2025/10/03 by Zhao Zhang, Qingyuan Liang, Zhang, Zhao +15 · 1 citation
Computer Science · #Abstract syntax #Abstract syntax tree #Code (set theory) #Code generation #FOS: Computer and information sciences #Model-Driven Software Engineering Techniques #Parse tree #Parsing #Python (programming language) #Representation (politics) #Software Engineering (cs.SE) #Software Engineering Research #Syntax #Topic Modeling
paper · pdf · doi:10.48550/arxiv.2510.02887
published in arXiv (Cornell University) (Cornell University)
openalex publication_date 2025/10/03 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/05
Code generation has shown great promise in assisting software development. A fundamental yet underexplored question is how the choice of code representation affects model performance. While existing studies employ various representations, such as treating code as plain text, grammar rule sequences, or syntax tree sequences, they lack a principled understanding of the relationship between parsing difficulty and model effectiveness. This paper proposes a conjecture: the easier a representation is to parse, the better performance the model achieves. We formalize this idea using grammar classes, where representations in simpler classes (e.g., LL(1)) are easier to parse. Through a controlled experiment on a Python-based DSL, we show that parsing difficulty strongly correlates with model performance. Motivated by this finding, we present GramTrans, a general approach that automatically transforms a context-free language into a representation within the LL(1) class. GramTrans introduces a novel hierarchical conflict elimination algorithm, enabling a flexible trade-off between syntactic simplicity and token efficiency. We evaluate GramTrans on both Python and Java using three code generation models: StarCoder 1B, DeepSeek-Coder 1.3B, and Qwen2.5 1.5B. Across multiple benchmarks, GramTrans consistently delivers significant improvements over baseline representations. Furthermore, our analysis of existing representations reconfirms the strong alignment between parsing difficulty and model performance, providing additional support for the conjecture.