Model-based Tool Support for the Development of Visual Editors

A Systematic Mapping Study

David Granada, Juan M. Vara, Francisco P

erez Blanco and Esperanza Marcos

Kybele Research Group, Rey Juan Carlos University, M

ostoles, Madrid, Spain

Keywords:

Model Driven Engineering, DSL, IDE, Systematic Mapping Study.

Abstract:

Visual Domain Speciﬁc Languages play a fundamental role in the development of model-driven software. The

increase in this type of visual languages and the inherent complexity as regards the development of graphical

editors for them has, in recent years, led to the emergence of several tools that provide technical support for

this task. Most of these tools are based on the use of models and increase the level of automation of software

development, which are the basic principles of Model Driven Engineering. This paper therefore reviews the

main features, potential advantages and current limitations of the main tools that exist for the development of

graphical editors for visual DSLs.

1 INTRODUCTION

Model-Driven Engineering (MDE) (Schmidt, 2006)

is a natural step in the historical tendency of software

engineering as regards raising the level of abstrac-

tion at which software is developed. In this context,

Domain Speciﬁc Languages (DSLs) (Mernik et al.,

2005) play a cornerstone role in almost any proposal

that applies the principles of MDE. The fact that DSLs

are targeted toward a particular domain contributes to

ease of use and greater expressiveness allowing the

distance between business users and developers to be

shortened (Ghosh, 2011). DSLs also allow us to ﬁnd

solutions by using appropriate terms and abstraction

level for the problem domain, which improves the

quality, productivity and usability of software.

During the last few years, provided that MDE is

approaching its slope of enlightenment in the tradi-

tional hype cycle experienced by any technological

proposal (Linden and Fenn, 2003), we have under-

taken several works oriented towards the adoption of

more rigorous and systematic approaches for the de-

velopment of MDE tools. They were ﬁrst focused on

identifying best practices for the development of DSL

toolkits (Vara and Marcos, 2012), while the focus lat-

ter shifted to the development of model transforma-

tions (Bollati et al., 2013) as the main assets with

which model-based proposals can be automated. We

have been working as well in the application of MDE

principles in other ﬁelds (Vara et al., 2012).

One of the main lessons learned during this time

is that probably the most attractive feature of MDE

for researchers from other ﬁelds is the ability to use

visual models to depict their ideas, represent plans or

processes, etc. In order for this feature to become re-

ally useful those DSLs should come accompanied by

the corresponding tooling, basically in the shape of

visual editors or diagrammers (Selic, 2012).

Therefore, in order to know what the possibilities

are when building visual editors for DSLs and to ob-

tain a clear understanding of the state of the art as

regards tools for the development of these editors, in

this paper we present the main results obtained after

carrying out a systematic mapping study according to

the guidelines proposed in (Biolchini et al., 2005) for

the development of this type of studies in the context

of Software Engineering, which is composed of three

main phases: Planning, Execution and Analysis of Re-

sults.

Note also that only model-based tools are consid-

ered. That is, tools that make some use of models for

the development of visual editors. We do so under

the premises that: we ﬁrmly believe in the advantages

brought by MDE and we acknowledge that it is time

to start eating our own dog food (use models to de-

velop solutions) if we want MDE to be effective and

deﬁnitively adopted by the industry.

The results obtained from this study were then

used in an attempt to gain a general idea about the

current state of the different technologies used in the

tools analysed. We also wished to check for possible

improvements based on certain criteria evaluated. For

330

Granada, D., Vara, J., Blanco, F. and Marcos, E.

Model-based Tool Support for the Development of Visual Editors - A Systematic Mapping Study.

DOI: 10.5220/0006430503300337

In Proceedings of the 12th International Conference on Software Technologies (ICSOFT 2017), pages 330-337

ISBN: 978-989-758-262-2

the sake of space, we do not include here all the data

gathered during the development of the study, which

can be found in (Granada, 2016).

