A Full Software Development Lifecycle Testing Project - Software Testing: Testing Across the Entire Software Development Life Cycle - page 225

larger software projects completed over several months, it would not be surprising to
experience a 5-10 day correction cycle. The intent should be to have all of the severe
defects corrected and most of the important defects corrected by the Preliminary
construction stage completion date.
CPI previously developed several software applications that are about the same
size and complexity as the DCPS using about the same size project team. As a result,
CPI has defect history that can be used as an expected pattern for the DCPS defect
discoveries. As we saw in Chapter 12, if the DCPS defect log tends to follow the
CPI project defect histories, then we gain a certain level of comfort that the testing
is proceeding well. If, on the contrary, the DCPS defect log diverges from the CPI
project defect histories, then we begin to ask “why?” Figure 13.3 shows the CPI de-
fect histories and the DCPS defect log progress as of Week 5, a little more than half
way through Preliminary construction.

30

25

20

CPI history
DSA project

15

10

5

0

1 4 7 10 13 16 19 22 25 28 31 34 37 40

Project days

Figure 13.3 Comparison of CPI defect history with DSA project defect log

Before raising any issues from the comparison, you fi rst need to recognize that
both plots represent small numbers of defects discovered on any given day. This recog-
nition should temper the issues and conclusions that you derive from the comparison.
The most obvious conclusion you can draw from the defect discovery compari-
son is that DSA testing is fi nding fewer defects and fi nding them later, normally not
a desirable situation. After conducting discussions with the developers, you surmise
that the later defect peak might be attributable to the fact that the print screen was
developed and tested before the data entry screen in the DSA project. It is reasonable
to expect that most defects will be found in the screens with the largest number of
input or output data fi elds; therefore, the delayed defect discovery peak might be at-
tributable to the delayed testing of the data entry screen relative to other CPI project
screen development sequences. Recall that DSA did request a change in the intended
sequence of software development.