lakshmi

• Nov 22nd, 2007

 

1.  It is a Good news to Developers, but bad news to a tester, coz a tester is supposed to exploit the application and digout the bugs in it.

2. I think all the 3

3.  Itz in the name itself Inconsistent

4.  First we plan for the Test (Test Plan to be Prepared).  Based on the plan Scenarios to be developed and based on scenarios test cases, then execution and reporting.

5 & 6 same as above

7.  When the number of bugs is high testing is said to be effective.
when your test cases cover up all the functionalities in a given application is said to effective.

8.  Test policy is to find bug

dilton_dalton

 Profile Answers by dilton_dalton 

• Dec 7th, 2007

 

1) As in most cases, it depends.  There are two situations where one would expect all of the tests to pass.  The first is where there are no defects (or very few defects) to find.  The second is when many test cases failed in the first few test passes and all of the detected problems were repaired.

The first case means that the developers did a good job.  The second means that the developers did a poor job and the system probably contains a lot more problems than were found by testing.  most development project fit into the latter category.

2) Automatic testing is necessary in two situations.  The first is in the case were the release being tested is one of N releases of a product.  The automated tests make regression testing of later releases much more economic.  The second case is where the system being developed is expected to contain a lot of defects and the tests have to be run again and again as defects are repaired.  Automatic testing does not reduce resource requirements in the short term because developing the automated tests is labor intensive. 

3) An inconsisent bug is one that doesn't always make itself known, even when the same test case is run.  Timing problems and race conditions are typical problems of this type.

4 through 6 - It depends on what the project is?

7) Test effectiveness is the ratio of the defects found to the defects present in a product.  Since all the defects are never going to be found, it is necessary to make an estimate of the number present at the start of testing.  There are techniques in the literature for making this estimate. 
Efficiency is the ratio of outputs to inputs.  I guess that in the case of testing that would be test cases executed per unit of effort spent on testing.

8) Test policy is just the management statement regarding testing.  A unit testing policy may state that all code must be unit tested using the McCabe criteria and that only code that has been inspected may be submitted for unit test.  A System testing  policy could state that 95% of all functional requirements must be covered by at least one test case and that only code that has been unit tested may be submitted for use in a test build.  Policy should be actionable.