The online community for software testing & quality assurance professionals
 
 
Calendar   Today's Topics
Sponsors:




Lost Password?

Home
BetaSoft
Blogs
Jobs
Training
News
Links
Downloads



Quality Engineering >> Quality Methodologies

Pages: 1
koranke
Junior Member


Reged: 12/18/04
Posts: 162
Loc: Seattle, WA
Test Change Management
      #717244 - 10/01/12 10:53 AM

Not sure what this would officially be called or if this is the right place for posting this. I'm wondering if anyone has a process in place that allows them to evaluate what regression tests need to be run when a change is introduced? For example, if a database table were changed, ideally all tests that touch on that table would be flagged as potentially needing to be rerun. Or if a code file, class or method were changed, likewise ideally all tests that touch on that code would be flagged as potentially needing to be rerun. I've not seen or heard of anyone actually having a system in place that allows for that. Perhaps because it would be too cumbersome to track all the potential relations. I've worked with a number of different test case managment tools including Quality Center and more recently TFS, but neither tool seems to have any built in support for this kind of thing.

Post Extras: Print Post   Remind Me!   Notify Moderator  
michaeljfModerator
Veteran


Reged: 09/17/01
Posts: 3979
Loc: Yankee Land
Re: Test Change Management [Re: koranke]
      #717333 - 10/02/12 08:44 AM

I still consider this Regression Testing, not sure there is a process about it, but to determine what Regression Tests need to be run you HAVE to understand the changes. Choosing Regression Tests to me is more art than science, sure you COULD run everything if you have time, but more often than not you need to be selective. For that you really need to understand the changes, talk to the Dev and anyone associated with the pieces undergoing change, then decide what Tests will give you coverage.

--------------------
- M

Nothing learns better than experience.

"So as I struggle with this issue I am confronted with the reality that noting is perfect."
- Unknown

Now wasting blog space at QAForums Blogs - The Lookout


Post Extras: Print Post   Remind Me!   Notify Moderator  
koranke
Junior Member


Reged: 12/18/04
Posts: 162
Loc: Seattle, WA
Re: Test Change Management [Re: michaeljf]
      #717338 - 10/02/12 09:11 AM

It's output is definitely regression tests, but the point of the process is to try to make regression testing more science and less art. As you said, you need to be selective, but what should you select? I worked at a place that for many years simply regression tested everything. Near the end of my time with them they finally switched to a more focused approach. Basically, we'd look at the changes made, talk to Dev and see what they thought was at risk or impacted by the changes and regression test that. However, Dev does not have perfect knowledge of dependencies and relationships. Maybe it's the best we can do, but I was reading a book (forget the author and title) where he was talking about regression testing based on evaluation of code file changes. That implies some link between code files and test cases, though he didn't go into any explanation on that.

Post Extras: Print Post   Remind Me!   Notify Moderator  
michaeljfModerator
Veteran


Reged: 09/17/01
Posts: 3979
Loc: Yankee Land
Re: Test Change Management [Re: koranke]
      #717410 - 10/03/12 07:28 AM

It's understanding the changes, much of it familiarity with the system and how it interacts. Basically QA is going to have to use their judgement, and knowledge, about the SUT to make a decision as to what Tests need to be run. You can simplify this by documenting certain systems, so you know what needs to be run, within your tests - or if you use a runner like NUnit add Tags to specific tests so if you have an area that needs testing you pass that value in when runinng the suite.

I've read similar books and articles but it comes down to knowing what Tests will Test certain functionality, or Check it depending on what term you like to use. If you don't have some kind of matrix, or way, to link a test to a function or app or requirement it'll definitely be more art than science. I got around a lot of this by continual revision and review of the cases so I knew what tested what and then if Code in X changed I knew to run the cases that were linked to X.

--------------------
- M

Nothing learns better than experience.

"So as I struggle with this issue I am confronted with the reality that noting is perfect."
- Unknown

Now wasting blog space at QAForums Blogs - The Lookout


Post Extras: Print Post   Remind Me!   Notify Moderator  
Pages: 1



Extra information
0 registered and 3 anonymous users are browsing this forum.

Moderator:  blueinatl, AJ, michaeljf, swt88 

Print Topic

Forum Permissions
      You cannot start new topics
      You cannot reply to topics
      HTML is disabled
      UBBCode is enabled

Rating:
Topic views: 2755

Rate this topic

Jump to

Contact Us | Privacy statement SQAForums

Powered by UBB.threads™ 6.5.5