Practitioners’ Experiences on Developing Graphical Modeling Editors:

A Survey

Mert Ozkaya

, Kamran Musayev

and Mehmet Alp Kose

Yeditepe University, Department of Computer Engineering, Istanbul, Turkey

DFDS, Istanbul, Turkey

Independent Researcher, Istanbul, Turkey

Keywords:

Survey, Practitioners, Meta-Modeling, Modeling, Graphical Modeling Editors.

Abstract:

Graphical modeling editors used for modeling and processing any information can be developed using either

programming technologies (e.g., software libraries and frameworks) or meta-modeling technologies. How-

ever, with the existing literature, it is not clear which technique is popular and what motivate and demotivate

practitioners using those techniques. In this paper, we conducted a survey among 76 practitioners (with 52

acceptable responses) to understand their experiences on developing graphical modeling editors. The sur-

vey led to interesting results. The top motivation for developing editors is the model-driven engineering and

model transformation. 62% of the participants use meta-modeling technologies for developing editors, while

the rest use programming languages. Sirius is the top-used meta-modeling technology, while C# and Python

are the top-used programming languages. The participants using programming languages emphasized the

reduced learning-curve with programming and advanced development platforms for developing portable ed-

itors. Many of those participants have no idea about meta-modeling. The participants using meta-modeling

technologies revealed the huge time and effort gain with no-code editor development. Also, enhanced main-

tenance of editors by just changing the meta-model without writing code is considered important. However,

those practitioners state challenges on the meta-modeling technologies’ support for extensibility and customi-

sation, developers’ community, and complex meta-modeling.

1 INTRODUCTION

Models are considered as the abstract representations

of any real systems for understanding and reasoning

about systems (Seidewitz, 2003; Kent, 2002; K

uhne,

2006). With models, complex systems can be de-

scribed from different viewpoints at a high-level of

abstraction. Today, modeling is used in diverse in-

dustries, including defense/military, avionics, auto-

motive, ﬁnance, telecommunications, manufacturing,

and logistics, so as to improve their software devel-

opment and businesses. Indeed, as Rumbaugh stated

in (Rumbaugh et al., 1999), models can be used for

various purposes including the precise understanding

and communications of the domain knowledge, mak-

ing architectural design decisions, separating design

from requirements, generating various useful busi-

ness products, accessing and manipulating informa-

tion about complex systems, exploring alternative so-

lutions, and managing complexity.

Models can be speciﬁed using modeling lan-

guages (Harel and Rumpe, 2000), which offer textu-

al/visual notation sets for specifying models at vary-

ing degree of abstractions. Using meta-modeling

technologies, such as Eclipse-based tools (e.g., Sir-

ius (Viyovi

c et al., 2014) and Xtext (Bettini, 2013)),

Metaedit+ (Kelly et al., 2013) and MPS (Pech et al.,

2013), one can easily deﬁne a textual/graphical

domain-speciﬁc modeling language (DSML) that fo-

cuses on solving any domain-speciﬁc problem with

modeling. Indeed, meta-modeling technologies pro-

vide frameworks for deﬁning the language concepts,

relationships, rules and constraints (i.e., the meta-

model deﬁnition) along with the symbols correspond-

ing to those concepts (if graphical language to be de-

veloped). Then, a modeling editor is generated auto-

matically for specifying models in accordance with

the language deﬁnitions. Also, any tools that pro-

cess models can easily be developed and integrated

into the editors through which models can be pro-

cessed automatically for many useful operations such

as transformation (e.g., code generation) and model

276

Ozkaya, M., Musayev, K. and Kose, M.

Practitionersâ

Z Experiences on Developing Graphical Modeling Editors: A Survey.

DOI: 10.5220/0012062400003538

In Proceedings of the 18th International Conference on Software Technologies (ICSOFT 2023), pages 276-286

ISBN: 978-989-758-665-1; ISSN: 2184-2833

 2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)

analysis (e.g., checking models for pre-deﬁned rules).

An alternative way for developing graphical mod-

eling editors can be using programming languages

(e.g., Java, C#, and Python) and their supporting tech-

nologies (e.g., advanced development platforms, GUI

libraries, and frameworks) for developing graphical

editors. Some examples are Java’s swing library

and

Python’s tkinter

Given the meta-modeling and programming tech-

nologies that can be used for developing graphical

modeling editors, it is however not so clear to what

extent those technologies are used in different indus-

tries. Indeed, as discussed in Section 2, the exist-

ing surveys on modeling and meta-modeling do not

aid in understanding practitioners’ motivations for de-

veloping graphical modeling editors, their choices of

using meta-modeling technologies or programming

technologies, the reasons that make practitioners pre-

fer either of the technologies and the reasons that

make them not prefer, and any challenges faced with.

Therefore, we intended to design and execute a practi-

tioners survey in this study so as to learn practitioners’

experiences with developing graphical editors.

