A Model-Driven Parser Generator, from Abstract Syntax Trees to Abstract Syntax Graphs

TL;DR

This paper introduces a model-driven parser generator that creates abstract syntax graphs from language models, enabling more flexible and conceptually aligned language processing without traditional grammar constraints.

Contribution

It presents a novel approach that derives syntax graphs directly from language models, supporting complex references and avoiding conventional grammar limitations.

Findings

Automatically generates abstract syntax graphs from models

Supports cyclic and recursive references in syntax graphs

Enhances language design flexibility

Abstract

Model-based parser generators decouple language specification from language processing. The model-driven approach avoids the limitations that conventional parser generators impose on the language designer. Conventional tools require the designed language grammar to conform to the specific kind of grammar supported by the particular parser generator (being LL and LR parser generators the most common). Model-driven parser generators, like ModelCC, do not require a grammar specification, since that grammar can be automatically derived from the language model and, if needed, adapted to conform to the requirements of the given kind of parser, all of this without interfering with the conceptual design of the language and its associated applications. Moreover, model-driven tools such as ModelCC are able to automatically resolve references between language elements, hence producing abstract syntax graphs instead of abstract syntax trees as the result of the parsing process. Such graphs are not confined to directed acyclic graphs and they can contain cycles, since ModelCC supports anaphoric, cataphoric, and recursive references.