A Full Software Development Lifecycle Testing Project - Software Testing: Testing Across the Entire Software Development Life Cycle - page 226

Regardless of the sequence of screen development and testing, you would expect
to see the same or very similar defect detection peak in both plots. The comparison is
trending toward the DSA having a lower defect discovery peak (20) than the other CPI
projects (25). Although the relative difference is 20%, the actual difference is only
fi ve defects, which may be too small a difference to prompt further investigation.
At day 26 the plots intersect. Based on the lag in the peak of DSA discoveries, it
is reasonable to expect the DSA plot to remain higher (larger number of defects dis-
covered per day) than the CPI projects for the remaining DSA project days. Although
this trend strongly implies missed defects by DSA project end, it does not indicate
what kinds of missed defects there might be. A root cause analysis of the 334 defects
discovered so far might shed some light on the expected nature of the missed defects.
Figure 13.4 shows a rough analysis of the code earmarks for these 334 defects on the
DSA defect log.

Root cause analysy of defects

334 total defects corrected to date

Defects
corrected per
code earmark

Percent defects
corrected per
code earmark

Code earmarks

DSA1.01S 122

70
61
56
25
334

37%
21%
18%
17%
7%

DSA1.02S

DSA3.01S
DSA3.02S
All others
Total

Figure 13.4

DSA Project showing defect tracking log corrected as of a day 28

As stated before, the screens with the largest number of input or output fi elds
such as the certifi cates data entry screen (DSA1.01S) are expected to contribute
the largest number of defects. Having said that, the screen with the second larg-
est number of defects is expected to be the certifi cate search screen (DSA1.03S)
or instructor data entry screen (DSA2.01S), but neither code earmark appears at
the top of the list. Instead, the screen with the second highest number of defects is
the certifi cate print screen (DSA1.02S). The picture becomes even more puzzling
with the weekly certifi cates backup (DSA3.01S) and weekly certifi cates transfer
(DSA3.02 S) screens having the third and fourth highest defect counts. When you
discuss your analysis results with the development team, you discover that they have