Lord Lotus Blog | Test Case review

Wednesday 25th, August 2004

Test Case review08/25/2004

Test Cases

The testing team I'm working with reviewed the 28 test cases I'd written as of 8/20/04. There were 44 sub-cases in all. The review itself went very well. Most of the changes requested were formatting/wording changes that I'd need to make to each test case/sub-case. I'll need to make all of the changes over the next several days. No rest for the wicked, though, as I need to write another three test cases today.

Comments :

1. Kiatchai05/02/2007 03:45:23 AM

for testing

2. Richard V08/07/2007 05:47:02 PM

I've written and executed tests for different companies for several years, but have never done a formal test review. What did it involve for your team? Did all of the reviewers study every single test case? Did someone read through all of the test cases in a meeting for a group of reviewers? Do you know of resources for helping me to plan and execute a test review? My company is asking for test case review, and I'm anticipating glazed eyes and a sense of ad nauseum amongst the reviewers for all of the detail in the test document. Thanks for listening!

3. Michael Sobczak08/23/2007 08:05:33 PM
Homepage: http://www.punkdbynotes.com

The project I described here and in other posts during this time-frame involved an identity management web service EDS developed for GM. Since the GM client required that the code be tested independently of the development team, EDS' Global Testing Organization was engaged to handle all aspects of testing, including creating the test plan, writing the test cases, executing the test cases and recording the results.

The team consisted of a testing administrator and two test case authors. The testing administrator served as the project leader for the testing requirements of the project. My role on the project was as a test case author.

Before we wrote the test cases, I created a test plan. This document described who would be responsible for writing the test cases, the testing environment, how issues would be logged, how the issues and overall testing status would be communicated to the interested parties, etc. We reviewed the test plan first as a team, then held a formal review with other members of GTO not directly involved with our project. This ensured that we had independent views on what we were planning to accomplish with our testing. After the technical project lead signed off on the test plan, the other test case author and I mapped out all of the test cases and estimated the amount of time it would take to create them. After our estimate was approved, we wrote all of the test cases over a period lasting several weeks. I think we wrote with over a hundred individual test cases.

Our test cases closely followed general Use Case structure. We described in detail each step the actor would take, as well as the expected outcome of that step. Fortunately, the code we were testing was a web service. We didn't have any User Interface to worry about. We did have to build our input XML streams from scratch, though. I used IBM WebSphere Studio's XML editor, which made creating the XML documents somewhat easier in that I was sure that the resulting XML document was valid according to the schema provided.

After the other test author and I finished writing all of the test cases, the three of us reviewed every test case off-line and discussed them in a conference call. When we were done with our review and revising our test cases, we sent the test cases to the lead developer for his review. After he had time to review them, we discussed his findings via a conference call. Finally, we sent the test cases to the project leader and the entire development team for their review. After about a week, we reviewed their findings in a formal review meeting conference call. (Most of the developers on the project worked at different EDS facilities, including several who were based out-of-state, so we couldn't get together in a single conference room.)

In each of our "formal" review meetings, we never went through each test case line-by-line. The understanding was that all reviewers would take the time to review the test cases and record their comments off-line. Then, when the formal review meeting was held, the moderator of the meeting would announce each test case by name and wait to see if there are any comments from any of the reviewers. If there were comments, those comments would be discussed and recorded on a meeting log form. After the meeting was over, the test administrator and the test case authors reviewed the comments and updated the test cases accordingly. We never scheduled any formal follow-up meetings that I can remember. I believe we just e-mailed the revised test cases to the reviewers in case they wanted to review them again for accuracy.

The business and technical aspects of the project followed GM SDP methodology, which I believe has changed over the past several years to something more CMM-oriented. EDS GTO followed their own procedures.
Since all of the processes and materials are probably copywrite-protected, they can't be provided to someone outside of GM and EDS without those parties agreeing to wave the copywrite-protection.

Hope this helps!