一句话：提出基于再生与交叉验证的SE工具，检测并缓解代码、文档与测试之间的不一致。
代码、文档与测试之间的不一致是软件工程中长期存在的问题，会降低软件质量、拖慢开发并增加维护难度。三类工件从不同但重叠的视角描述同一软件行为，构成三角关系，原则上任一工件都可用于再生其他工件。该博士项目旨在利用这种关系，通过基于再生的技术检测并最终缓解工件间的不一致。具体地，项目将构建一个SE工具，利用原始版本与再生版本进行三角定位，并通过交叉验证提高可靠性；同时特别关注降低误报，确保报告给开发者的问题可信且可操作，尤其是在真实项目中。
Inconsistencies between different software artifacts, such as source code, documentation, and tests, are a common and long-standing problem in software engineering. These misalignments can degrade software quality, slow down development, and hinder maintenance. Each of these artifacts provides a distinct yet overlapping perspective of the same software behavior, forming a triangular relationship in which any one artifact can, in principle, be used to regenerate the others.In this PhD project, we aim to exploit these relationships through regeneration-based techniques to detect and ultimately mitigate inconsistencies across software artifacts. The approach focuses on building an SE-tool that leverages original and regenerated versions of artifacts to triangulate inconsistencies and improve reliability through cross-validation. A particular emphasis is placed on reducing false positives, ensuring that reported issues are trustworthy and actionable, especially when presented to developers in real-world projects.
DOI 原文 ·
@inproceedings{paperbot9515,
title = {Detecting and Mitigating Inconsistencies Between Code, Documentation and Tests},
author = {Tobias Kiecker},
booktitle = {ASE},
year = {2025},
doi = {10.1109/ase63991.2025.00397}
}