The remainder of this paper is structured as fol-

lows: Section 2 describes the main phases of the

method followed to carry out the systematic mapping

study; Section 3 presents its main results; Section 4 is

devoted to answer the research questions of the study

and Section 5 to compare this work with existing lit-

erature. Finally, Section 6 summarizes the main con-

clusions derived from this work.

2 METHOD

In this section, we shall focus on the planning of the

study (Biolchini et al., 2005), in which the research

objectives and how the study was conducted are in-

troduced. More speciﬁcally, Section 2.1 presents the

research questions deﬁned for the study; Section 2.2

presents the digital libraries and query strings used;

and ﬁnally, Section 2.3 enumerates the inclusion and

exclusion criteria.

2.1 Research Questions

The ﬁrst step conducted in the planning phase was

that of deﬁning the main objective of the systematic

mapping study, i.e., to identify and analyse the state

of the art as regards the generation of editors that sup-

port DSLs in the context of Model-Driven Engineer-

ing. To achieve this goal, we deﬁned the two research

questions (RQ) that gave rise to this study.

• RQ1: How many model-based tools with which

to generate editors for visual DSLs exist?

• RQ2: What are the main features and functionali-

ties of these tools?

2.2 Digital Libraries and Query Strings

The next phase includes the identiﬁcation of the dig-

ital libraries in which the search for studies will take

place (Biolchini et al., 2005). It is also necessary to

deﬁne a set of query strings that will be used in these

digital libraries.

Table 1 shows the digital libraries that were se-

lected for this study. It is worth noting that, although

Table 1 shows only six data sources, we initially con-

sidered another four digital libraries that were dis-

carded during the search process, given the lack of

relevant information and data obtained from them.

The box below shows the query string used to

search for the studies. Note that given that each of the

Table 1: Digital libraries.

digital libraries supports a slightly different syntax,

the query string had to be adapted to each search

engine.

(”development” OR ”generation”) AND (”visual editor” OR ”di-

agrammer” OR ”graphical editor” OR ”IDE”) AND (”DSL” OR

”DOMAIN SPECIFIC LANGUAGE”)

2.3 Inclusion and Exclusion Criteria

In any systematic mapping study, most of the results

ﬁrst retrieved from the digital libraries are not related

to the research questions posed. To specify inclusion

and exclusion criteria directly derived from the re-

search questions are therefore needed in order to ﬁlter

those studies (Biolchini et al., 2005). The inclusion

criterion used in this study was formulated as follows:

the abstract indicates that it is a study related to mod-

elling tools that can be used to develop editors for vi-

sual DSLs. We therefore checked the abstract (as well

as the title and keywords of each study) in order to

conﬁrm whether they satisﬁed the inclusion criterion.

Resulting studies were then subject to a detailed

reading in order to exclude non relevant ones accord-

ing to the following exclusion criteria: studies whose

main purpose is to classify other articles or are them-

selves systematic literature reviews; studies whose

main purpose is not related to graphical editors or vi-

sual DSLs, although these concepts are mentioned in

the content of the study; studies related to other types

of graphical editors, tools and/or technologies that are

not relevant to the purpose of this analysis.

The remaining set of studies were then considered

as potential candidates to become primary studies for

this review.

3 RESULTS

This section presents the results of this systematic

mapping study. First, section 3.1 provides some data

related to the number of works retrieved from the dig-

ital libraries. Section 3.2 then presents the main ﬁnd-

ings and conclusions gathered after analysing the set

of primary studies considered here.

Model-based Tool Support for the Development of Visual Editors - A Systematic Mapping Study

331

3.1 Studies Selection

Running the search in each digital library (using the

adapted version of the query string) yielded the results

shown in (Table 2) which computed 10,116 works.

First two columns on the left show the number of re-

sults for each digital library. Third column shows the

number of relevant studies, i.e., those that fulﬁlled the

inclusion criterion. Fourth column shows the number

of primary studies: those remaining that were not ex-

cluded by the exclusion criteria. Last column shows

the percentage of primary studies found in each data

source. Note that most of the primary studies were

found in Google Scholar (27.82%) and ISI Web of

Knowledge (21.8%).

Table 2: Distribution of selected studies.

