Scholarlink

Software testing is essential for ensuring system reliability. However, some systems face the Oracle Problem, where the correctness of outputs cannot easily be determined. In cybersecurity, password-based authentication is widely used, and tools such as John the Ripper (JtR) can expose vulnerabilities. This paper applies Metamorphic Testing (MT) and Metamorphic Exploration (ME) to JtR, identifying flaws and enhancing system understanding. A web-based Open Educational Resource (OER) was also developed with interactive features to promote the use of MT and support learner engagement. The OER includes quizzes for self-assessment and a feedback mechanism explaining incorrect answers. An initial attempt was made to use Large Language Models (LLMs) to generate on-demand quiz questions and feedback. However, survey participants noted that AI-generated questions were too simple and often contained context clues. This paper will be of interest to educators and researchers in software testing, cybersecurity, and AI-assisted education.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Enhancing Learning Through a Web-Based OER: Application of Metamorphic Testing to John the Ripper

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Lindsay Elle Chiara Song,
  • Anson Hwong Lee,
  • Dave Towey

摘要

Software testing is essential for ensuring system reliability. However, some systems face the Oracle Problem, where the correctness of outputs cannot easily be determined. In cybersecurity, password-based authentication is widely used, and tools such as John the Ripper (JtR) can expose vulnerabilities. This paper applies Metamorphic Testing (MT) and Metamorphic Exploration (ME) to JtR, identifying flaws and enhancing system understanding. A web-based Open Educational Resource (OER) was also developed with interactive features to promote the use of MT and support learner engagement. The OER includes quizzes for self-assessment and a feedback mechanism explaining incorrect answers. An initial attempt was made to use Large Language Models (LLMs) to generate on-demand quiz questions and feedback. However, survey participants noted that AI-generated questions were too simple and often contained context clues. This paper will be of interest to educators and researchers in software testing, cybersecurity, and AI-assisted education.