Have been stuggling with this for some time, so i thought, let's see what i can get from the 'community'
Issue is as follows, i am setting up tests for a highly batch oriented application under test, and am having difficulties how to set this up logically.
Assume the following
10 testcases with each 10 teststeps
Each of these testcases consist of manual user actions and batch processing jobs.
the manual user actions are unique to the testcase but the batch processing jobs are generic for all testcases, although not always applicable to each testcase
Taking a business process approach for each testcase (A-Z execution per testcase), results in a 'stand-alone' testcase, but as well in running duplicates for the batch processing jobs, hence unefficient
Doing testcases in parralel, e.g. do manual step 1 for all testcases, run batchprocess, manual step 2 for all testcases, results in an efficient batch process, but in an difficult to organise testset.
2 testcases, Hire2Retire process
Testcase 1 A new employee is hired
testcase 2 An employee gets a raise
Both testcases require several manual steps, however, you only want to run the payroll process once. and validatie the individual results afterwards
Interested in other points of view on this.
(obviously above is a simplified form, if there were only 10 testcases in total, i would be fine one way or another, thruth is there are just a few more....)
The pessimistic part of me says that actually I want to do both - test each end to end individually to make sure they work in isolation and then do a whole heap together to make sure they can run together. I appreciate this makes a lot more work.