Information Technology Reference
In-Depth Information
Time
Trend-lines for bugs
Figure 12.6
Frequency of bugs.
The first round of QA typically generates a lot of bugs, some “real”
and critical, and others that are not so serious. The discovery of bugs
follows an upward trend initially, and then descends. This does mean that
the software is ready to launch after the initial batch of bug detection.
Figure 12.6 shows how the process generally works.
QA teams discover the most obvious problems in the beginning (first
test cycle). These are typically related to form rather than content: empty
fields, mandatory and nonmandatory checks, user-level validations, bound-
ary conditions, improper messages, blank reports, etc. Once these bugs
are fixed, QA starts to focus on the logical aspects of the product, using
more detailed test cases with proper test data. During this phase of QA,
one is testing not just the stability of the code, but also the behavior
expected from the system. It is here that the next increase in bug detection
occurs.
Most bugs require the attention of the original engineering team. At
the time of the first code drop to QA, the Engineering manager could be
under pressure to release some of the developers because they are done
with their main task; the code has moved on to QA. However, to reallocate
the engineering team at this time is premature. A substitute team may
unintentionally add problems because they will not understand the entire
code and its dependencies across the product. The original developers
should be retained until one complete cycle of QA has been completed.
QA teams use a test tracking tool that records tests and classifies bugs.
These tools allow development to focus on bugs classified as “High
Priority.” Even with proper guidelines, this classification may not be very
reliable because there is some subjectivity in determining the seriousness
