A Review on Detecting and Managing Documentation Drift in Software Development

Maintaining alignment between software code and specifications is a persistent challenge in the software development lifecycle. Outdated or inaccurate documentation hinders effective use, misleads developers, and increases vulnerability risks. This paper reviews literature on code-documentation drift, highlighting the need for robust synchronization and traceability. It examines solutions ranging from heuristic methods and synchronization algorithms to AI-driven tools, multi-agent systems, and machine learning for sustaining continuous consistency between evolving code, comments, and formal documentation. The review also adds a comparative analysis of approaches, discusses industry adoption and barriers, and summarizes recent advances in 2024-2025-particularly those leveraging Large Language Models (LLMs).