However, many of the relevant, and therefore, the

primary studies were available in more than one dig-

ital library. Table 3 shows the eventual number of

primary studies found after removing all the repeated

studies. These data show that more than 36% of the

primary studies found in this review were duplicated

in different digital libraries. We therefore removed

the copies and consequently obtained a list of 85 non-

duplicated primary studies.

Table 3: Duplicated primary studies.

3.2 Data Extraction

The data extraction phase of the systematic map-

ping study allowed us to collect the most relevant

information from the primary studies. Regarding

RQ1, the main model-based tools for the generation

of editors for visual DSLs resulted to be: DiaGen

(Minas and K

oth, 2000), EuGENia (Kolovos et al.,

2009), GMF (Gronback, 2009), Graphiti (Brand

et al., 2011), MetaEdit+ (Tolvanen and Kelly, 2009),

Obeo Designer (Juliot and Benois, 2010), Sirius

(Viyovic et al., 2014) and Tiger (Ehrig et al., 2005).

Note also that this is indeed a never-ending task, since

new proposals arise every so often.

In order to analyse each of these tools in a sys-

tematic manner and gather the data needed to answer

RQ2, the following criteria were used, derived from

the features analyzed by the primary studies consid-

ered:

Scope. One of the main issues to consider regarding

software engineering tools is whether the tool is

commercial or open source. The type of license is

also considered here.

Framework. Frameworks are technological struc-

tures that can provide the basis for the develop-

ment of software. In the context of MDE, Eclipse

is the most widely used framework (Vara and

Marcos, 2012). Given this, it is interesting to

identify whether the tool analysed is isolated or

runs on an existing framework.

Distinction between Abstract and Concrete Syntax.

The abstract syntax deﬁnes the set of abstractions

provided by a language and how they may be

combined to create models, while the concrete

syntax provides a notation that facilitates the

representation of models expressed with such

language.

Abstract Syntax. The ﬁrst step toward the deﬁnition

of a DSL is the speciﬁcation of its metamodel. It

is therefore relevant to identify the way abstract

syntaxes must be speciﬁed for the reviewed tool

to generate the visual editor.

Concrete Syntax. One of the main distinctive fea-

tures between these tools is the mechanisms pro-

vided for the deﬁnition of the concrete syntax of

the DSL for which the editor is been developed.

Editing capabilities. Once the editor has been devel-

oped, the ability to manipulate and modify the ele-

ments of the models that can be created with such

editor in a simple, complete and functional man-

ner becomes the most relevant need. Editing capa-

bilities of the editors produced with the reviewed

tool is something that is worth attention as well.

Use of Models. Given that we focus on model-based

tools, the extent to which models are used in

the development process implemented by the re-

viewed tool need attention.

Automation. The level of automation is one of the

most relevant features of MDE proposals. More-

over, one of the main factors for a software en-

gineering tool to success is the ability to perform

its job without adding external complexity to the

ICSOFT 2017 - 12th International Conference on Software Technologies

332

process. All this given, the level of automation

with which these tools are able to produce visual

editors is also considered in this study,

Usability. This criterion was included to check the

ease of use of each tool. In other words, the

simplicity with which users could perform all

the steps required to complete the development

of a visual editor for their DSLs. We check

for instance whether the tools provide ofﬁcial

documentation, wizards, guided tours, examples

and/or any other form of assistance.

Methodological Basis. Finally, when designing a vi-

sual notation, the issues related to the quality of

that notation should be considered. Several mech-

anisms with which to assess the quality of the

metamodel that collects the abstract syntax of a

DSL can be employed for this purpose (L

opez-

Fern

andez et al., 2014). There are as well dif-

ferent metrics with which to assess whether the

graphical symbols composing the concrete syntax

could be correctly processed by the human mind.

We analyse consequently whether the tools fol-

lows or applies any kind of scientiﬁc theory, em-

pirical evidence or any given method to guide, de-

rive or deﬁne visual notations.

