??“ Developers would probably appreciate a QuickTime movie of your tester??™s screen travels too.
??? The probability of the existence of more bugs in a section of a program is proportional to the
number of bugs already found in that section [Myers G].
??? Bugs remain in software after testing because both the testers and the developers have the same
view of the way the software will be used.
??? The use of randomly-generated test data reduces the likelihood of shared oversights remaining.
??? Testing carried out by selected
test cases
, no matter how carefully and well-planned can provide
nothing but anecdotes [Mills 87].
??? Mathematical modeling will not reveal discrepancies between the model being used and the real
world. Only testing will do that [Parnas].
??? 20% of the units contain 80% of the bugs [Endres, Boehm 75].
??? In testing nothing succeeds like failure. If you??™re trying to convince people of the need to test
longer, harder, or with more tool support, the evidence of failure in terms of enraged, fee-paying
users is the best support you can get. But you??™ll get it too late.
??? Software exhibits weak-link behavior: failures in even unimportant parts of the code can have
important repercussions elsewhere.
??? You never know the environment in which software will have to work throughout its life [Voas].
??? The existence of a tool changes the nature of the problem.
3
Therefore systems as yet unbuilt will
be used in ways we cannot imagine.
Pages:
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159