About

Domain-specific modeling (DSM) makes software more accessible and inclusive by leveraging the expertise and knowledge of domain experts from various backgrounds, such as insurance or healthcare, instead of relying solely on programmers and software engineers. This modeling practice allows the solution to be specified directly using problem domain concepts. Ref. In the activity of DSM, models are the primary artifact manipulated and produced using domain-specific languages (DSL) Info DSL overview . Contrary to general-purpose programming languages (GPLs) E.g. Java, Python , DSLs are languages tailored to a specific application domain. The domain may be restricted to a device (embedded system), a specific project E.g. meeting scheduler application or it may be common to a larger group of experts and practitioners, such as the hardware description language VHDL or the web languages CSS and HTML. They offer substantial gains in expressiveness and ease of use compared with GPLs in their domain of application. Ref.

Those modeling and language activities are supported by language workbenches. They support the efficient definition, reuse, and composition of DSLs and their IDEs Ref.. One key aspect of their differences lies in their modeling editors, which can be classified as parser-based and projectional.
Most language workbench editors use a parser-based approach, with a textual or graphical syntax E.g. Xtext, Spoofax, MetaEdit+ . A parser-based editing approach relies on a parser to continuously scan the input to build the an abstract syntax tree (AST), validated with a grammar. However, this approach limits the evolution of DSLs and is prone to syntactic errors.
With a projectional editing approach E.g. Jetbrains MPS, Whole, Scratch the AST is modified directly as the user edits a program, thus preventing syntactic errors. The approach enables the support of notations that cannot be easily parsed, such as tables, forms or diagrams, and the composition of any language without introducing syntactic ambiguities.