We believe that the survey results will be useful

for many stakeholders. Meta-modeling tool vendors

can improve their tools for better addressing the needs

of practitioners and the challenges faced with. Pro-

gramming technology providers (e.g., framework de-

velopers) can also get convinced about the lack of in-

terest on the programming technologies and ﬁgure out

how the existing libraries and frameworks can be im-

proved to address the needs of practitioners. Practi-

tioners can gain insights on alternative technologies

for building editors (i.e., strong and weak points) and

thus better make decisions on which technology to

prefer for their problems under speciﬁc constraints.

In the rest of the paper, we ﬁrstly discuss similar

works in the literature. Then, we introduce the re-

search methodology for our survey study. Next, we

give the analysis of the survey responses. Lastly, we

discuss the key ﬁndings and the threats to the validity

of the survey results.

2 RELATED WORK

The literature includes several industrial surveys on

modeling and meta-modeling. The existing survey

studies aid in learning

(i) practitioners’ experiences with modeling in

particular domains, e.g., embedded systems (Akdur

et al., 2018),

https://docs.oracle.com/javase/6/docs/api/

https://docs.python.org/3/library/tkinter.html

(ii) the extent to which modeling is used in par-

ticular countries (e.g., (Torchiano et al., 2011; Agner

et al., 2013)),

(iii) the particular aspects of modeling (e.g., vari-

ability modeling (Berger et al., 2013), UML usage

(Ozkaya and Erata, 2020a), and formal modeling

(Ozkaya, 2018b)),

(iv) practitioners challenges in modeling and

meta-modeling (Ozkaya and Akdur, 2021; Ozkaya

and Erata, 2020b), and

(v) practitioners’ experiences with the modeling

languages (Malavolta et al., 2013).

Moreover, the literature also includes analytical

studies that analyse and compare the existing model-

ing languages and tools with regard to some require-

ments of interest. In several works such as (Cabot and

Teniente, 2006; P

erez-Medina et al., 2007), the au-

thors analysed a set of well-known modeling tools for

a speciﬁc sub-set of requirements. In (Ozkaya, 2019),

Ozkaya analysed UML-based modeling tools for a set

of requirements that are important for practitioners.

In (Ozkaya, 2018a), Ozkaya analysed 124 different

architectural modeling languages for the practition-

ers’ needs. In (El Kouhen et al., 2012; Erdweg et al.,

2015; Kern et al., 2011), the authors analysed the ex-

isting meta-modeling technologies.

However, none of the existing surveys and analyt-

ical studies consider practitioners’ experiences with

the graphical modeling editors, which is the main fo-

cus of the study discussed in this paper.

3 RESEARCH METHODOLOGY

We followed the online survey method in our study

and thus reached our participants remotely over inter-

net so as to collect their responses and analyse them

in the quickest way possible (Groves et al., 2009).

3.1 Research Questions

In our survey, we investigate three research questions

to achieve the paper goal introduced in Section 1.

RQ1: What motivate practitioners for developing

graphical modeling editors? The goal here is to un-

derstand the reasons that lead practitioners to develop

graphical modeling editors. To this end, we intend

to learn why practitioners develop graphical model-

ing editors and which other facilities (e.g., code gen-

eration, model analysis, simulation, and collaborative

modeling) practitioners need to perform. We also in-

tend to learn the most and least important reasons.

RQ2: Which techniques and technologies practi-

tioners prefer for the graphical modeling editor de-

Practitionersâ

Z Experiences on Developing Graphical Modeling Editors: A Survey

277

velopment and their reasons for using those tech-

niques? We consider two techniques for the graphical

modeling editor development - the programming lan-

guages and meta-modeling technologies. Our goal is

to understand which technique(s) are preferred more.

Also, we intend to understand what make practition-

ers prefer (and not prefer) the meta-modeling tech-

nologies (and programming technologies).

RQ3: What are the challenges that practitioners

face with while developing graphical modeling edi-

tors? In this research question, the goal is to under-

stand any challenges that practitioners face with while

using either of the technologies (i.e., programming

technologies or meta-modeling technologies) for the

graphical modeling editor development.

3.2 Survey Design

To design our survey, we used our expertise on

modeling, modeling languages, and meta-modeling

(Ozkaya, 2018b; Ozkaya and Akdur, 2021) and ex-

amined the past surveys discussed in Section 2. So,

we proposed a list of questions in a draft form. We

performed a pilot study with a group of practition-

ers and academics who are expert on modeling and

meta-modeling. We got valuable feedback on the

structuring of the sections and questions, the answer

types (i.e., free-text, single choice, multiple choices),

ambiguous questions/answers, missing questions/an-

swers and duplicate answers. Finally, we ended up

with the survey questions given in Table 1. The ques-

tions 1-4 are for learning the demographic informa-

tion. The questions 5-6 are for addressing the research

question RQ1 given in Section 3.1. The questions 7-

10 and 12-14 are for addressing the research question

RQ2. The questions 11 and 15 are for addressing the

research question RQ3.