4 DISCUSSION

In this section, we present the detailed answers to the

research questions posed in Section 2.1. For each of

the criteria deﬁned in the Data extraction results sec-

tion, a brief analysis will be presented.

4.1 Scope

We discovered that most of the tools analyzed are

open source with different license types. Only

two of these proposals have a commercial license:

MetaEdit+ and Obeo Designer.

4.2 Framework

Frameworks are characterized by the fact that they

provide various types of support that facilitate the re-

duction of development time. In this context, most

of the proposals under discussion are designed to be

used in the Eclipse development environment. This

framework is widely accepted by the development

community owing to the ease with which its function-

alities can be expanded through the creation of new

extensions that can be added to its core. It is notewor-

thy that MetaEdit+ is the only tool that provides its

own framework independent of Eclipse.

4.3 Distinction between Abstract and

Concrete Syntax

With regard to the distinction between abstract syn-

tax and concrete syntax, most of the tools analyzed

allow users to perform this process. For example, GMF

uses the component called GMF Notation. This com-

ponent has an EMF-standard mechanism to provide

annotations in an Ecore metamodel, which allows in-

formation on graphical decisions and the deﬁnition of

elements in the domain model to be maintained as two

different concepts.

Another interesting proposal in this regard is

Eugenia, which provides not only the ability to deﬁne

visual annotations during the deﬁnition of the meta-

model (Kolovos et al., 2009), but also the deﬁnition

of these visual annotations by means of external li-

braries written in EOL, signifying that the two syn-

taxes can be kept conceptually separated. Of the tools

analyzed, the only proposals that do not provide such

a separation are Diagen and Tiger.

4.4 Abstract Syntax

Most of the proposals analyzed use Ecore to deﬁne

the abstract syntax of DSLs, and of these propos-

als, some may also use other mechanisms to deﬁne

the abstract syntax. For example, Graphiti can use

objects deﬁned in Java, while Diagen can do so via

UML diagrams. The tools that do not use Ecore to

deﬁne the abstract syntax are: MetaEdit+ and Tiger.

Tiger uses the graphic transformation integrated en-

gine AGG (Attributed Graph Grammar) to deﬁne the

abstract syntax. Finally, MetaEdit+ uses the GOP-

PRR (Graph Object Property Port Role Relationship)

language, which also provides a support tool with

which to import and use other modeling languages.

4.5 Concrete Syntax

After specifying the abstract syntax it is necessary to

associate a visual representation with each of its ele-

ments. GMF and Graphiti have in common the use of

Draw2D tool, which is based on the use of Java ob-

jects to deﬁne the graphical representation of the do-

main elements. In Obeo Designer and Sirius, the

graphical aspects of the domain model elements are

deﬁned by means of templates called Odesign. These

templates allow us to deﬁne some visual properties

through the use of intuitive contextual menus. In

Eugenia, the deﬁnition of the concrete syntax is con-

ducted by means of textual annotations on the meta-

model, but it also provides the ability to deﬁne these

annotations using the EOL language in an external

Model-based Tool Support for the Development of Visual Editors - A Systematic Mapping Study

333

ﬁle. MetaEdit+ uses an internal library and a symbol

editor that allow the user to assign a graphical repre-

sentation to each of the elements of the metamodel.

Finally, Tiger uses a component called ShapeFig-

ures, which allows the user to assign the shape, color

and size to each of the elements of the domain model.

4.6 Editing Capabilities

Most of the proposals provide the common operations

needed to edit the elements of the models that can be

created with the editor developed. However, only a

few proposals have sophisticated and advanced op-

tions with which to edit these elements.

In this context, the tools that provide the greatest

editing capabilities are: GMF, Graphiti, MetaEdit+,

Obeo Designer and Sirius. For example, GMF pro-

vides some advanced editing features such as: a de-

sign palette, a browser, decorators, an assistant di-

agram, editing tips, ﬁlters, drag and drop, animated

zoom and exportation to other formats.

Likewise, Obeo Designer and Sirius, in addi-

tion to providing a wide variety of styles by default,

make it possible to extend these styles. Most of the

editing options can be easily performed through the

use of intuitive contextual menus and the properties

table of the elements related to the generated editor.

