Framework for Composition of Domain Specific Languages and the Effect of Composition on Re-use of Translation Rules

DSLs are programming languages that have been designed to be used to solve problems in a specific domain. They provide constructs that are high-level and domain-specific to make it easier to implement solutions in the given domain. They frequently also limit the language to the domain, avoiding general purpose constructs. One of the main reasons for using a DSL is to reduce the amount of work required for implementing new programs. To make the use of DSLs feasible, the cost of developing a new DSL for a domain has to be less than the total amount of cost saved by having the DSL. Thus, reducing the cost of developing new DSLs means that introducing DSLs becomes feasible in more situations. One way of reducing costs is to use composition techniques, where new languages are created from existing ones. This includes defining new language constructs in terms of existing ones, combining the constructs from one or more existing languages, and redefining existing constructs. We present a framework for composing languages on the abstract level and discuss to which degree one can ensure that languages produced by the composition language are valid. In particular, we look at how translation rules for translating from a composed language to a GPL are affected by the composition. That is, to which degree can a language composed from other languages reuse the translation rules of the languages it is composed from. We use a patience game suite as a case-study to show how our composition techniques can be used and demonstrate the short-comings of the techniques. We also show how a tool for composing languages can be created using DSLs produced by composition. The implementations are all in Java.