DESIGN FOR SIX SIGMA (DFSS) TEAM AND TEAM SOFTWARE PROCESS (TSP) - Software Design for Six-Sigma: A Roadmap for Excellence - page 269

TABLE 10.7 Complex and Large Size-Project Result

Results using PSP and TSP
Complex and Large Size Project

Project Plan Actual

Size (LOC) 90000 95000

Effort (People) 20 22

Schedule (Weeks) 20 26

Project Quality
(Defect/KLOC removed in phase)
Complex and Large-Size Project

Integration 0.005 Defects/KLOC 0.006 Defect/KLOC

System Test 0.0025 Defect/KLOC 0.002 Defect/KLOC

Field Trial 0 Defects/KLOC 0.001 Defect/KLOC

Operation 0 Defects/KLOC 0.001 Defect/KLOC

Following PSP and TSP provided a very good initialization during the early stage
of the project, whereas it also was realized that various important aspects of the
software process method during the middle and later stages were not going to be
fulfilled as observed during previous applications of PSP and TSP for moderate and
medium-sized software projects. Since the project did not have a long life cycle,
it was agreed to follow the concepts of other software process and methods. The
shortcomings and possible improvisation to PSP and TSP are discussed in Chapter
2. In addition to the above views, while following PSP and TSP, it posed challenges
to use the process methods while working with cross-functional teams and suppliers
that were based globally. As shown in Table 10.7, the results were near to the plan
but not encouraging compared with Six Sigma. The reliability was less than industry
acceptable standards, which was proved during the series of vehicle-level testing.
It was then determined to analyze current design, find out the flaws, and determine
possible resolutions.

10.5 THE RELATION OF SIX SIGMA TO CMMI/PSP/TSP
FOR SOFTWARE

Various researchers have experience with PSP/TSP, CMMI, and Six Sigma in the
area of software systems in terms of complexity affecting reliability and safety,
human errors, and changing regulatory and public views of safety. Although PSP/TSP
covers the engineering and project management process areas generally well, they do
not adequately cover all process management and support process areas of CMMI.
Although a few elements of the Six Sigma for Software toolkit are invoked within
the PSP/TSP framework (e.g., regression analysis for development of estimating
models), there are many other tools available in the Six Sigma for Software toolkit
that are not suggested or incorporated in PSP/TSP. Although PSP/TSP refers to and
may employ some statistical techniques, specific training in statistical thinking and