Prev | Current Page 109 | Next

Peter Farrell-Vinay

"Manage Software Testing"

???
???Yes, but we didn??™t have any specifications, and we only had two extra features added yesterday.
We??™ve had no time to write the tests.???
???So? You should have said so earlier.???
(I??™m not exaggerating much. Even if you only heard the
news 10 minutes ago there are those who will say anything, no matter how silly, to deflect blame
from themselves.)
The Big Questions You Need Answers To
21
2.
Yet another example:
it doesn??™t matter how many bugs they fix, the developers always seem to
leave new ones in the code. Management has realized that the situation is getting terminal. Try
to resolve this by:
??? Demonstrating it??™s true (see section 2.7.2 and section 18.10.2.2)
??? Proposing setting up an ad-hoc group to assess the code and unit-test the most-changed parts
in an effort to stem the tide
Avoid the hysteria, risk, and disruption of personal relations by having a release exit criteria defined,
understood, and accepted well in advance. See section 8.5.9
.
Make sure the Directors know this and are
happy with it. This way they have a simple question for you, and you can say ???
Yes, we met the criteria???
or ???
No.???
2.5 What Do We Test It Against?
Here??™s a selection:
1. Our
belief in real-world events.
You can have all the specifications you want and they still won??™t
define every possible event in the real world, like the last example.
2.
Specifications.
What kind of specifications?
a.


Pages:
97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121