Prev | Current Page 113 | Next

Peter Farrell-Vinay

"Manage Software Testing"


There??™s no user interface. There??™s a system A, which processes something from
system B and generates something to be used by system C. Only the data is real. You have to test
it. Naturally there??™s no specification for the thing in between. Write an interface specification as
discussed in Chapter 8.
Look at server log entries to find what real traffic has occurred.
6.
Gossip and hearsay.
Don??™t ignore it. Developers can be very touchy about being refused permission
by a pressuring project manager to fix unstable and otherwise risky bits of code. If this
happens they??™ll often love to tell you where the really risky bits are. Similarly a hard-to-repeat
bug can be discussed with a sympathetic developer who can offer insights on why it might have
appeared.
2.6 What Do We Test it With?
1. Tools which simulate inputs.
2. Real-world sample and faked data.
3. Tool harnesses which simulate an environment.
4. Tools which manage tests.
5. Test scripts which simulate users.
See Chapter 7 for more on tools and data.
2.7 Are We Getting Better or Worse at Finding Bugs?
The only absolute answer to this question is to identify the bugs reported from the field as a proportion
of those found in-house for any release. This measure, however:
The Big Questions You Need Answers To
23
1. Ignores those bugs users cannot be bothered to report, perhaps because they are so bad that they
are about to abandon use of the system or product.


Pages:
101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125