The proﬁle question about the participants’ expe-

riences and the participants’ frequency of developing

editors are single-answered questions. To maximise

the precision here, we ask the practitioners to choose

one of the related set of answers, e.g., the ques-

tion 3 for learning the participant experience (None,

Less than 2 years, 2-5 years, 6-10 years, and 10+

years). The questions for understanding the motiva-

tions, and reasons for preferring or not preferring any

technologies are multiple-answer questions and sup-

ported with pre-determined answer lists. We deter-

mined the answer lists for each multiple-answer ques-

tion through our expertise and the pilot study. Note

that participants can type their own answers unless

the existing answer list includes their desired answer.

Lastly, the questions for understanding any challenges

are free-text questions, which promote the partici-

pants to type their answers freely. We analyse any

free-text answers using the coding strategy (Popping,

2015). That is, we ﬁrstly check if the answer is con-

sistent with the question. If so, we check if the answer

is close to any other free-text answers, and thus we ei-

ther consider the answer as a new answer or increment

the frequency of the existing answer.

3.3 Survey Execution

Our survey has been accessible online via Google

forms for 2 months in between January 2023 and

February 2023. We shared the survey link with sev-

eral people involved in the graphical editor devel-

opment in diverse industries. In executing the sur-

vey, we initially used our personal contacts whom

we know through our consultancy services, R&D

projects, and past/present work experiences. In to-

tal, we sent e-mails to approximately 180 individ-

uals. We also used Google Scholar to determine

any practitioners who have contributed to the well-

regarded conference/journal papers that are associ-

ated with such relevant topics as software and systems

engineering, modeling, meta-modeling, and model-

ing languages, compiler/parser design and develop-

ment, and modeling toolsets. Besides, we sent posts

to several mailing-lists that we think are related to

the graphical modeling editor development. These

include the Eclipse modeling platform mailing-lists

where we could reach developers with meta-modeling

experiences (e.g., sirius-dev, graphiti-dev, papyrus-

rt-dev, emf-dev, emft-dev, agileuml-dev, papyrus-ic,

etc.) and the mailing lists where we could reach de-

velopers ( e.g., Netbeans mailing list, and IEEE ar-

chitecture description). Last but not least, we left

a message on several related Linkedin groups in-

cluding DSM forum, INCOSE, domain-speciﬁc lan-

guages, software developers, engineers, and program-

mers, software/technology, and software design pat-

terns and architecture, and software engineering pro-

fessionals, and MBSE. We also sent hundreds of mes-

sages to the individuals who are enrolled on those

linkedin groups and we think are likely to show in-

terest on our survey.

3.4 Survey Sampling

In our survey, we performed the non-probability sam-

pling technique as we did not have the chance for

reaching each practitioner in the population (Fricker,

2008). That is, we sent e-mails to hundreds of dif-

ferent practitioners whom we know as discussed in

Section 3.3. Also, to minimise any biases here due

to the non-randomness, we shared the survey link

ICSOFT 2023 - 18th International Conference on Software Technologies

278

Table 1: The survey questions.

Res.

Que.

Survey Questions

Multiple

Answers

Free

Text

Single

Answer

Proﬁle

1-Which industry(ies) do you work in? X X

2-What is (are) your current job position(s)? X X

3-How many years of experience do you have in the sector? X

4-Do you develop graphical modeling editors? X

RQ1

5-Which other tool(s) or services do you integrate into your graphical modeling editor?

X X

RQ1

6-Which of the following motivation(s) make you develop a graphical modeling editor?

X X

RQ2

7-How do you develop a graphical modeling editor and its supplementary tools (e.g.,

validators, generators, exporters/importers)?

RQ2

8-Which programming languages do you use for developing a graphical modeling editor?

X X

RQ2

9-What is(are) the reason(s) that make you prefer programming languages for developing a

graphical modeling editor?

X X

RQ2

10-What is(are) the reason(s) that make you NOT prefer meta-modeling technologies

for developing a graphical modeling editor?

X X

RQ3

11-Please tell us any challenges that you face with while using programming languages for the

graphical modeling editor development.

RQ2

12-Which meta-modeling technologies do you use for developing a graphical modeling editor?

X X

RQ2

13-What is(are) the reason(s) that make you prefer meta-modeling technologies for

developing a graphical modeling editor?

X X

RQ2

14-What is(are) the reason(s) that make you NOT prefer programming languages for

developing a graphical modeling editor?

X X

RQ3

15-Please tell us any challenges that you face with while using meta-modeling technologies

for the graphical modeling editor development.

across several actively-used mailing lists and Linkedin

groups where each interested practitioner who are en-

rolled has equal chance to participate.

3.5 Data Analysis

We have collected 76 responses from diverse indus-

tries. Among those, 2 participants stated in question-3

that they do not have any experiences in industry and

thus we directed the participants to submit the form

without ﬁlling any questions. Indeed, in our survey,

we are interested in learning from the practitioners

who have some experience about developing editors

in industry. Moreover, 22 responses belong to the par-

ticipants who stated in question-4 that that they have

never developed a graphical editor before. So, we di-

rected those participants to submit the survey too. So,

we ended up with 52 acceptable responses that have

been analysed as discussed in Section 4. To analyse

the data here, we performed the voting approach for

each question answer and made statistical inferences

using Google form and MS Excel.

