Information Technology Reference
In-Depth Information
What should you test?
Critical to success
Functionally complex
Core components
Visible to customer
Has financial impact
Late additions
Figure 12.2
Prioritizing testing efforts
.
Given that we cannot test everything, we have to use a combination
of risk analysis, prior industry benchmarks, and one's own experience
in determining what and how much testing would be adequate (Figure
12.2).
There are certain areas of the application and development scenarios
that stand out as candidates for increased testing. More testing resources
should be devoted to:
Features considered critical by the customer
. Every system or appli-
cation has one or more critical functionalities. If these do not work,
then it matters little if the others do — the system is a failure. The
decision about what is critical must be made by the customer, not
the developer or by the analyst.
. There is an increased likelihood of
bugs being present in applications or modules that are complex.
Functionally complex modules
Core components
. Components that are heavily used or reused
throughout the application are crucial. A bug here has widespread
effects if it were to fail.
Features that are most visible to the customer
(even if they appear
to be of cosmetic or of mere “nuisance” value). Users' first impres-
sions are often difficult to change. These impressions are generally
the result of interactions with, maybe, some trivial aspects of the
application. Users assign a lot of importance to the user interfaces
of the applications with which they interact. While it is understand-
able that one would like to focus on critical and severe bugs, it
may not be advisable to prioritize these over those that are visible.
Any feature likely to have a
financial impact
, however small.