4.7 Use of Models

With regard to the extent to which models are used

for the production of graphical editors, Eugenia, GMF,

Graphiti, Obeo Designer and Sirius are those

that use different types of models in the development

of editors. GMF is speciﬁcally a proposal that pro-

vides a model-driven approach with which to gener-

ate graphical editors. Its core is the Graphical Deﬁni-

tion Model, which contains information related to the

graphical elements that will appear in a GEF-based

runtime (editor), while the tool palette are deﬁned by

the Tooling Deﬁnition Model. The latter two mod-

els are related by means of the Mapping Deﬁnition

Model, which links the graphical and tooling deﬁni-

tions to the domain model selected. The other tools

that have a high level of use of models are based on

GMF, and therefore inherit its approach.

4.8 Automation

It is important to obtain the editor, or at least the in-

termediate stages, automatically in order to provide

the user with a clear and efﬁcient process. In some

of the tools analyzed it is possible to achieve this

goal by means of a few initial instructions and the se-

lection of some default options. The proposals that

offer a higher level of automation during the pro-

cess of generating an editor are Eugenia, Graphiti,

MetaEdit+, Obeo Designer and Sirius. More

speciﬁcally, in the case of Eugenia, the deﬁnition of

the abstract syntax and visual annotations for each el-

ement of the domain makes it possible to generate all

the models involved in the editor-generation process

automatically. Similarly, Graphiti is another tool

that has an editor-generation process in which reﬁne-

ments and modiﬁcations made to the models created

during the process are not required. Obeo Designer

and Sirius go one step further, because these tools

provide the user with the ability to generate the editor

without having to either manually launch any inter-

mediate process or use a second instance of Eclipse.

4.9 Usability

In the case of usability, it is interesting to identify the

facilities provided by the proposals employed to cre-

ate and modify new editors. In this context, DiaGen,

EuGENia, Graphiti, MetaEdit+, Obeo Designer,

Sirius and Tiger provide a better usability. For ex-

ample, Diagen provides a set of interfaces that guide

users when deﬁning the abstract syntax and show

them how to use the editor generated. Moreover,

one of the main goals of Graphiti is to provide the

user with easy and intuitive methods when generat-

ing graphical editors. This goal is achieved by means

of an interface based on an API that allows the user

to build an editor without having prior knowledge of

GEF or Draw2D.

On the contrary, as mentioned previously, GMF is

one of the most important projects for the generation

of graphical editors, although one of its weaknesses is

the complexity of the learning curve when compared

to other tools.

4.10 Methodological Basis

In this case we can conclude that none of the tools

studied follows a methodological basis in the process

of construction of graphical editors, nor use a theoret-

ical basis for assessing the quality of the visual nota-

tions generated. In this respect, none of the tools an-

alyzed takes into account the basic principles of Hu-

man Computer Interaction (HCI) in order to obtain

concrete (visual) syntaxes that can be understood and

correctly analyzed by the human mind. The deﬁni-

tion of the concrete (visual) syntaxes therefore usu-

ally consists of the arbitrary assignment of graphical

symbols to the concepts deﬁned, of which the abstract

ICSOFT 2017 - 12th International Conference on Software Technologies

334

syntax of the language is composed.

In this context, only a few tools such as

MetaEdit+, Obeo Designer and Sirius provide

certain facilities related to the use of visual variables

(location, shape, color, brightness, size, orientation

and texture) (Bertin, 1983). These three tools have a

properties panel from which it is possible to select,

in a way that is intuitive for the user, all the possible

values of the visual variables that can be used in the

elements of the concrete syntax generated.

4.11 Summary Results

Having analyzed the different proposals, and in order

to provide a rapid overview of them, Table 4 shows

a summary of the comparative study according to the

characteristics deﬁned in the previous subsections.

