Prev | Current Page 183 | Next

Peter Farrell-Vinay

"Manage Software Testing"

Non-safety-critical projects need to ensure it happens because:
??“ Developers (and their project manager) may want to get releases out fast, don??™t want to hear
how many bugs there are, and don??™t face the users or the shareholders.
??“ By the time that senior management realizes there??™s a problem it will be too late to save the
project, the product or the company.
??? Release management: The test manager or his boss should be able to prevent a release to the field.
If they can??™t, then they can bear no responsibility for any bugs found later. Essentially testing is a
way of answering the question ???is the system ready to be released.??? If the answer is ???no,??? then
ignoring that answer is another way of saying ???we don??™t need to test.??? At this point the wise test
manager polishes his CV and sends it off. Since the project manager is ultimately in charge of
quality, a release must ultimately be his responsibility.
??? Why might the answer be ???no????
??“ Because you still have some severity-1 or -2 bugs in the system
??“ Because the testers tell you it??™s not ready.
??“ Because you haven??™t exercised enough of the code.
??“ Because there should be x bugs, and you haven??™t found them yet. See Chapter 18.
??“ Because code turmoil is still too high (see section 18.9.6).
??“ Because the bug-detection rate is still too high. Look at the curve in the Figure 4.11.
The curve shown in Figure 4.11 is of a test cycle which is finding a growing number of bugs.


Pages:
171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195