study shows that the number of papers with, as main

contribution, a concrete transformation solution, is not as high as expected. This number increased to reach a

peak in 2010 and is decreasing since then. Our results also include a characterization and an analysis of the

published proposals following a rigorous classiﬁcation scheme.

1 INTRODUCTION

Model-Driven Engineering (MDE) has been gaining

much popularity in the last decade (Whittle et al.,

2014). It is an area of software engineering where

problems are expressed at levels of abstraction closer

to the problem domain, rather than the domain of

code, and software is realized through automated

transformation of domain models.

MDE, like any new technology, follows different

stages of adoption as described by (Moore, 2002). At

each stage, new contributions in research solve part of

the problems that represent major adoption obstacles.

One of the early obstacles to MDE adoption is the

difﬁculty to write and reuse model transformations

for many concrete problems. Indeed, the availabil-

ity of automated transformations is a prerequisite and

a founding principle of MDE.

In MDE, a model transformation is the automatic

manipulation of a model following a speciﬁcation de-

ﬁned at the level of metamodels (L

ucio et al., 2014).

It is an operation that accepts a source model as input

and produces a target model as output, where each

model conforms to its respective metamodel. Typ-

ically, a model transformation is deﬁned by a set

of declarative rules to be executed (Czarnecki and

Helsen, 2006). As an example of transformation,

in (Syriani and Ergin, 2012), UML activity diagrams

are transformed into Petri nets for simulation and

analysis purposes.

With the concern of the adoption of the MDE

paradigm, the MDE research community has pro-

posed solutions to concrete model transformation

problems. These proposals can be classiﬁed into two

categories: transformations that improve the MDE in-

frastructure, such as code generators for programming

languages (Funk et al., 2008; Di Marco and Pace,

2013), and transformations for domain-speciﬁc prob-

lems (Winkler et al., 2012; Yue and Ali, 2012).

However, in recent years, two phenomena

changed the landscape of MDE. First, MDE is more

and more used in industry. Many studies showed

that this new technology is used on strategic projects

in many industrial organizations (Mohagheghi et al.,

2013).The second phenomenon is the advent of new

initiatives for the creation of publicly available repos-

itories of metamodels, models, and transformations

(e.g., ReMoDD (France et al., 2007) and ATL Trans-

formations Zoo

). In particular, the Transformation

Tool Contest (TTC)

proposes publicly available so-

lutions to speciﬁc problems expressed as model trans-

formations. Both phenomena change the needs in re-

search towards, among others, automated approaches

www.eclipse.org/atl/atlTransformations/

www.transformation-tool-contest.eu/

176

Batot, E., Sahraoui, H., Syriani, E., Molins, P. and Sboui, W.

Systematic Mapping Study of Model Transformations for Concrete Problems.

DOI: 10.5220/0005657301760183

In Proceedings of the 4th International Conference on Model-Driven Engineering and Software Development (MODELSWARD 2016), pages 176-183

ISBN: 978-989-758-168-7

# exported records

(query)

1187

# exported records

(filters)

544

# records excluded

(filters)

643

# manually added

records

591

# screened records

1135

# papers assessed

for eligibility

# papers included in

the study

# excluded records

1040

13 papers excluded

• Concrete

transformations not a

main contribution

Identification

Screening Eligibility Included

Figure 1: Flow of information during the selection process.

to derive MDE artifacts such as metamodel well-

formedness rules (Faunes et al., 2013) or model trans-

formations (Baki et al., 2014). In particular, it is legit-

imate to question whether proposals for speciﬁc arti-

facts, such as metamodels and transformations of con-

crete problems can still be considered as research con-

tributions, or if they respond to a practical need.

To help answering such a question, we conducted

a systematic mapping study (Petersen et al., 2008),

covering the last decade (2005-2014), that aims at un-

derstanding the trends and characteristics of model

transformations proposed for concrete problems and

published in research forums. We opted for a system-

atic empirical process to minimize bias and maximize

reproducibility of the study. In addition to study the

evolution of the amount of published material during

the considered period, we are also interested in var-

ious characteristics of these transformations such as

their nature, the used languages, the involved meta-

models, and the relation to industry.

