Estudo Comparativo de Parsers LL(1): desempenho e eficiência em compiladores

Abstract

Since 1952, compilers have played a significant role in computing. As such, concerns about performance, reliability, and the evolution of their core functions are ongoing, being a central topic

of discussion and innovation both in the software development industry and in academia. Major technology companies like Google and Microsoft develop their own programming languages and continuously invest in compiler research. Syntax analysis, an essential phase of compilation, transforms the source code into a structure that can be interpreted and optimized. In this context, parsers play a central role, ensuring that the code conforms to a defined grammar. Grammar defines all the rules that validate whether a program is accepted by a language, with the LL(1) class representing deterministic grammars that can be parsed predictively, reading input from left to right and deriving using a single lookahead symbol. Among the established models in the literature, both Recursive Descent Parsers and Table-Driven Parsers have been compared in terms of performance, memory consumption, and implementation complexity, with both being implemented in a project that uses a language defined by a deterministic grammar. Thus, this work aims to assist in choosing the most suitable parser based on the language’s grammar and the project’s requirements.