As will be observed in the summary of the com-

parative study presented in Table 4, most of the pro-

posals solve the problem of the conceptual separation

between the metamodel (abstract syntax) and the de-

sign features that are attached to each of its elements

(concrete syntax), without adding information related

to the domain of the solution in the context of the

problem domain. There are also proposals that au-

tomatically generate many of the intermediate phases

during the process of generating an editor. However,

most of these tools have to reﬁne each intermediate

phase through the use of mechanisms that are com-

plex and unintuitive for the user.

As regards the use of models, only a few of the

tools use models in each of the editor-generation

phases, which considerably facilitates the usability of

the tool and allows the user to focus on the impor-

tant aspects of the domain. Furthermore, in this work

we have found that most of the proposals have been

designed to be used in Eclipse, which is one of the

most popular frameworks in the software develop-

ment community. This feature facilitates integration

and interoperability with other software systems de-

veloped under Eclipse. Regarding the use of a scien-

tiﬁc basis to design visual notations, we have identi-

ﬁed that virtually no tool provides mechanisms with

which to achieve this goal, and only a few of them

provide the possible values of visual variables in a

way that is orderly and intuitive for the user.

Furthermore, the values of the last row in Table

4 represent the percentage of the maximum possi-

ble score for the qualitative criteria evaluated (ticks

obtained / maximum possible ticks). As this data

shows, the highest values are obtained for the us-

ability (91.6%) and the use of models (87.5%) crite-

ria, while it is worth noting that a very small value

(12.5%) is obtained for the criterion related to the

use of a methodological basis so as to take into ac-

count the quality of the visual notations generated.

Given that, we can state that of the commercial tools,

MetaEdit+ and Obeo Designer are the most com-

plete proposals according to the characteristics evalu-

ated in this work. Of the open source tools, Eugenia,

GMF, Graphiti and Sirius can, meanwhile, be con-

sidered as the most complete and functional for the

purpose of generating editors for DSL. And this may

be understandable, considering that these four tools

have technological bases in common, such as the fact

that they use EMF.

Many of the proposals discussed in this paper

therefore have interesting features as regards gener-

ating graphical editors for DSL, although several of

their features have room for improvement and reﬁne-

ment. In order to obtain tools that are able to create

editors with a minimal difﬁculty and transparency for

the user, it is important that these tools take into ac-

count aspects such as the use of frameworks that fol-

low the standards of the software development com-

munity, the use of models in each phase of editor gen-

eration, the separation between abstract and concrete

syntax, automation in the development process and

the usability of the tool itself.

Finally, we have concluded that none of the tools

follows a methodological basis in the process of con-

struction of editors that support DSLs, nor use a theo-

retical basis for assessing the quality of the visual no-

tations generated, in order to guide to the user in the

generation of visual notations that are easier to pro-

cess by the human mind, without being limited to the

classic geometric shapes like boxes and arrows.

5 RELATED WORKS

In this section, we brieﬂy present some related works

(Ehrig et al., 2005; Pelechano et al., 2006; Amyot

et al., 2006; Kelly, 2004; Kern et al., 2011). More

speciﬁcally, some works, such as (Ehrig et al., 2005),

present the Tiger tool, but they also present a state-

of-the-art generation of a graphical editor divided into

two generation categories: the Eclipse-based editor

generation and the Graph-transformation based editor

generation. The choice of presenting a state of the art

divided into these two categories is made because the

tool presented combines the advantages of formal vi-

sual language speciﬁcations using graph transforma-

tions with the facilities offered by Eclipse/GEF in or-

der to generate graphical editors.

In (Pelechano et al., 2006) the authors present

a comparative study of tools for model-driven de-

velopment, and more speciﬁcally a comparison be-

Model-based Tool Support for the Development of Visual Editors - A Systematic Mapping Study

335

Table 4: Overview of tools for graphical editors development.

tween Microsoft DSL tools and Eclipse Modeling

plugins. This study presents the main differences be-

tween these two categories of tools according to cer-

tain criteria, such as how to collect the abstract syntax,

the format of its models, the type of concrete syntax

used and different types of transformations between

models. It also presents an experiment with real users

of the tools, which allowed some signiﬁcant values to

be obtained as regards user preferences and the effec-

tiveness of using each of the different approaches.