Our results show that the number of papers with as

main contribution a concrete transformation increased

since 2005 to reach a peak in 2010, but has been de-

creasing since then. Among other results, the major-

ity of the proposals deals with general modeling lan-

guages as compared to proposals for domain-speciﬁc

problems. An interesting number of the proposals are

industry-oriented.

The remainder of the paper is organized as fol-

lows. In Section 2, we describe the paper selection

and analysis procedure. We report the results of the

systematic mapping using a predeﬁned classiﬁcation

scheme in Section 3. We brieﬂy outline related work

in Section 4 and ﬁnally conclude in Section 5.

2 PROCEDURE AND SELECTION

The following subsections describe the main activities

we performed (from (Petersen et al., 2008)) in order

to discover the trends in model transformation.

2.1 Research Objectives

As motivated in Section 1, the modeling community

values to know whether producing model transforma-

tions that solve concrete problems is useful and still

considered a research contribution. We, therefore,

formulate our research objectives with the following

two research questions:

• RQ1: What are the trends in concrete model

transformations?

• RQ2: What are the characteristics of these trans-

formations?

2.2 Paper Selection

With these two objectives in mind, we determine the

scope of the search to be contributions in the litera-

ture, published between 2005 and 2014, that present

a model transformation for a concrete problem, e.g.,

transformation from UML activity diagrams to Petri

nets (Syriani and Ergin, 2012).

To retrieve the paper of interest, we queried the

Scopus database

. Moreover, we manually added pa-

pers to the corpus from the following specialized fo-

rums, which are likely to have papers of interest to

this study: ICMT, ECMFA, MODELS, and SOSYM.

As shown in Fig. 1, after the querying and ﬁlter-

ing, we obtained 544 papers from Scopus. For the

manual addition, we considered all the 591 long pa-

pers published in the four MDE forums between 2005

and 2014. After combining the two sources, a cor-

pus of 1 135 papers was considered for the screening

phase.

Screening is the most crucial phase in the system-

atic mapping process (Petersen et al., 2008). In order

to avoid the exclusion of papers that should be part of

the ﬁnal corpus, we followed a strict screening proce-

dure. With four reviewers at our disposal (co-authors

of this paper), each article is screened by two review-

ers independently.

Scopus archives over 55 million records from over 5,000

publishers www.elsevier.com/online-tools/scopus.

Systematic Mapping Study of Model Transformations for Concrete Problems

177

Table 1: Classiﬁcation scheme.

Category Description

Transformation

kind

Does the transformation operate on the

structure, mainly the syntax (e.g., migra-

tion), or on the behavior, mainly the se-

mantics (e.g., simulation), of the models

involved?

Metamodel

kind

Do the input and output metamod-

els describe a general-purpose language

(e.g., UML, source code) or domain-

speciﬁc language?

Model kind

Do the transformed models come from in-

dustrial data (e.g., private), from publicly

available data (e.g., open-source), or from

made-up toy examples?

Intent

Under which intent category does the

transformation fall, as deﬁned in (L

ucio

et al., 2014)?

Transformation

language

Is the model transformation expressed us-

ing a dedicated language for transfor-

mations (e.g., ATL), a programming lan-

guage (e.g., Java), or in another way

(e.g., formally, without implementation)?

Validation

Is the transformation veriﬁed and vali-

dated formally (e.g., proving properties),

empirically (e.g., case study, validated on

multiple models), or is there no validation

(e.g., informal argumentation)?

Scope

Is the transformation exogenous (deﬁned

on different metamodels), outplace (op-

erates on different models, but deﬁned

on the same metamodel), or inplace (op-

erates on the same models) as deﬁned

in (Mens and Van Gorp, 2006)?

Orientation

Does the transformation involve an au-

thor from industry or only academic au-

thors are concerned?

