一句话：提出 ADATracer，面向 Ada 安全关键系统，自动恢复需求、源码、提交与问题之间的追溯链接。
在航空电子、航天、汽车等安全关键领域，Ada 常用于认证驱动开发，但将自然语言需求、Ada 源码、版本控制提交与问题跟踪工件自动关联仍多靠人工，Ada 追溯工具也较少。本文提出 ADATracer：它结合关键词检索、语义向量嵌入与 Neo4j 知识图谱，并含 Ada 感知解析器，可处理包规范、包体与强类型等特性；支持追溯链接恢复和图探索，以检查从需求到实现与问题解决的证据链。作者将其应用于开源 Ada Web Application 框架，识别追溯链接，为认证驱动的 Ada 开发提供支持，并为安全关键系统研究提供框架。
Traceability is an important part of safety-critical software development in areas such as avionics, space, and automotive engineering. Ada is often used in these systems. It shows how requirements are implemented in code and how changes are reflected in commits and issue reports. Automated traceability tools for Ada are not as common as those for other programming languages. Currently, the systematic linking of requirements to Ada source files, version-control history, and issue artifacts is often a manual process that requires effort to maintain throughout the software lifecycle. We present ADATracer, a software traceability tool designed to recover links between natural language requirements, Ada source files, version-control commits, and issue-tracking artifacts. The tool uses keyword-based retrieval, semantic vector embeddings, and a Neo4j-based knowledge graph. It includes an Ada-aware parser that accounts for language features such as package specifications, bodies, and strong typing. ADATracer supports traceability recovery and graph-based exploration, allowing for the inspection of evidence chains from requirements to implementation and issue resolution. We apply ADATracer to the open-source Ada Web Application (AWA) framework, where the tool identifies traceability links. This work facilitates traceability in certification-driven Ada development and provides a framework for further study in safety-critical software systems. The source code and demonstration video are publicly available at https://github.com/rishalab/ADATracer and https://youtu.be/sb66cEffF-o.
@inproceedings{ChepuriTTCS26,
title = {ADATracer - Towards A Software Traceability Pipeline for ADA Systems},
author = {Nithin Chepuri and Rajeev Thota and Venkat Kaushal Thippisetty and Sridhar Chimalakonda and Sumith Shankar},
booktitle = {FSE},
year = {2026},
doi = {10.1145/3803437.3806431}
}