| Summary: | 碩士 === 大同工學院 === 資訊工程研究所 === 81 === Program testing is an important step for improving software
quality and reliability. Many test strategies are proposed and
used today, but how to compare them to show which one is more
cost-effective than others ? In this article, we establish a
formula to evaluate the averaged fault-detecting-ability (FDA)
per test case of test strategy M under some test sets of M, to
be an index of the efficiency of a test strategy. In this
formula , we consider not only the effectiveness of test
strategy but also the cost of test. Since the size of test set
is often too large, we propose a method, partitionong the
input domain into subdomains, to reduce the size of test
set for the tester. Finally, we experiment with various
test strategies for some programs, getting the comparative
basis from the results, and concluding which test strategy
is better than others among these test strategies.
|
|---|