A paper is included if its title or abstract explicitly

mentions a model transformation in its peculiar con-

text. It is excluded if either the transformation is not

the main topic of the paper or the abstract indicates

the paper proposes a generalization of a transforma-

tion technique rather than an actual concrete trans-

formation. Among the 1 135 screened papers, 1 040

were excluded, 83 were included directly (both re-

viewers agreed), and 12 were included after conﬂict

resolution, for a total of 95 papers. After screening,

the full text of the 95 papers were read in order to de-

tect situations where the abstract suggested to be in

favor of including the paper, whereas the content of

the paper suggested otherwise. This step excluded 13

additional papers from the study. Indeed, although

the reviewers concluded from the abstract that the pa-

pers satisfy the four criteria, the in-depth examination

showed that the proposed transformations did not rep-

resent the main contribution of the papers. The ﬁnal

number of papers considered for the study is then 82

The complete list is available online geodes.

iro.umontreal.ca/modeltransformSMS/

When we determined the objectives of this study,

we already had in mind some of the criteria with

which we were going to evaluate the papers. While

reading all the abstracts during screening, we re-

ﬁned these criteria until we obtained the classiﬁca-

tion scheme in Table 1. It will be used to classify all

retained papers along different categories that are of

interest in order to answer our research questions.

3 ANALYSIS

In this section, we analyze the retained papers accord-

ing to the classiﬁcation scheme in order to answer the

research questions stated in Section 2.

8 8

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Specialized forums

General forums

Total

Figure 2: Evolution of concrete model transformations.

3.1 Evolution of Concrete Model

Transformations

The number of papers with a concrete model trans-

formation as a main concern increased since 2005 to

reach a peak in 2010 and has been decreasing since

then, as depicted in Fig. 2. We distinguish between

papers published in general software engineering fo-

rums and forums specialized in MDE. We note that

both general and specialized forums follow a similar

trend. However, we notice a shift towards specialized

forums around 2010. This is a recurring pattern that

often occurs when a new research community gains

popularity. The drop in 2013 may indicate that pro-

ducing concrete model transformations is becoming a

development task rather than a research endeavor.

3.2 Characteristics of Concrete Model

Transformations

Now that we shed light on a general trend in the model

transformation community, we investigate the charac-

teristics of concrete model transformations in order

to answer RQ2, using the classiﬁcation scheme pre-

sented in Table 1.

MODELSWARD 2016 - 4th International Conference on Model-Driven Engineering and Software Development

178

3.2.1 Intent

Language

Translation

33%

Refinement

22%

Semantic

Definition

12%

Abstraction

10%

Analysis

Model

Composition

Editing

Model

Visualization

Figure 3: Distribution of intent category.

Three classes of intents account, each, for more than

10% of the classiﬁed papers, as depicted in Fig. 3.

The most popular transformation intent class is lan-

guage translation, which encompasses a third of the

analyzed transformations. Such transformations es-

tablish a bridge between a source and a target mod-

eling language to achieve tasks that are difﬁcult (or

impossible) to perform on the source language (Yang

et al., 2014) or to make use of a speciﬁc tool (Winkler

et al., 2012). The second most frequent class of intent

is reﬁnement with 22%. Code synthesis is the domi-

nant transformation as in (Di Marco and Pace, 2013),

where wireless sensor network code is generated from

a UML proﬁle. Nevertheless, other kinds of reﬁne-

ments are also present, such as reﬁning a require-

ment model into a platform-independent model of a

multi-agent system (Harbouche et al., 2013). Seman-

tic deﬁnition is another predominant intent class with

12%. On the one hand, there are simulation trans-

formations that encode the operational semantics of a

language. On the other hand, there are translational

semantics transformations, whose purpose is to trans-

late one metamodel into another in order to deﬁne its

semantics, as in (Wagelaar et al., 2012). The lesser

popularity of the remaining intents (10% or less) is

due to the fact that approaches to perform, for exam-

ple, analysis and visualization, already exist in non-

MDE technologies. Therefore, the community has

not been focusing on explicitly modeling these tasks

by means of model transformation.

