首页    期刊浏览 2024年07月08日 星期一
登录注册

文章基本信息

  • 标题:A Framework for Automatically Mining Source Code
  • 本地全文:下载
  • 作者:Shaheen Khatoon ; Guohui Li ; Rana Muhammad Ashfaq
  • 期刊名称:Journal of Software Engineering
  • 印刷版ISSN:1819-4311
  • 电子版ISSN:2152-0941
  • 出版年度:2011
  • 卷号:5
  • 期号:2
  • 页码:64-77
  • DOI:10.3923/jse.2011.64.77
  • 出版社:Academic Journals Inc., USA
  • 摘要:Mining source code by using different data mining techniques to extract the informative patterns like programming rules, variable correlation, code clones and frequent API usage is an active area of research. However, no practical framework for integrating these tasks has been attempted. To achieve this objective an integrated framework is designed that can detect different types of bugs to achieve software quality and assist developer in reusing API libraries for rapid software development. Proposed framework automatically extracts large variety of programming patterns and finds the locations where the extracted patterns are violated. Violated patterns are reported as programming rule violation, copy paste code related bugs and inconsistent variable update bugs. Although, the bugs are different but the framework can detect these bugs in one pass and produces higher quality software systems within budget. The framework also helps in code reusing by suggesting the programmer how to write API code to facilitate rapid software development. Proposed framework is validated by developing a prototype that developed in C# (MS Visual Studio, 2008) and evaluated on large application like ERP. Results shows proposed technique greatly reduced time and cost of manually checking defects from source code by programmers.
国家哲学社会科学文献中心版权所有