Furthermore, (Amyot et al., 2006) present a com-

parative study of 5 tools that support the development

of DSML environments: GME, Telelogic Tau G2,

Rational Software Architect, XMF-Mosaic and

Eclipse EMF/GEF. In this study, the following evalu-

ation criteria are used: graphical completeness, editor

usability, effortlessness, language evolution, integra-

tion and transformation. The authors of this paper as-

sessed these criteria by developing a graphical editor

for a case study with each of the ﬁve tools analyzed.

Likewise, (Kelly, 2004) shows a comparison be-

tween Eclipse EMF/GEF and MetaEdit+ for Domain-

Speciﬁc Modeling. This paper presents a brief de-

scription of these types of tools and a comparison be-

tween them, focused on comparing only those aspects

related to the development time required and the num-

ber of lines of code generated.

Finally, (Kern et al., 2011) present a comparison

of a set of metamodeling languages like ARIS, Ecore,

GOPPRR, GME or MS DSL Tools, which are using

the heavyweight metamodelling approach (Frankel,

2003) and are available as tools.

In conclusion, the main differences between these

articles and the study presented here is that these ar-

ticles are mainly focused on presenting one (or sev-

eral) tools in detail, and showing only some differ-

ences among these tools. Furthermore, those works

that show a wider comparison do not follow a rigor-

ous literature review process, and the research criteria

are different and fewer than those evaluated here.

6 CONCLUSIONS

Visual DSLs are becoming increasingly more impor-

tant and this implies that there is clearly a need for

tools with which to create supporting tools for these

DSLs as quickly and effectively as possible (V

olter,

2009). In order to address this need, the ﬁndings

of this systematic mapping study have enabled us to

gather some data and ideas that provide a complete

picture of existing tool support for the development of

visual editors for DSLs. This study has also served to

identify some of the main challenges that the genera-

tion of these editors should address in the forthcoming

years. Some of them are summarized as follows:

Giving more relevance to the two basic principles

of MDE when developing visual editors, i.e., leverag-

ing the role of models throughout the entire develop-

ment process and enhancing the level of automation

in the development process.

Usability plays a cornerstone role for this type of

tools. Note that domain experts have not to be neces-

sarily familiarized with software engineering tools at

all. Ease of use is then key to enable users to develop

visual editors for the domain of interest without even

noticing that they are applying model-based princi-

ples or developing a DSL.

ICSOFT 2017 - 12th International Conference on Software Technologies

336

According to the ﬁndings of this study, none of

the analysed tools consider quality aspects during the

development of visual editors. This might be related

with the fact that most of these tools were built by

experienced developers not used to deal with HCI is-

sues.

It is worth noting that this type of study is a never-

ending task since new tools appear every some often.

One of our future goals is to periodically update this

study in order to identify new tools that enable the

generation of visual editors from a domain model.

ACKNOWLEDGEMENTS

This research has been funded by the Govern-

ment of Madrid under the SICOMORo-CM project

(S2013/ICE-3006), the ELASTIC project (TIN2014-

52938-C2-1-R), ﬁnanced by the Spanish Ministry of

Science and Innovation, and by the GES2ME Re-

search Excellence Group, co-funded by URJC and

Banco Santander

REFERENCES

Amyot, D., Farah, H., and Roy, J.-F. (2006). Evaluation of

development tools for domain-speciﬁc modeling lan-

guages. In International Workshop on System Analysis

and Modeling, pages 183–197. Springer.

Bertin, J. (1983). Semiology of graphics: diagrams, net-

works, maps. University of Wisconsin press.

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

G. H. (2005). Systematic review in software engineer-

ing. COPPE/UFRJ, Technical Report ES, 679(05):45.

Bollati, V. A., Vara, J. M., Jim

enez, A., and Marcos, E.

(2013). Applying MDE to the (semi-)automatic de-

velopment of model transformations. Information and

Software Technology, 55(4):699 – 718.

Brand, C., Gorning, M., Kaiser, T., Pasch, J., and Wenz, M.