3.2.2 Transformation Kind

As depicted in Fig. 4 (right), a striking majority of the

papers deal with structural transformations. Behav-

ioral transformations are scarce, mostly designed to

analyze software evolution and perform simulations.

This accords with the distribution of intents. (G

ırba

et al., 2005) were precursors and showed how “one

can manipulate time information just like structural

information”. In fact, analysis or simulation of the

behavior of systems through transformations are of-

ten preceded by a structural translation of models,

in order to reuse existing tools: e.g., transforming a

domain-speciﬁc language or UML (Anastasakis et al.,

2007) into Alloy for analysis purposes.

3.2.3 Scope

Structural

87%

Behavioral

13%

Outplace

Inplace

Exogenous

87%

Figure 4: Distribution of scope and transformation kind cat-

egory.

The lack of experience of the community in terms

of endogenous transformation is put to light in Fig. 4

(left). Only 13% of the transformations were en-

dogenous, inplace (e.g., simulating the token behav-

ior of Petri nets (Syriani and Ergin, 2012)) or out-

place (e.g., performing row and column manipula-

tions on spreadsheets (Cunha et al., 2012)). In fact,

this corroborates with the intent distribution, since the

three most popular intents are typically implemented

by means of exogenous transformations. Many of

the proposed endogenous transformations are imple-

mented by mean of graph transformations (see, for

example, (Amelunxen and Sch

urr, 2008)). In addition

to refactoring, simulation and analysis, metamodel-

model co-evolution and metamodel-transformation

co-evolution (e.g., (Garc

es et al., 2014)) are naturally

favored application domains, as they require modify-

ing existing models. Other papers deal with model

synchronization (under the model composition intent

class), which is a special case as one can see it as an

exogenous transformation. Indeed, the synchroniza-

tion is performed between two models, generally be-

longing to two different metamodels. However, syn-

chronization can also be seen as the evolution of a

given model to handle new constraints resulting from

the modiﬁcation of another model. This is the case,

for example, in (Xiong et al., 2013) where the au-

thors propose an algorithm for synchronizing concur-

rent models.

Systematic Mapping Study of Model Transformations for Concrete Problems

179

3.2.4 Metamodel Kind

GPL / GPL

65%

DSL / DSL

18%

DSL / GPL

12%

GPL / DSL

Dedicated

53%

Other

29%

18%

Programming

Figure 5: Distribution of metamodel-kind and

tansformation-language category.

Fig. 5 (left) shows the distribution of combina-

tions between general-purpose and domain-speciﬁc

source and target metamodels of a transforma-

tion. Overall, two thirds of the approaches fa-

vor reusing existing tooling to simulate their sys-

tems, thus transforming between general-purpose lan-

guages (e.g., (Denil et al., 2014)). In particular, (a

subset of) UML is the most frequently used meta-

model. General-purpose languages, such as Ecore

(UML), XML (Eichberg et al., 2010), and Petri

nets (Syriani and Ergin, 2012), are mostly used as

target, again to favor reuse of existing technologies.

Domain-speciﬁc metamodels are typically used in

transformations to translate from one language to an-

other (Acher et al., 2009). Most of the papers involv-

ing source and/or target domain-speciﬁc metamodels

deal with business concerns, like business models in

(Siqueira and Silva, 2014). Other domain-speciﬁc to

general-purpose language transformations range from

particular aspects of application development such as

user-interface generation (Pastor, 2007) to very spe-

ciﬁc domains like conﬁguration of video surveillance

systems (Acher et al., 2009). Finally, a representa-

tive case of DSL-to-DSL transformations is the one

in (Selim et al., 2012). In this work, the transforma-

tion aims at migrating models expressed using a Gen-

eral Motors domain-speciﬁc language to AUTOSAR.

