Domain Specific Languages (DSLs) provide interesting characteristics that align well with the goals and mission of model-driven software engineering. However, there are still some issues that hamper widespread adoption. In this chapter, the authors discuss two of these issues. The first relates to the vagueness of the term DSL, which they address by studying the individual terms: domain, specificity, and language. The second is related to the difficulty of developing DSLs, which they address with a view to making DSL development more accessible via processes, standards, and tools.
TopIntroduction
The concept of Domain Specific Languages (DSLs) is not new, and the advantages to using them have long been highlighted in the literature:
We must develop languages that the scientist, the architect, the teacher, and the layman can use without being computer experts. The language for each user must be as natural as possible to her/him. The statistician must talk to his terminal in the language of statistics. The civil engineer must use the language of civil engineering. When a man learns his profession he must learn the problem-oriented languages to go with that profession (Martin, 1967, p. 89).
We must constantly turn to new languages in order to express our ideas more effectively. Establishing new languages is a powerful strategy for controlling complexity in engineering design; we can often enhance our ability to deal with a complex problem by adopting a new language that enables us to describe (and hence to think about) the problem in a different way, using primitives, means of combination, and means of abstraction that are particularly well suited to the problem at hand (Abelson, Sussman, & Sussman, 1996, pp. 359-360).
The use of DSLs is not limited to information technology, as they are used in many other areas as well, such as finance (Arnold, Van Deursen, & Res, 1995), chemistry (Murray-Rust, 1997), biology (Hucka et al., 2003), music (Boulanger, 2000), for example. Among the DSLs most commonly used in computer science today are the Structured Query Language (SQL) (Chamberlin & Boyce, 1974; ISO, 2008), the regular expression language for manipulating strings (Friedl, 2006), Microsoft Office Excel (Microsoft, 2011), and the eXtensible Markup Language (XML) (ISO, 1986).
A DSL is usually designed to solve a specific class of problems in a particular domain. The focus of DSLs on a particular domain facilitates the creation of languages that best represent the domain concepts. This convergence between problem domain and solution domain has many benefits, in terms of the expressiveness and precision (semantics) of the DSL. In addition, DSLs have also shown good potential in terms of productivity, reusability, and reliability (Kelly & Tolvanen, 2008; Kleppe, 2008).
In this chapter, we discuss the concept of DSLs (section 1), and present the types of DSLs (section 2) and the tools used to develop them (section 3). Section 4 presents some of the standards that can be used as a foundation for developing DSLs. Section 5 describes the development process itself. Finally, we summarize this work in section 6.
TopBackground
As mentioned earlier, DSLs have been a part of the computing world for decades. SQL and the UNIX languages awk and make are a few examples. These languages were developed by specialists with solid language development skills and good knowledge of the DSL domain. However, the emergence of the model driven approach and the need for languages capable of producing precise models that can be processed by machines, has opened up new horizons for DSLs. In the future, these languages could play a central role in the software development cycle, which would take DSLs from the arena of specialists to that of software developers.