Neural Code Translation of Legacy Code: APL to C#

TL;DR

This paper explores neural translation of APL to C# using large language models, introducing a novel framework with guided strategies and an automated evaluation pipeline to improve translation quality.

Contribution

It presents a new APL-to-C# translation framework with guided strategies and a comprehensive evaluation pipeline, addressing challenges of syntax sparsity and limited data.

Findings

Guided strategies outperform baseline direct translation.

Automated evaluation verifies both syntax and functionality.

Neural translation effectively bridges APL and C# for diverse programs.

Abstract

Automatic translation between programming languages remains a challenging problem, particularly when the source language is highly concise and specialized. This paper investigates the translation of APL into C# using large language models. The task is difficult due to APL's sparse syntax, the scarcity of large-scale parallel corpora, and the requirement for specialized knowledge to interpret APL programs. To address these challenges, we introduce a novel framework for APL-to-C# translation by comparing three guided strategies, namely natural language description-mediated, retrieval-augmented, and iterative refinement, against a baseline direct translation model. We constructed multiple datasets of functionally equivalent code pairs spanning various levels of complexity, and to rigorously assess translation quality, we developed an automated evaluation pipeline that verifies both syntactic compilation and functional execution of the generated C# code. Our results demonstrate that neural code translation can successfully bridge the gap between APL and C# for a wide range of programs, and that incorporating additional context and guidance significantly improves model performance.