3.2.5 Transformation Language

As Fig. 5 (right) illustrates, more than half of the

transformations are implemented in a language dedi-

cated to model transformations. Half of those are im-

plemented in languages considered de facto standards

(ATL (Wagelaar et al., 2012) and QVT (Siqueira

and Silva, 2014)), and the other half with less pop-

ular languages (Pastor, 2007).Only 18% of the papers

still used programming languages for their transfor-

mations. These are mainly written in Java for the

Eclipse platform (Buchmann et al., 2011), but also

Prolog (Eichberg et al., 2010) and XSLT. The remain-

ing 29% of the papers only described the transfor-

mation, either without implementation or by imple-

menting it in a one-time use tool (Yue and Ali, 2012).

These observations reveal that model transformation

is being recognized as a paradigm in itself.

3.2.6 Model Kind and Orientation

On the one hand, in the vast majority of cases, we

found that scalable examples are not a priority: over

two-thirds of the papers illustrate their results with

toy examples, often made up for the particular work.

Although these toy examples cannot provide com-

pelling evidence about the quality of the proposed

transformation, they have the advantage of clearly il-

lustrate the transformation and then favor its adop-

tion by the potential users (Whittle et al., 2014). On

the other hand, industrial data shows how strong-built

and scalable the technology is. Among the few cases

where large sets of industrial data are involved, it

is worth mentioning the work by (Hermann et al.,

2014), in which data provided by satellite Astra is

used to validate the proposed transformation. Some-

times, although industrial data is used, its size is too

small to produce compelling evidence of the qual-

ity/usefulness of the proposed transformation (Selim

et al., 2012). Larger data models, from industry or

that are publicly available, have been slowly gain-

ing momentum in the past few years (Yue and Ali,

2012), as indicated in Fig. 6. This is certainly inﬂu-

enced by the fact that an industrial stakeholder was

involved for 20% of the papers, which has happened

predominantly in the 2010-2012 period. It is interest-

ing to note that papers with industrial authors follow

the same trend as those with academic authors only:

most publications in 2010, mainly exogenous trans-

formations, but with more industrial models.

3.2.7 Validation

Concrete model transformations are being validated

empirically more often with time (see Fig. 6). This

observation corroborates with a previous study in

2011 (Carver et al., 2011). Indeed, since 2011, MOD-

ELS has increased the number of pages for submis-

sions in order to give more space for a discussion

about validation. Half of the papers have validated

their work empirically as in (Yue and Ali, 2012), with

a peak in 2012. Nevertheless, most validations were

performed on small examples (Siqueira and Silva,

2014) which reduces the scalability of the validation

MODELSWARD 2016 - 4th International Conference on Model-Driven Engineering and Software Development

180

Validation

Model kind

Orientation

Empirical

Formal

No Validation

2005

2007

2006

2008

2009

2010

2011

2012

2013

2014

43%

51%

15%

16%

69%

20%

80%

Industrial

Toy

Public

Academic

Industrial

2005

2007

2006

2008

2009

2010

2011

2012

2013

2014

Figure 6: Evolution of validation, model kind and orientation, and their inﬂuences.

to its peculiar context. Furthermore, Fig. 6 shows

that a drop in studies using toy models correlates

with the increase of empirical validation. With re-

spect to the forms of validation, case studies are the

most used with, as mentioned earlier, small illustrat-

ing examples. Still, some contributions, such as (Her-

mann et al., 2014) and (Yue and Ali, 2012), pro-

pose strongly built validations, addressing both per-

formances and accuracy aspects. Validation is also

made by simulation as in (Denil et al., 2014) but they

remain scarce. Finally, in the very few papers that use

theoretical validation, this is implicit as the proposed

transformation is itself described formally.

