引入错误级别的测试集压缩模型研究

开发者每一次对软件的修改都会导致测试集的增长。在进行回归测试时，如果重新运行所有的测试案例，成本将是极其昂贵的。于是，一些压缩测试集的算法被提出。这些算法能够找出一个与原测试集有相同覆盖度的压缩测试集，但这些压缩后的测试集在查错方面的性能不十分令人满意。引入错误级别的概念可以对测试案例的查错能力进行更为为细致的划分。本文通过引入错误级别，改进了一个双标准测试集压缩模型，并提出了一个新的测试集压缩算法。改进后的双标准测试集压缩模型能提供关于测试案例查错能力的更具体的信息。新算法有一个重要的特点：在任何时候中断算法，仍然能得到一个满足覆盖度的压缩测试集。

A Study of the Test-Suite Reduction Models with Fault Severity Levels

Every change to an evolving program will cause the size of the test-suite to grow. Rerunning all of the test cases in   nsive. So some algorithms are invented to reduce the test-suite. These algorithms can find out a reduced test-suite that provides the same coverage degr ee as the original test-suite. However, the capability of the reduced test-suite to reveal errors is not very satisfying. The capability of the test cas es to detect errors can be given a more accurate evaluation by incorporating the fault severity levels. In this paper, we incorporate the fault severity levels into a bicriteria model for test-suite reduction to improve the model＇s performance on fault-detection, and then propose a new testsuite reducttion algorithm. The improved model can provide more detailedalgorithm has a feature that the reduced test suite satisfies the adequate coverage degree no matter when the algorithm is terminated.