Discuss Agilepublic

Community Agile Discussions

A variety of ongoing discussions about Agile development

This is a public Discussion Area  publicRSS

Thread

    Tracking your Progress when running Test Cases
    Thread posted 4/23/09 by Neil Addison
    18058 Views, 5 Comments
    Title:
    Tracking your Progress when running Test Cases
    Content:

    We are rapidly spinning up our manual test case execution using RQM.  Formerly, we used QC's Test Director which tracks test progress at the Test Step level.

    We really like the way that Rally aggregates to the TEST CASE level, but feel that a user should have at least a Check Box at the Test STEP level, especially for those cases that might have 20+ steps.

    Does anyone else think this would be useful?

    ~Neil~

    CSM - Brinks Inc.

    Comments

    • posted 6/19/09 by pete skrebetz

      Neil,

         We ran into the same issue when some of our team that used QC's Test Director started using Rally.   We came to an agreement that any Steps that didn't meet the Expected Outcome should cause the Test Case to Fail.   The tester then needed to manually document the step it failed on in the Results Notes.  

          Another option you might consider is the integration that Rally has with QC's Test Director.

      =pete

      Reply to this Comment

      Attach File

      • posted 7/20/09 by Neil Addison

        Pete, sorry I'm so late responding.  I followed your advise here and it's work beautifully.  We decided to make the cases much more succinct going forward. Your post / response was very helpful. THANKS a MILLION !

        Neil

        Brinks, Inc.

        Reply to this Comment

        Attach File

    • posted 7/17/09 by John Overbaugh

      Neil, another thing to think about is a serious metaphor change. Test Director is really big into what I call the monolithic approach to testing... Long test cases which are more like scenarios than cases. In my worldview, test cases should be relatively atomic and have only one validation point. Rather than having a case with steps like this:

      1. Open browser, navigate to www.facebook.com
      2. Log in using valid credentials
      3. Navigate to inbox
      4. Create a new message
      5. Address it to somoene in the address book
      6. WRite a long message
      7. Click send
      8. Log out
      9. Validate that you are logged out

      I would have separate test cases for the login functionality (including negative cases), the inbox's new message and address book functionality, and for log out. It's much more reflective of the overall functionality of the project, as well as the project's quality. If the above scenario failed because log out failed, management would think the entire app is broken when, in fact, 2/3 of the features under test are working just fine.

      Reply to this Comment

      Attach File

      • posted 7/20/09 by Neil Addison

        John,  I think you've got the right approach.  As I look back at what was created in QC, many of these test cases read like War and Peace.  Atomic is the way to go.  Thanks for the good feedback. 

        Neil

        Brinks, Inc.

        Reply to this Comment

        Attach File

    • posted 7/21/09 by Sudheer K

      I agree with this approach, I also had same problem, for instnance I copy existing test case and try to update. It is being hard to delete test cases one by one. If we have check box next to the each test case that really helps, and we need separate delete button though.

      Reply to this Comment

      Attach File