Browsing by Author "Srikanth, Hema Lakshman"

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • No Thumbnail Available
    ADaPT: Adaptive Development and Prototyping Technique
    (2002-08-28) Srikanth, Hema Lakshman; Dr. Ana (Annie) I. Anton, Committee Chair
    Adaptive prototyping focuses on developing software for rapidly changing environments while improving delivery speed. Traditional methodologies are not effective in highly volatile environments; thus, agile methodologies have gained acceptance recently. Although agile methodologies offer less bureaucratic software processes, they fail to adequately support requirements engineering best practices. Adaptive prototyping provides a balance between heavy and ad hoc processes, aims to eliminate the drawbacks inherent in Agile Methodologies and traditional prototyping; it also incorporates requirements engineering best practices. This thesis proposes a methodology, ADaPT (Adaptive Development and Prototyping Technique), which employs scenario analysis to elicit and validate requirements; maintains the spirit of CMM-level 2; iteratively re-examines system requirements; and only documents essential artifacts. Significant emphasis is placed on testing; and acceptance tests are written before implementation. Initial validation efforts, in the form of post-project surveys, suggest that ADaPT can improve system delivery speed and quality. Surveys were administered to three groups: customers, instructors, and students. Customers surveyed agreed that sponsored teams: delivered their system on time, developed a high quality system, and produced useful artifacts. Instructors surveyed strongly agreed that all projects were completed successfully, met course requirements and ensured a highly satisfied customer. The collective application of these techniques appears to improve software quality, reduce software cost, and improve system delivery speed while enforcing requirements engineering best practices as compared with previous experiences in student projects.
  • No Thumbnail Available
    Value-Driven System Level Test Case Prioritization
    (2006-04-17) Srikanth, Hema Lakshman; Laurie Williams, Committee Chair
    Software testing is an expensive process often consuming at least 50% of the total development cost. Among the types of testing, system testing is the most expensive and complex as it involves configuring multiple complete integrated systems to closely emulate customer needs. System testing involves validating system compliance with its specified requirements. Companies are frequently faced with budgetary constraints, which may limit their ability to effectively complete testing efforts before delivering a software product. The Prioritization of Requirements for Test (PORT), a system-level approach to test case prioritization, builds upon prior test case prioritization research. PORT prioritizes system test cases based on four factors for each requirement: requirements volatility, customer priority, implementation complexity, and fault proneness. Test cases for requirements with higher priority based upon a weighted average of these four factors are executed earlier in the system test. A feasibility study was conducted on four similar student projects at graduate software testing class at North Carolina State University. Also, three post-hoc industrial case studies of PORT were conducted with Tekelec and I-Cubed. The results indicate that PORT can be used to improve the rate of failure detection when compared with a random and operational profile-driven random approach. Furthermore, the contribution of the prioritization factors towards the improved rate of failure detection was investigated and the results show that customer priority was the most significant contributor, followed by implementation complexity and fault proneness. The results indicate that PORT can be used to improve test efficiency. Tool support is provided for the PORT scheme via the Requirements-Based Testing (ReBaTe) tool. ReBate provides a ranking of tests and also allows for end-to-end traceability between requirements, test cases, and defects.