Towards a Corpus of Use-Cases for Model-Driven Engineering Courses - Paper 03

Sep 11, 2026

Uploaded by Bintang

Drawing on years of teaching Model-Driven Engineering, the authors examine why choosing the right use-cases is crucial for student engagement. The paper diagnoses common weaknesses in typical course examples—such as unrealistic domains, artificial development processes, and non-iterative scenarios that undermine cost-effectiveness. It then proposes alternative use-cases and offers guidelines to keep them pragmatic, motivating, and achievable without unnecessary complexity.

LanguageEnglish

CategoryPresentations

FormatPDF

Pages5

Reading Time2 min

AccessFree to download

Abstract

1 Introduction

2 Background

Unrealistic Domains

Artiﬁcial Development Processes

Non-Iterative Development Scenarios

3 Towards a Corpus of Use-Cases for MDE Courses

3.1 Auto-Synchronised (Opposite) References in Java

Towards a Corpus of Use-Cases for Model-Driven Engineering Courses Dimitrios S. Kolovos 1 and Jordi Cabot2 1 University of York, UK – [dimitris.kolovos@york.ac.uk](dimitris.kolovos@york.ac.uk) 2 ICREA UOC, Spain – [jordi.cabot@icrea.cat](jordi.cabot@icrea.cat) Abstract. Having taught Model-Driven Engineering courses for a number of years, in this paper we reﬂect on the importance of selecting appropriate use-cases for students to explore related principles and technologies. We discuss examples on both ends of the spectrum and we present guidelines for selecting use-cases that are pragmatic and motivating without being excessively complex. 1 Introduction While the value of teaching software modelling and model-driven engineering at University level is increasingly recognised [1], eﬀective teaching continues to bea challenging task. The authors of this paper have been somewhat unsuccessful in the past in convincing students about the potential of MDE. We believe a key reason for this is the poor choice of use-cases and examples we show in class. Based on our experience and on extensive discussions with colleagues, in this paper we wish to propose examples of alternative use-cases that have a potential to overturn the current situation. 2 Background In our experience, to maintain the interest of students in a course they need to see practical value in the technologies and techniques they are taught. For software engineering courses this makes the selection of motivating use-cases a critical point. On top of the bad modelling practices discussed in [2], we have identiﬁed that use-cases we use often suﬀer from the following weaknesses that can end up demotivating students instead of sparking their interest. Unrealistic Domains In an attempt not to overwhelm students with the complexity of developing fully-functional systems, educators often use unrealistic application domains. For example, a common use-case – variants of which we have used in the past – is to ask students to develop a DSL for modelling book libraries 1 and then write model-to-text transformations that generate HTML reports from library models or deﬁne additional validation constraints. This usecase has two advantages: the DSL contains concepts that students can relate to, [1](1 help.eclipse.org/mars/topic/org.eclipse.emf.doc/tutorials/clibmod/clibmod.html)[ help.eclipse.org/mars/topic/org.eclipse.emf.doc/tutorials/clibmod/clibmod.html](1 help.eclipse.org/mars/topic/org.eclipse.emf.doc/tutorials/clibmod/clibmod.html) and it exercises most of the features of metamodelling technologies (inheritance, containment/non-containment/opposite references) . On the other hand however, it is rather unlikely that one would implement a real-world library management system using models instead of e.g. a relational database as a data persistence format. Artiﬁcial Development Processes Another common use-case to demonstrate model-to-model transformation is the (infamous) UML to RDBMS example. Here, students are typically asked to develop a transformation that produces a relational schema from a UML class diagram. This is a complex transformation that exercises many features of contemporary transformation languages, however, at the end of the process students end up with a model of a relational schema that is of little practical use. An additional model-to-text transformation can be used to produce SQL that can set up a database, however, this isnot very helpful either as in practice students would then still need to interact with the database using low-level SQL commands. Extending the use-case and asking students to also generate code that can provide a high-level object oriented interface to the database from ﬁrst principles is a rather complex task and the results are unlikely to be of comparable quality to object-relational frameworks like Hibernate. An alternative would be to ask students to produce e.g. Hibernate-based code, however this means that a co

FAQs

Why does use-case selection matter in Model-Driven Engineering courses?

Effective teaching depends on students seeing practical value in what they learn. Poorly chosen use-cases can fail to convince students about MDE’s potential.

What weaknesses in common MDE use-cases can demotivate students?

The paper highlights unrealistic domains, artificial development processes, and non-iterative scenarios that make students question cost-effectiveness.

What approach does the paper propose for improving course use-cases?

The authors propose alternative use-cases and provide guidelines that aim to be pragmatic and motivating while avoiding excessive complexity.

Sep 11, 2026

Uploaded by Bintang