Bridging the Gap Between AI and Reality • Rhodes, Greece
Time: Sunday, 2.11
Room: Room A
Authors: Daniel Sami Mitwalli
Abstract: The development of graphical domain-specific languages (DSLs) typically in- volves implementing the language using either general-purpose languages (GPLs) or specialised meta-DSLs [5] [2]. Throughout this process, developers define ab- stractions, their concrete syntactic representations, and the associated semantics. While effective, this approach can be time-consuming due to the time required to learn how to implement with the underlying meta-DSLs/GPLs, as well as the manual work involved in specifying and refining language details. In response to these challenges and the growing popularity of the Model Context Protocol (MCP) [1], this paper examines how Large Language Models (LLMs) can assist language engineers in developing graphical DSLs by leveraging meta-DSLs within an MCP application. Furthermore, it is evaluated to what extent natural-language-driven specification and refinement can reduce the effort required in language engineering.