Understanding the Successes and Challenges of Model-Driven Software Engineering - A Comprehensive Systematic Mapping

Abstract

Model-Driven Software Engineering (MDSE) is a development method in which models are used to generate software. Despite documented advantages, projects employing MDSE may fail due to development challenges. In this paper, we study and document these challenges by conducting an up-to-date systematic mapping that goes beyond reviewing MDSE studies: we also include two derived paradigms (Model-Oriented Programming and Models at Run-time). Therefore, we present a systematic mapping with two objectives: The first objective was to identify specific domains in which MDSE is successful, while the second objective was to identify what are the challenges to apply this methodology to general purpose development processes. Following the review of 4859 studies (3727 are unique), we have identified the application and technological domains in which MDSE projects are more likely to succeed. We also discuss challenges presented by 17 primary studies. The analysis of the results indicate that MDSE application is consolidated in specific domains. A common feature identified among studies related to general purpose MDSE processes is that, initially, authors reported lack of proper methods and training. After new techniques have risen, it has been pointed that MDSE projects still face maintenance problems that can discourage their usage in other domains.