(2011). Graphiti - development of high-quality graph-

ical model editors. Eclipse Magazine.

Ehrig, K., Ermel, C., H

ansgen, S., and Taentzer, G. (2005).

Generation of visual editors as eclipse plug-ins. In

Proceedings of the 20th ASE International Confer-

ence, pages 134–143. ACM.

Frankel, D. S. (2003). Model Driven Architecture Applying

Mda. John Wiley & Sons.

Ghosh, D. (2011). Dsl for the uninitiated. Communications

of the ACM, 54(7):44–50.

Granada, D. (2016). Desarrollo dirigido por modelos de ed-

itores grﬁcos cognitivamente eﬁcaces para Lenguajes

Especﬁcos de Dominio. PhD thesis, Rey Juan Carlos

University.

Gronback, R. (2009). Eclipse Modeling Project: A Domain-

Speciﬁc Language (DSL) Toolkit. Eclipse Series. Pear-

son Education.

Juliot, E. and Benois, J. (2010). How to build eclipse dsm

witwith being an expert developer? Obeo Designer

whitepaper.

Kelly, S. (2004). Comparison of eclipse emf/gef and

metaedit+ for dsm. In 19th ACM SIGPLAN confer-

ence.

Kern, H., Hummel, A., and K

uhne, S. (2011). Towards a

comparative analysis of meta-metamodels. In Pro-

ceedings of DSM’11, pages 7–12. ACM.

Kolovos, D. S., Rose, L. M., Paige, R. F., and Polack, F. A.

(2009). Raising the level of abstraction in the devel-

opment of gmf-based graphical model editors. In Pro-

ceedings of the 2009 ICSE Workshop on Modeling in

Software Engineering, pages 13–19.

Linden, A. and Fenn, J. (2003). Understanding gartners

hype cycles. Strategic Analysis Report N

R-20-1971.

Gartner, Inc.

opez-Fern

andez, J. J., Guerra, E., and de Lara, J. (2014).

Meta-model validation and veriﬁcation with metabest.

In Proceedings of the 29th ASE international confer-

ence, pages 831–834. ACM.

Mernik, M., Heering, J., and Sloane, A. M. (2005). When

and how to develop domain-speciﬁc languages. ACM

computing surveys (CSUR), 37(4):316–344.

Minas, M. and K

oth, O. (2000). Generating diagram editors

with diagen. Applications of Graph Transformations

with Industrial Relevance, pages 539–541.

Pelechano, V., Albert, M., Mu

noz, J., and Cetina, C. (2006).

Building tools for model driven development. In Com-

paring Microsoft DSL Tools and Eclipse Modeling

Plug-ins, volume 227, pages 1613–0073.

Schmidt, D. C. (2006). Model-driven engineer-

ing. COMPUTER-IEEE COMPUTER SOCIETY-,

39(2):25.

Selic, B. (2012). What will it take? a view on adoption of

model-based methods in practice. Software and Sys-

tem Modeling, 11(4):513–526.

Tolvanen, J.-P. and Kelly, S. (2009). Metaedit+: deﬁning

and using integrated domain-speciﬁc modeling lan-

guages. In Proceedings of the 24th ACM SIGPLAN

conference, pages 819–820. ACM.

Vara, J. M., Andrikopoulos, V., Papazoglou, M. P., and Mar-

cos, E. (2012). Towards model-driven engineering

support for service evolution. Journal of Universal

Computer Science, 18(17):2364–2382.

Vara, J. M. and Marcos, E. (2012). A framework for model-

driven development of information systems: technical

decisions and lessons learned. Journal of Systems and

Software, 85(10):2368–2384.

Viyovic, V., Maksimovic, M., and Perisic, B. (2014). Sirius:

A rapid development of dsm graphical editor. In Intel-

ligent Engineering Systems (INES), pages 233–238.

olter, M. (2009). Md* best practices. Journal of Object

Technology, 8(6):79–102.

Model-based Tool Support for the Development of Visual Editors - A Systematic Mapping Study

337