Generally by sprint 2 you would have your regression test suite from sprint 1 automated; then in sprint 3 your sprint 2 regression test cases automated, and so on... allowing them to be re-run in each new sprint, and allowing your manual testers to focus on the new/changed functionality. If you are doing all your testing manually, I feel you may struggle to keep up in later sprints.
If you are new to Agile (which I am assuming you are based on some of your recent posts), may I suggest an excellent book written by Lisa Crispin and Janet Gregory: 'Agile Testing: A Practical Guide for Testers and Agile Teams'
Enlighten me please. [img]/images/graemlins/smile.gif[/img]
Regression testing is regression testing regardless of any development methodology - correct? I suspect that Agile makes no difference - correct? Regression testing was alive and well long before Spiral, RUP, Agile, etc.
To me, if there is a new build of software, one must consider regression testing.
Ramesh, what I have seen done with (mostly unit but can be system) testing in Scrum is:
* Stores are turned into tests (TDD)
* These tests are automated and included into a big test harness
* code is written to pass these tests
* Some sort of Constant Integration is run daily and kicks off the automation harness once the build compiles to test it.
So as the test harness is growing every iteration, the regression is being performed on a daily basis by the CI environment.
This is of course an imperfect science. Not everything can be automated and some manual regression is still on the cards. However if it cuts the manual workload say 75% its a massive bonus on short iterations
Dont go from Waterhocfall to Scrumhoc [img]/images/graemlins/smile.gif[/img]
QA Manager, ISTQB, ISEB cert.
I strongly support psychometric testing QA staff