Quantitative process models for measurement and control of software test
Nancy Eickelmann · 1997
This dissertation focuses on improving the software test process, which is consistently identified as deficient in industry surveys. To address software test process deficiencies, software test experts such as Gelperin, Hetzel, and Beizer have espoused a life cycle software test process with a defect prevention focus. Moreover, the National Institute of Standards and Technology (NIST) adopted the life cycle software test process under the Federal Information Processing Standard 101 in 1983. Over a decade later, however, it is still atypical for industry to employ a life cycle software test process. To better understand why this gap exists between state of the art and state of the practice, the Software Test Environment Pyramid (STEP) Model is developed and used to evaluate the functionality necessary to support a life cycle test process; test planning and test configuration management are found to be essential. Moreover, a glaring deficiency found in software test planning is the inability to estimate and allocate required test resources accurately. To address that deficiency, the Software Test Estimation Model (STEM) is developed as part of this dissertation. STEM provides the critically needed support for more accurate test planning, thus providing an enabling technology for a life cycle software test process.