4 ANALYSIS OF RESULTS

In this section, we give the analysis of the survey re-

sponses for each question.

Figure 1: Participants’ work industries.

4.1 Proﬁle Questions

In this section, we analyse the responses given for the

questions 1-4.

The industries in which the participants work are

displayed in Figure 1. So, top popular industry that

show interest on the survey is computer (33%), which

is followed by the research industry (25%) including

the participants from the research centers and univer-

sities and IT and Telecommunications (24%).

As shown in Figure 2, nearly half of the partici-

pants (43%) hold the software developer/programmer

positions in their company. Also, 30% of the partici-

pants are researchers.

Practitionersâ

Z Experiences on Developing Graphical Modeling Editors: A Survey

279

Figure 2: Participants’ job positions.

Figure 3: Participants’ experience in the sector.

Figure 4: Participants’ experience in developing graphical

modeling editors.

According to Figure 3, most of the participants

are highly experienced in their sector - 43% have 10+

years of experiences and 21% have 6-10 years of ex-

periences. Note that a very small number of the par-

ticipants who have no experiences in their sector have

been directed to submit the survey without answering

any of the technical questions so as to minimise any

biases.

As shown in Figure 4, 70% of the the practitioners

who participated in the survey develop (or developed)

graphical modeling editors to some extent. Those par-

ticipants (30%) who never develop graphical model-

ing editors have been directed to submit the survey so

as to minimise any biases.

Figure 5: The types of tools that the participants integrate

into their graphical modeling editors.

4.2 Developing Graphical Modeling

Editors

In this section, we analyse the questions 5-7, which

have been responded by the participants who indi-

cated that they develop(ed) graphical modeling edi-

tor(s).

As indicated in Figure 5, most of the participants

developing graphical modeling editors (87%) tend to

integrate their editors with model transformation tools

(e.g., code generators) so as to enable the automated

transformation of models speciﬁed with the editors

into some useful artefact. Integrating editors with

some exporter/importer tools (e.g., exporting and im-

porting models in XML, HTML ﬁle formats and ex-

porting models in picture formats) is the second-top

motivation for the participants (75%). While using

model analysis and validation tools (e.g., checking

user errors for incomplete, inconsistent, and incorrect

models) together with the editors is also important for

many participants, versioning models and accessing

the models collaboratively are rarely preferred.

Figure 6: Participants’ motivations for developing graphical

modeling editors.

Figure 6 gives the participants’ motivations for

the graphical modeling editor development. So,

most of the participants (77%) aim to perform the

model-driven software engineering activities with

their graphical editors, which promotes the speciﬁ-

cation of models, their analysis and transformation

into useful artefact (e.g., code) (Kent, 2002). 65% of

the participants develop editors because they want to

model any information in an understandable and man-

ageable way. Indeed, editors can be developed for

specifying multiple-viewpoints models where differ-

ent models that each focus on a separate concern can

ICSOFT 2023 - 18th International Conference on Software Technologies

280

easily be speciﬁed and related to each other (Rozan-

ski and Woods, 2011). Also, sub-diagramming can be

performed, where an element drawn can be clicked

for specifying another model via the newly open-

ing sub-editor. The third and fourth top motivations

for the graphical modeling editor development are

the need for visualising any information (e.g., using

boxes and lines) (60%) and modeling any information

abstractly (i.e., one of the main goals of modeling)

(56%).

Figure 7: The technologies that the participants prefer to

use for developing graphical modeling editors.

Figure 8: The industries that use the programming lan-

guages and meta-modeling technologies for developing

graphical modeling editors.

Figure 7 shows how the participants prefer to de-

velop the graphical modeling editors and their sup-

porting tools (e.g., code generators, model validators,

and exporter/importer tools). So, while 62% prefer to

use the meta-modeling technologies (e.g., Metaedit+,

Sirius, and MPS), the rest of the participants use the

programming languages and technologies (e.g., Java,

Python, and C++). Figure 8 shows the correlation be-

tween the participants’ work industries and the tech-

nology that they use for developing graphical model-

ing editors. So, the biggest portion of the participants

who use meta-modeling technologies are from the re-

search industry while the programming languages are

top-used by the computer industry.

4.3 Using Programming Languages

In this section, we analyse the responses given for the

questions 8-11, which are concerned with the experi-

ences of the participants who use programming lan-

guages to develop graphical modeling editors.

Figure 9 shows the programming languages used

by the participants. So, the top-preferred program-

Figure 9: The programming languages that are preferred

by the participants who use programming technologies for

developing graphical modeling editors.

ming languages are C# (40%) and Python (40%),

which is followed by Java (30%).

Figure 10: The reasons of the participants using program-

ming languages for developing graphical modeling editors.

