It's all about the answers!

Ask a question

RQM: Test case and multiple test scripts

Jörg Werner (3033579) | asked Jan 24 '14, 5:39 a.m.
We have setup test cases with multiple test scripts. One test script covers a positive test and one a negative test.
We use no test environment and have create just a dummy one "None", so that we are able to create TER within a test plan.

Currently only one TER with one test script (the positive one ) is created.
There is nothing which indicates that also the negative test should be performed.

Bug or Feature?

It seems that we can solve this only
a) create for each test script a test case.
b) Merge the test scripts to one big (and ugly?) test script?

In case that the answer is "Feature", is it possible to prevent, that a user creates multiple test scripts for a test case?

  Thanks, Joerg

Accepted answer

permanent link
Piotr Aniola (3.7k11638) | answered Jan 24 '14, 5:45 a.m.
There should be only be one test script per test case. If there are multiple, these should perform the same test (but perhaps on different platforms, etc).

Jörg Werner selected this answer as the correct answer

Georg Kellner commented Jan 24 '14, 7:06 a.m. | edited Jan 24 '14, 7:11 a.m.

Does this make sense?
Easy example: A train is allowed to pass a yellow signal with 60 km/h.
I would create a testcase:
"Check speed for yellow signal"
with testscript "Valid speed" and testscript "Invalid speed"

But this isn't usable with RQM.
So I have the possibility to use a script "check speed" with one testcase, but it doesn't make sense in executing, because you don't have two yellow signals one after another.
Using two testcases, describing exactly the same test execution with a small change of one parameter is definitively overkill.

greetings georg.

Piotr Aniola commented Jan 24 '14, 7:12 a.m.

My personal opinion on this is that (regardless of RQM architecture) there should be two test cases - what happens if the train is driving by the yellow light at 60km/h and what happens at 61km/h

Simon Washbrook commented Jan 24 '14, 7:22 a.m.

As stated in the above blog, Test Cases are tied to requirements, not Test Scripts. That is why best practices are to use one test script per test case. Then you can see which requirements have been covered by test cases.
In the above example of the train, you would have 2 requirements, for example
1/ When a train goes through a yellow signal, if the speed is less than or equal to 60 Km/H then no alarm will be raised
2/ When a train goes through a yellow signal, if the speed is greater than 60 Km/H then an alarm will sound in the cabin.
You would then map these requirements to Test Cases and within each Test Case you would have a script to test that the requirements is met.

David Mehaffy commented Jan 27 '14, 2:57 p.m.

I disagree with one test script per test case and we have many examples where this is not true.  For example you have a generic test case to test a function but the procedure to run the test is slightly different for say running on Linux versus Windows.  You don't want to have to create 2 test cases because what you are testing is the same and this could cause an explosion of test cases - which would basically have the same design and other information the only thing that would differ would be how to do it.  Also if you had a manual test that you convert to automated - that would have two test scripts associated as well - so the suggestion of only one test script per test case is invalid.  

I have a couple of enhancements opened on this problem and we would like to see this fixed as well. 

Piotr Aniola commented Jan 28 '14, 5:03 a.m. | edited Jan 28 '14, 5:03 a.m.

Actually, the Linuxx vs Windows case is covered by RQM, and it's perfectly valid usage to have two scripts per test case. In fact, I believe this is the very reason adding more than one script to a test case is possible to begin with. This case is actually mentioned in the article:

Your answer

Register or to post your answer.