Apr 25 2009
Selecting and maintaining the regression test suite.
- 1 Comment
As all of us knows, the purpose of regression testing is to ensure that, by adding new features or modifying existing features to the existing software has not negatively impacted on any functionality that it offered previously.
Selecting regression suite
The issue is how to select effective regression test suite so that the number of tests does not grow too large and minimize the running cost of regression
There are many methods to select regression test suite. One of the effective method which I feel best is
Categorize all functional tests base on functionality. From each category select the minimum number of tests to be rolled into regression suite. Selection should be base on
· All tests which covers multiple functionalities.
· All tests which covers critical functionalities
· All Boundary value test cases
· Some of positive test cases
· Some of negative test cases
Maintaining regression suite
Most Important thing controlling the growth of regression suite (Adding tests into regression suite)
For example: suppose 100 new tests resulted from a new feature testing. Instead of dumping all 100 tests into regression suite, categorize the 100 tests cases based on functionality and select minimum number of tests (i.e. 30% to 40%) into regression suite from each category
While designing a regression test suite, it might help to analyze the recent production/UAT remedies, and why they were missed in QA testing. It may help identify critical areas which should be focused upon.
Also when adding new cases to regression suite, it’s important to remove old cases which are no longer critical/applicable, otherwise the cost of running regression will keep on growing, which is a big concern for some systems.