Code Understandability

Code understandability, crucial for software maintainability and collaboration, is a growing research area focusing on improving the readability and comprehension of automatically generated code, particularly from large language models (LLMs). Current research employs various techniques, including fine-tuning LLMs for code refactoring and developing novel metrics based on user feedback and code characteristics to predict understandability, often leveraging machine learning models like Random Forests. These efforts aim to enhance the quality of AI-assisted software development and improve the usability of computational notebooks, ultimately impacting software engineering practices and data science workflows.