As Figure 10 shows, maintainability is the top rea-

son of the participants using programming languages

for developing graphical modeling editors (65%). In-

deed, any programmers can essentially maintain the

tool whenever needed and no any additional abilities

(e.g., modeling and language engineering) and knowl-

edge (e.g., meta-modeling) are needed. Also, the ben-

eﬁts that are gained with the programming languages

such as platform-independency (e.g., jar applications)

and advanced software libraries are found motivating

by nearly half of the participants for the graphical ed-

itor development.

Figure 11: The reasons of the participants using program-

ming languages for NOT preferring meta-modeling tech-

nologies.

Practitionersâ

Z Experiences on Developing Graphical Modeling Editors: A Survey

281

In Figure 11, the reasons that make the par-

ticipants using programming languages avoid meta-

modeling technologies are addressed. 30% of the par-

ticipants who use programming languages have no

idea about meta-modeling technologies. 25% of the

participants stated that the team in which they work

have no knowledge on meta-modeling.

Some other issues selected by 25% of the partici-

pants are to do with meta-modeling technologies’

(i) lack of support for open-source develop-

ment for being extended by the developers whenever

needed,

(ii) lack of support for the customisation of the

editor user interfaces as the meta-modeling technolo-

gies may not allow practitioners to change the user-

interface layouts (e.g., changing the toolbar location

in the editor) and

(iii) immaturity (e.g., being very new products).

The participants did not state any challenges re-

garding their experiences with the programming lan-

guages for developing graphical modeling editors.

Figure 12: The meta-modeling technologies preferred by

the participants for developing graphical modeling editors.

4.4 Using Meta-Modeling Technologies

In this section, we analyse the responses given for the

questions 12-15, which are concerned with the ex-

periences of the participants who use meta-modeling

technologies to develop graphical modeling editors.

As Figure 12 shows, the top-used meta-modeling

technology for developing graphical modeling editors

is Sirius

(50%), which is based on Eclipse modeling

framework (EMF). Sirius is followed by Metaedit+

(35%), offered by Metacase

. Obeo Designer

that

is based on Sirius is used by 32% of the participants

and MPS

used by 23% of the participants. The

rest of the technologies including Eugenia

, GEMS

https://www.eclipse.org/sirius/

https://www.metacase.com/

https://www.obeodesigner.com/

https://www.jetbrains.com/mps/

https://www.eclipse.org/epsilon/doc/eugenia/

https://wiki.eclipse.org/GEMS

GME

, AtoMPM

, Gentleman

, and Enterprise Ar-

chitect’s model-driven generation technologies

are

rarely used by a few participants (1-4) only. Note that

a few of the participants (9%) indicated that they build

their own tools to build graphical editors.

Figure 13: The reasons of the participants using meta-

modeling technologies for developing graphical modeling

editors.

As Figure 13 shows, the top-reason that make the

participants use meta-modeling technologies is the

automated generation of an editor from the language

deﬁnitions (i.e., meta-model) without writing single

line of code (84%). The automated integration of

the editor with code generators and model validators

is second top-reason here (77%). Indeed, the meta-

modeling technologies offer frameworks for develop-

ing model transformation tools (e.g., code generators)

and automatically integrate the transformation tools

with the editor - no any effort required for integra-

tion here. Also, having a running editor quickly with-

out coding at all for the editor implementation & test-

ing and maintaining the editor (e.g., adding/removing

concepts) without coding are quite popular reasons

for using meta-modeling technologies (65-68%). On

other hand, meta-modeling technologies’ support for

the collaborative modeling (i.e., multi-user access),

hybrid modeling (i.e., modeling using different types

of notation sets e.g., visual, textual, and tabular), and

versioning (e.g., versioning models and meta-models)

are rarely considered by the participants.

In Figure 14, we give the reasons that make

the participants who use meta-modeling technologies

avoid programming languages for the graphical mod-

eling editor developments. So, most of those par-

ticipants (84%) ﬁnd the programming languages re-

quiring too much time and effort for the editor devel-

opment. Another crucial issue (63%) is to do with

the difﬁculties in maintaining the editor at code level

whenever some changes are required (e.g., adding

new concepts and relationships).

The participants stated several of their challenges

faced while using meta-modeling technologies and

https://webgme.org/

https://atompm.github.io

https://github.com/geodes-sms/gentleman

https://sparxsystems.com/resources/mdg tech/

ICSOFT 2023 - 18th International Conference on Software Technologies

282

Figure 14: The reasons of the participants using meta-

modeling technologies for NOT preferring programming

languages.

the top ones are

(i) the inadequate support for customising the ed-

itor user interfaces,

(ii) the inadequate support for specifying complex

meta-model concepts and transformation algorithms,

and

(iii) lack of community support (e.g., mailing lists,

tutorials, forums, case-studies, etc.). Other challenges

include the lack of support for

(i) the development of web-based modeling edi-

tors,

(ii) the development of editors that enable the use

of textual and graphical notations together,

(iii) integrating the editors with other systems and

technologies,

(iv) concurrent modeling and meta-modeling,

(v) managing changes on the meta-model,

(vi) re-using shared models acting as patterns,

(vii) deﬁning editors using an existing editor (or

its parts),

(viii) understanding the needs of the users and re-

ﬂecting them on the meta-model deﬁnitions that the

tools are based on,

(ix) the change management (i.e., changing meta-

models).