3.3 Discussion

With respect to RQ1, after tracing correspondences

between the statistical results from this classiﬁcation,

there are two possible explanations for the trend ob-

served in Fig. 2: either the interest in model trans-

formation is decreasing or the model transformation

community has reached an appropriate level of ma-

turity and adoption since 2013. However, the for-

mer should be discarded because of the continued

popularity of transformation-exclusive venues: ICMT

and TTC. Indeed, its main artifacts—concrete model

transformations—have been pulling out from the re-

search literature since 2009 and are becoming con-

sidered as development tasks. This observation is

corroborated by the recent survey in (Whittle et al.,

2014), in which, a large number of surveyed actors

use MDE with concrete domain-speciﬁc artifacts. By

no means should one conclude that all transformation

problems have been solved. Instead, this indicates

that signiﬁcant scientiﬁc contributions are now being

exploited to solve practical problems.

Answering RQ2 suggests that the model transfor-

mation community has favored exploring exogenous

transformations, that are structural in order to trans-

late, reﬁne/synthesize code, or to give precise mean-

ing to models. This is mainly a consequence of want-

ing to reuse existing non-modeled software, as op-

posed to modeling the solution in a model transfor-

mation for behavioral transformations, such as anal-

ysis and simulation. As a result, a plethora of tools

have emerged for implementing model transforma-

tion, without one clear outlier. Although languages

dedicated to model transformations are preponderant,

programming languages are still used. From another

perspective, transformations are seizing to be applied

on toy models and, since 2010, are becoming more

applied in larger case studies. Finally, research in

model transformations is heading for a more stable

and grounded validation. The number of studies val-

idated empirically has been gradually increasing in

the past decade. This conﬁrms the level of maturity

reached by model transformation, as stated for RQ1.

4 RELATED WORK

Several works attempted to classify model transfor-

mations. However, to the best of our knowledge, this

paper is the ﬁrst to propose a systematic mapping

study for model transformations.

(Mens and Van Gorp, 2006) proposed a taxonomy

of model transformation based on how models are

manipulated and their execution strategies. Related to

this contribution, (Czarnecki and Helsen, 2006) clas-

siﬁed the features offered by languages to express

model transformations. For a separate but related

matter, (L

ucio et al., 2014) have cataloged the differ-

ent use cases and intents where a model transforma-

tion can be used.

Nevertheless, there have been several empirical

studies about various aspects of MDE. Concerning

the MDE adoption, (Hutchinson et al., 2011; Mo-

hagheghi et al., 2013) performed user studies in the

form of interviews and surveys among developers to

investigate how MDE technologies are applied in in-

dustry. Similarly and with respect to the development

Systematic Mapping Study of Model Transformations for Concrete Problems

181

process, (Mart

ınez et al., 2014) compared the perfor-

mance of maintenance tasks when using an MDE ap-

proach against a code-centric approach.

From a product perspective, (Vanderose and

Habra, 2008) propose an approach to empirically

evaluate quality factors of developed artifacts in

MDE. In the same vein, (Fernandez et al., 2013) per-

formed a series of user studies to assess the usability

of web applications developed in an MDE process.

The closest study to the one in this paper

is (Carver et al., 2011): a systematic survey focused

on MODELS in order to understand the frequency with

which empirical evaluations have been reported.

5 CONCLUSION

In this paper, we report on a systematic mapping study

to understand the trends and characteristics of model

transformations for concrete problems. Our study

uses a major online database, Scopus, along with the

published articles in the four major MDE forums.

This study, which covers the period 2005-2014, was

conducted following the systematic mapping process.

First, we collected all publications found by query-

ing the database and by gathering the papers of the

specialized forums. Then, we screened them using

their abstract to ensure that they were eligible for our

analysis. Finally, we analyzed every included article

to classify the proposed transformation according to a

predeﬁned scheme.

In addition to the ﬁndings discussed throughout

this paper, our study is a contribution to a global

assessment of the state of research and adoption of

MDE. Indeed, as for any new technology, it is our

duty as a research community to reﬂect globally on

what is relevant for research and what should be

treated as technical problems. In this context, we plan

to periodically repeat this study to have an up-to-date

portray of the situation. We also plan to perform a

similar study on additional MDE artifacts, in partic-

ular, metamodels. Finally, the classiﬁcation scheme

will be evolved to take into consideration new re-

search results.

