A two-agent LLM pipeline with hierarchical merging explains COBOL functions, files, and projects, outperforming zero-shot baselines on several text-quality metrics.
X-COBOL: A Dataset of COBOL Repositories
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
Despite being proposed as early as 1959, COBOL (Common Business-Oriented Language) still predominantly acts as an integral part of the majority of operations of several financial, banking, and governmental organizations. To support the inevitable modernization and maintenance of legacy systems written in COBOL, it is essential for organizations, researchers, and developers to understand the nature and source code of COBOL programs. However, to the best of our knowledge, we are unaware of any dataset that provides data on COBOL software projects, motivating the need for the dataset. Thus, to aid empirical research on comprehending COBOL in open-source repositories, we constructed a dataset of 84 COBOL repositories mined from GitHub, containing rich metadata on the development cycle of the projects. We envision that researchers can utilize our dataset to study COBOL projects' evolution, code properties and develop tools to support their development. Our dataset also provides 1255 COBOL files present inside the mined repositories. The dataset and artifacts are available at https://doi.org/10.5281/zenodo.7968845.
citation-role summary
background 1
citation-polarity summary
fields
cs.SE 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Enhancing COBOL Code Explanations: A Multi-Agents Approach Using Large Language Models
A two-agent LLM pipeline with hierarchical merging explains COBOL functions, files, and projects, outperforming zero-shot baselines on several text-quality metrics.