FOCUS: PROGRAMMING LANGUAGES Evolution of Software Systems with Extensible Languages

offers a process for gradually integrating domain-specific languages into existing software systems, thereby improving system maintenance. // SOFTWARE SYSTEMS CONTINU-OUSLY change and grow in com-plexity.1 It comes as little surprise, then, that most development time and money isn’t spent on the initial design and implementation of new software systems but on maintaining existing ones.2,3 Domain-specific languages (DSLs) offer a way to reduce maintenance costs. A DSL is a programming language specifically designed to support an application by provid-ing domain concepts as language constructs. This enables program-mers to map knowledge into source code and vice versa, which simpli-fies the creation, comprehension, and maintenance of domain-specific programs.4–7 The vast majority of existing software systems don’t use DSLs. Instead, the long-standing suc-cess of C, C++, and Java has led to large procedural and object-ori-ented systems that don’t benefit from the maintenance advantages that DSLs provide. We propose a new direction for programming lan-guage research—namely, to develop mechanisms and tools that support integrating the powerful DSL tech-nologies we’re already developing into existing software systems whose maintenance is so important. In this article, we describe how the extensible programming lan-guage SugarJ supports the gradual evolution of existing software sys-tems to apply DSLs and their main-tenance advantages, we report on our experience of evolving and im-proving two existing software sys-tems written in Java (Java Pet Store and Eclipse) by applying three DSLs, and we outline a research roadmap for developing tools that support maintainers in introducing DSLs into existing software systems