REFERENCES

Acher, M., Lahire, P., Moisan, S., and Rigault, J.-P. (2009).

Tackling high variability in video surveillance systems

through a model transformation approach. In Models

in Soft. Eng.

Amelunxen, C. and Sch

urr, A. (2008). Formalising model

transformation rules for UML/MOF 2. IET Software,

2(3).

Anastasakis, K., Bordbar, B., Georg, G., and Ray, I. (2007).

UML2Alloy: A Challenging Model Transformation.

In Proc. of the Int. Conf. on Model-Driven Engineer-

ing Languages and Systems.

Baki, I., Sahraoui, H., Cobbaert, Q., Masson, P., and

Faunes, M. (2014). Learning Implicit and Explicit

Control in Model Transformations by Example. In

Proc. of the Int. Conf. on Model-Driven Engineering

Languages and Systems.

Buchmann, T., Westfechtel, B., and Winetzhammer, S.

(2011). ModGraph-A Transformation Engine for

EMF Model Transformations. In Int. Conf. on Soft-

ware and Data Technologies.

Carver, J. C., Syriani, E., and Gray, J. (2011). Assessing the

Frequency of Empirical Evaluation in Software Mod-

eling Research. In EESSMod, volume 785.

Cunha, J., Fernandes, J. P., Mendes, J., Pacheco, H., and

Saraiva, J. (2012). Bidirectional Transformation of

Model-Driven Spreadsheets. In Theory and Practice

of Model Transformations.

Czarnecki, K. and Helsen, S. (2006). Feature-Based Survey

of Model Transformation Approaches. IBM Systems

J., 45(3).

Denil, J., Mosterman, P. J., and Vangheluwe, H. (2014).

Rule-based model transformation for, and in simulink.

In Symposium on Theory of Modeling and Simulation.

Di Marco, A. and Pace, S. (2013). Model-driven approach

to Agilla Agent generation. In Wireless Communica-

tions and Mobile Computing Conf.

Eichberg, M., Monperrus, M., Kloppenburg, S., and

Mezini, M. (2010). Model-Driven Engineering of Ma-

chine Executable Code. In Proc. of the Eur. Conf. on

Modelling Foundations and Applications.

Faunes, M., Cadavid, J., Baudry, B., Sahraoui, H., and

Combemale, B. (2013). Automatically Searching for

Metamodel Well-Formedness Rules in Examples and

Counter-Examples. In Proc. of the Int. Conf. on

Model-Driven Engineering Languages and Systems.

Fernandez, A., Abrah

ao, S., and Insfran, E. (2013). Em-

pirical validation of a usability inspection method for

model-driven Web devt. Int. J. on Soft. and Systems

Modeling, 86.

France, R., Bieman, J., and Cheng, B. (2007). Repository

for Model Driven Devt (ReMoDD). In Models in Soft-

ware Engineering, volume 4364.

Funk, M., Nyßen, A., and Lichter, H. (2008). From UML to

ANSI-C - An Eclipse-Based Code Generation Frame-

work. In Int. Conf. on Software and Data Technolo-

gies.

Garc

es, K., Vara, J. M., Jouault, F., and Marcos, E. (2014).

Adapting transformations to metamodel changes via

external transformation composition. Int. J. on Soft.

and Systems Modeling.

ırba, T., Favre, J.-M., and Ducasse, S. (2005). Using

Meta-Model Transformation to Model Software Evo-

lution. Electronic Notes in Theoretical Computer Sci-

ence, 137(3).

Harbouche, A., Erradi, M., and Mokhtari, A. (2013). Deriv-

ing Multi-Agent System Behavior. Int. J. of Soft. Eng.

and Its Applications, 7(4).

MODELSWARD 2016 - 4th International Conference on Model-Driven Engineering and Software Development

182

Hermann, F., Gottmann, S., Nachtigall, N., Ehrig, H.,

Braatz, B., Morelli, G., Pierre, A., Engel, T., and Er-

mel, C. (2014). Triple Graph Grammars in the Large

for Translating Satellite Procedures. In Int. Conf. on

Theory and Practice of Model Transformations.