5 DISCUSSION

Our survey attracted 52 acceptable responses from the

interested practitioners of diverse industries. The top

industries from which we attracted participants are

computer, research, and IT and telecommunications.

Nearly half of the participants work as a software

developer/programmer, and 30% of the participants

work as a researcher.

In the rest of this section, we discuss the key ﬁnd-

ings that we obtained with the analysis of the survey

responses.

Model-Driven Engineering is the Practitioners’

Top-Motivation. Model-driven software engineering

- i.e., specifying, analysing, and transforming mod-

els - is the top-motivation of the participants devel-

oping graphical modeling editors (77%). Almost all

the participants develop graphical modeling editors

and perform the activities of model-driven engineer-

ing including model speciﬁcation, model analysis,

and model transformation. Also, most of the partic-

ipants (87%) integrate transformation tools to the edi-

tors, revealing that model transformation is the partic-

ipants’ top interest concerning the activities involved

in model-driven engineering.

Meta-Modeling Technologies are Used Much More

Than Programming Languages. 62% of the partic-

ipants who develop graphical modeling editors do so

using meta-modeling technologies, while the rest use

traditional programming technologies. The biggest

portion of the participants who use meta-modeling

technologies are from the research industry.

In the rest of this section, we discuss the practi-

tioners’ thoughts on developing graphical modeling

editors. We ﬁrstly consider the practitioners who

use programming languages (PP, standing for pro-

grammer practitioner) and then the practitioners using

meta-modeling technologies (MP, standing for meta-

modeler practitioner).

PP: Programming Languages Require no Learn-

ing Curve. Concerning the participants using pro-

gramming languages for developing graphical model-

ing editors, the top industries are the computer and IT

and telecommunications. The top-preferred program-

ming languages here are C# (40%), Python (40%),

and Java (30%). The participants’ top reason for us-

ing programming language is to do with the main-

tainability of the editors that can always be changed

and corrected by any programmer without any meta-

modeling knowledge (65%). Indeed, the survey re-

sults show that many participants preferring program-

ming languages have no idea about meta-modeling

technologies at all and consider meta-modeling tech-

nologies as causing difﬁcult-to-customise editors.

PP: Programming Technologies are Found Ad-

vanced and Portable. Many of the participants who

prefer programming technologies (45%) give empha-

sis on the advanced software libraries that come with

the programming technologies and facilitate the edi-

tor development via re-usuable software modules and

patterns. Also, many of the participants (50%) care

about building platform-independent editors that can

run on any operating systems (e.g., jar application),

which could be possible if the programming technolo-

gies were used.

PP: Meta-Modeling Technologies are Difﬁcult to

Extend. The main concern of the participants

who prefer programming languages over the meta-

Practitionersâ

Z Experiences on Developing Graphical Modeling Editors: A Survey

283

modeling technologies is about the meta-modeling

technologies’ limited extension support for better ad-

dressing the stakeholder needs. Indeed, customis-

ing the editor user interfaces (e.g., changing the lay-

out, fonts, and language) may not always be possi-

ble with the meta-modeling technologies. Also, the

participants pointed out the lack of support for the

open-source code availability, which could prevent

the meta-modeling technologies to be extended when

needed.

MP: Meta-Modeling Technologies Save a Lot of

Time and Effort. The top-used meta-modeling tech-

nology by meta-modelers is Sirius (50%), followed

by Metaedit+ (38%). Most of the meta-modeler prac-

titioners (75-84%) use meta-modeling technologies as

they save huge time and effort with the automation

support - automated generation of editors from the

language deﬁnitions and integration of the editor with

any code generators and model validators.

MP: Programming Languages Require Too Much

Time and Effort to Develop and Maintain Ed-

itors. Almost all the meta-modeler practitioners

(84%) agree that programming languages require too

much time and effort for developing editors by writ-

ing code. Also, maintaining the editor at code-level is

not found easy either (63%).

MP: Challenges Faced About the Meta-Modeling

Technologies. The meta-modeler participants face

with some issues while using the meta-modeling tech-

nologies. The commonly faced challenges are the in-

adequate support for customising the editor user inter-

faces, specifying complex meta-model concepts and

transformation algorithms, the meta-modeler com-

munity (e.g., mailing lists, tutorials, forums, case-

studies, etc.).

6 THREATS TO VALIDITY

We considered the construct, internal and external

threats to the validity of the survey results (Wohlin

et al., 2012). To minimise any threats against the con-

struct validity, we gave clear explanations at the be-

ginning of the survey indicating the purpose of the

survey and anonymity of the personal data. We also

performed a detailed pilot study to ensure that the sur-

vey questions are consistent and complete with regard

to the research questions of the study. To minimise

any internal validity threats, we did our best to choose

the participants as randomly as possible so as to avoid

any unknown variables that could affect the results.

