CodeOCR: On the Effectiveness of Vision Language Models in Code Understanding - Explained Simply | ArXiv Explained