Hutchinson, J., Whittle, J., Rounceﬁeld, M., and Kristof-

fersen, S. (2011). Empirical assessment of MDE in

industry. In Proc. of the Int. Conf. on Software Engi-

neering.

ucio, L., Amrani, M., Dingel, J., Lambers, L., Salay,

R., Selim, G., Syriani, E., and Wimmer, M. (2014).

Model transformation intents and their properties. Int.

J. on Soft. and Systems Modeling.

Mart

ınez, Y., Cachero, C., and Meli

a, S. (2014). Empiri-

cal study on the maintainability of Web applications:

Model-driven Engineering vs Code-centric. Empirical

Soft. Eng., 19(6).

Mens, T. and Van Gorp, P. (2006). A Taxonomy of Model

Transformation. In Int. Work. on Graph and Model

Transformation, volume 152.

Mohagheghi, P., Gilani, W., Stefanescu, A., and Fernandez,

M. (2013). An empirical study of the state of the prac-

tice and acceptance of mde in four industrial cases.

Empirical Soft. Eng.

Moore, G. (2002). Crossing the Chasm: Marketing

and Selling Disruptive Products to Mainstream Cus-

tomers.

Pastor,

O. (2007). Generating User Interfaces from Con-

ceptual Models: A Model-Transformation Based Ap-

proach. In Computer-Aided Design of User Interfaces

Petersen, K., Feldt, R., Mujtaba, S., and Mattsson, M.

(2008). Systematic Mapping Studies in Software En-

gineering. In Int. Conf. on Evaluation and Assessment

in Soft. Eng., EASE’08.

Selim, G. M. K., Wang, S., C., J. R., and Dingel, J. (2012).

Model Transformations for Migrating Legacy Models:

An Industrial Case Study. In Proc. of the Eur. Conf.

on Modelling Foundations and Applications.

Siqueira, F. L. and Silva, P. S. M. (2014). Transforming

an enterprise model into a use case model in business

process systems. Int. J. on Soft. and Systems Model-

ing, 96.

Syriani, E. and Ergin, H. (2012). Operational Semantics

of UML Activity Diagram: An Application in Project

Management. In RE 2012 Workshops.

Vanderose, B. and Habra, N. (2008). Towards a generic

framework for empirical studies of Model-Driven En-

gineering. In Work. on Empirical Studies of MDE.

Wagelaar, D., Iovino, L., Di Ruscio, D., and Pierantonio,

A. (2012). Translational Semantics of a Co-evolution

Speciﬁc Language with the EMF Transformation Vir-

tual Machine. In Theory and Practice of Model Trans-

formations.

Whittle, J., Hutchinson, J., and Rounceﬁeld, M. (2014). The

state of practice in model-driven engineering. Soft-

ware, IEEE, 31.

Winkler, U., Fritzsche, M., Gilani, W., and Marshall, A.

(2012). BOB the Builder: A Fast and Friendly Model-

to-PetriNet Transformer. In Proc. of the Eur. Conf. on

Modelling Foundations and Applications.

Xiong, Y., Song, H., Hu, Z., and Takeichi, M. (2013). Syn-

chronizing concurrent model updates based on bidi-

rectional transformation. Int. J. on Soft. and Systems

Modeling, 12(1).

Yang, Z., Hu, K., Ma, D., Bodeveix, J.-P., Pi, L., and Talpin,

J.-P. (2014). From {AADL} to Timed Abstract State

Machines: A veriﬁed model transformation. Int. J. on

Soft. and Systems Modeling, 93.

Yue, T. and Ali, S. (2012). Bridging the Gap between

Requirements and Aspect State Machines to Support

Non-functional Testing: Industrial Case Studies. In

Proc. of the Eur. Conf. on Modelling Foundations and

Applications, volume 7349.

Systematic Mapping Study of Model Transformations for Concrete Problems

183