Indeed, we shared the survey in many social platforms

(e.g., diverse linkedin groups) and give each enrolled

and interested participant an equal chance to partici-

pate. Lastly, the threats to the external validity were

minimised by reaching practitioners with various pro-

ﬁles that vary in terms of their industries, job posi-

tions, and experience levels.

7 CONCLUSION

In this paper, we conducted a survey among practi-

tioners so as to understand their experiences in de-

veloping graphical modeling editors. The survey at-

tracted 76 participants from diverse industries holding

diverse job positions and 52 of them were acceptable

for analysis.

According to the survey results, model-driven en-

gineering is the participants’ top-motivation for de-

veloping graphical modeling editors and it is also re-

vealed that most participants use editors to not only

specify some models but also transform the mod-

els into useful artifacts (e.g., software code, docu-

mentation, etc.). Given two techniques for devel-

oping graphical modeling editors (i.e., programming

languages vs. meta-modeling technologies), meta-

modeling technologies are much more preferred by

the participants (62%). Note that it is the research in-

dustry (and thus researchers) who show the greatest

interest on the meta-modeling technologies. Those

participants using programming languages prefer so

because programming languages require no learning

curve for them and are supported with advanced soft-

ware libraries. Also, editors developed with program-

ming languages can be used in any platform as a

standalone application portably. The same partici-

pants ﬁnd meta-modeling requiring a steep learning

curve and the corresponding technologies as difﬁcult

to use due to such limitations as customising the ed-

itor user interfaces and extending the technologies

when needed. The participants using meta-modeling

technologies ﬁnd the programming technologies as

requiring too much time and effort for the graphi-

cal editor development and reducing maintainability

due to the need for making all changes at code-level

which is error-prone. Lastly, the participants using

meta-modeling technologies face with several chal-

lenges too, including the inadequate support for editor

customisation, specifying complex meta-models and

transformation algorithms and the developers com-

munity (e.g., forums and tutorials).

In the near future, we are planning to validate the

survey results via an empirical study. That is, we

will determine two groups of practitioners where one

group has intermediate level of experience on pro-

gramming languages and the other group on the meta-

modeling technologies. Each group will be given

ICSOFT 2023 - 18th International Conference on Software Technologies

284

the same case-study and asked to develop a graphical

modeling editor. So, we will observe the effective-

ness and productiveness of the groups and see to what

extent the observed data justify the survey results.

REFERENCES

Agner, L. T. W., Soares, I. W., Stadzisz, P. C., and Sim

aO,

J. M. (2013). A brazilian survey on uml and model-

driven practices for embedded software development.

J. Syst. Softw., 86(4):997–1005.

Akdur, D., Garousi, V., and Demir

ors, O. (2018). A survey

on modeling and model-driven engineering practices

in the embedded software industry. Journal of Systems

Architecture - Embedded Systems Design, 91:62–82.

Berger, T., Rublack, R., Nair, D., Atlee, J. M., Becker, M.,

Czarnecki, K., and Wasowski, A. (2013). A survey of

variability modeling in industrial practice. In Proceed-

ings of the Seventh International Workshop on Vari-

ability Modelling of Software-intensive Systems, Va-

MoS ’13, pages 7:1–7:8, New York, NY, USA. ACM.

Bettini, L. (2013). Implementing Domain-Speciﬁc Lan-

guages with Xtext and Xtend.

Cabot, J. and Teniente, E. (2006). Constraint support in

MDA tools: A survey. In Rensink, A. and Warmer, J.,

editors, Model Driven Architecture - Foundations and

Applications, 2nd European Conference, ECMDA-FA

2006, Bilbao, Spain, July 10-13, 2006, Proceedings,

volume 4066 of Lecture Notes in Computer Science,

pages 256–267. Springer.

El Kouhen, A., Dumoulin, C., G

erard, S., and Boulet, P.

(2012). Evaluation of Modeling Tools Adaptation.

Technical report.

Erdweg, S., van der Storm, T., V

olter, M., Tratt, L.,

Bosman, R., Cook, W. R., Gerritsen, A., Hulshout, A.,

Kelly, S., Loh, A., Konat, G. D. P., Molina, P. J., Palat-

nik, M., Pohjonen, R., Schindler, E., Schindler, K.,

Solmi, R., Vergu, V. A., Visser, E., van der Vlist, K.,

Wachsmuth, G., and van der Woning, J. (2015). Eval-

uating and comparing language workbenches: Exist-

ing results and benchmarks for the future. Comput.

Lang. Syst. Struct., 44:24–47.

Fricker, R. D. (2008). Sampling methods for web and e-

mail surveys. The SAGE handbook of online research

methods, pages 195–216.

Groves, R. M., Fowler Jr, F. J., Couper, M. P., Lepkowski,

J. M., Singer, E., and Tourangeau, R. (2009). Survey

methodology. John Wiley & Sons, 2 edition.

Harel, D. and Rumpe, B. (2000). Modeling languages: Syn-

tax, semantics and all that stuff, part i: The basic stuff.

Technical report, ISR.

Kelly, S., Lyytinen, K., and Rossi, M. (2013). Metaedit+

A fully conﬁgurable multi-user and multi-tool CASE

and CAME environment. In Seminal Contributions to

Information Systems Engineering, 25 Years of CAiSE,

pages 109–129.

Kent, S. (2002). Model driven engineering. In Butler,

M. J., Petre, L., and Sere, K., editors, Integrated For-

mal Methods, Third International Conference, IFM

2002, Turku, Finland, May 15-18, 2002, Proceedings,

volume 2335 of Lecture Notes in Computer Science,

pages 286–298. Springer.

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

uhne, S. (2011). To-

wards a comparative analysis of meta-metamodels. In

Proceedings of the Compilation of the Co-Located

Workshops on DSM’11, TMC’11, AGERE! 2011,

AOOPES’11, NEAT’11, & VMIL’11, SPLASH ’11

Workshops, page 7–12, New York, NY, USA. Asso-

ciation for Computing Machinery.

uhne, T. (2006). Matters of (meta-)modeling. Software

and Systems Modeling, 5(4):369–385.

Malavolta, I., Lago, P., Muccini, H., Pelliccione, P., and

Tang, A. (2013). What industry needs from architec-

tural languages: A survey. IEEE Trans. Software Eng.,

39(6):869–891.

Ozkaya, M. (2018a). The analysis of architectural lan-

guages for the needs of practitioners. Softw., Pract.

Exper., 48(5):985–1018.

Ozkaya, M. (2018b). Do the informal & formal software

modeling notations satisfy practitioners for software

architecture modeling? Information & Software Tech-

nology, 95:15–33.

Ozkaya, M. (2019). Are the UML modelling tools power-

ful enough for practitioners? A literature review. IET

Softw., 13(5):338–354.

Ozkaya, M. and Akdur, D. (2021). What do practitioners

expect from the meta-modeling tools? A survey. J.

Comput. Lang., 63:101030.

Ozkaya, M. and Erata, F. (2020a). A survey on the practical

use of UML for different software architecture view-

points. Inf. Softw. Technol., 121:106275.

Ozkaya, M. and Erata, F. (2020b). Understanding practi-

tioners’ challenges on software modeling: A survey.

J. Comput. Lang., 58:100963.

Pech, V., Shatalin, A., and Voelter, M. (2013). Jetbrains

mps as a tool for extending java. In Proceedings of

the 2013 International Conference on Principles and

Practices of Programming on the Java Platform: Vir-

tual Machines, Languages, and Tools, PPPJ ’13, page

165–168, New York, NY, USA. Association for Com-

puting Machinery.

erez-Medina, J. L., Dupuy-Chessa, S., and Front, A.

(2007). A survey of model driven engineering tools

for user interface design. In Winckler, M., John-

son, H., and Palanque, P. A., editors, Task Models

and Diagrams for User Interface Design, 6th Interna-

tional Workshop, TAMODIA 2007, Toulouse, France,

November 7-9, 2007, Proceedings, volume 4849 of

Lecture Notes in Computer Science, pages 84–97.

Springer.

Popping, R. (2015). Analyzing open-ended questions by

means of text analysis procedures. Bulletin of Soci-

ological Methodology/Bulletin de M

ethodologie Soci-

ologique, 128(1):23–39.

Rozanski, N. and Woods, E. (2011). Software Systems Ar-

chitecture: Working With Stakeholders Using View-

points and Perspectives. Addison-Wesley Profes-

sional, 2 edition.

Practitionersâ

Z Experiences on Developing Graphical Modeling Editors: A Survey

285

Rumbaugh, J. E., Jacobson, I., and Booch, G. (1999).

The uniﬁed modeling language reference manual.

Addison-Wesley-Longman.

Seidewitz, E. (2003). What models mean. IEEE Software,

20(5):26–32.

Torchiano, M., Tomassetti, F., Ricca, F., Tiso, A., and Reg-

gio, G. (2011). Preliminary ﬁndings from a survey

on the md state of the practice. In Proceedings of

the 2011 International Symposium on Empirical Soft-

ware Engineering and Measurement, ESEM ’11, page

372–375, USA. IEEE Computer Society.

Viyovi

c, V., Maksimovi

c, M., and Perisi

c, B. (2014). Sir-

ius: A rapid development of dsm graphical editor. In

IEEE 18th International Conference on Intelligent En-

gineering Systems INES 2014, pages 233–238.

Wohlin, C., Runeson, P., H

ost, M., Ohlsson, M. C., and

Regnell, B. (2012). Experimentation in Software En-

gineering. Springer.

ICSOFT 2023 - 18th International Conference on Software Technologies

286