Tuesday 5 February 2013

Regression Test Trap #492

I didn't call this "the" trap, as I think there are several. This is one aspect.

In June last year I had a little twitter rant/exploration (or transpective investigation, [1]) about regression testing. It was triggered when I saw a LinkedIn skill labelled as "Regression Testing". I wanted to know which skills that people considered specific to regression testing.

Now, one can take some of the LinkedIn labels with a pinch of salt, but I wanted to test the water about this one.
Note, I really enjoy these "bursts of activity" I have on twitter. There are many great minds that immediately question, probe, ask for explanation or example. I love it and am very appreciative to anyone that asks me to explain myself. It makes me stronger and clarifies my thoughts!
The twitter discussion dump, via twimemachine, in reverse order:
@Arborosa you have that problem in non-regression-specific testing also Thu Jun 28 06:54:20 +0000 2012@jamesmarcusbach Oh, the NAP is a great resource - was looking at this the other day http://t.co/Tjn3b18L Wed Jun 27 10:00:53 +0000 2012@jamesmarcusbach That's great, thanks - I need to spend more time on it - reminds me of Göranzon's work-> must dig that out again Wed Jun 27 09:49:44 +0000 2012@jamesmarcusbach True, it's one of my aids to avoid the trap of going from "application in context" to "application in general" Wed Jun 27 09:42:56 +0000 2012@jamesmarcusbach I try to distinguish between skills and application of skills, reminds me of "Dialogue Skills and Tacit Knowledge" Wed Jun 27 09:25:44 +0000 2012@jamesmarcusbach I call the workshops a feature walkthrough, see slides 20-27 in http://t.co/XnF5Ubnk - usual health warning with ppt slides Wed Jun 27 08:50:05 +0000 2012@jamesmarcusbach In my context, these skills are not specific to regression testing Wed Jun 27 08:37:44 +0000 2012@jamesmarcusbach Test idea maintenance - not specific to regression testing 4me - can be triggered by feature or test harness change/growth Wed Jun 27 08:36:42 +0000 2012@jamesmarcusbach Interviewing and analysis - I do this as part of feature interaction analysis - sometimes 1-2-1, sometimes as a workshop Wed Jun 27 08:36:00 +0000 2012@DuncNisbet yes, start from a picture of what you want to learn + where risks are - then let that point to what's "interesting" to know Tue Jun 26 22:08:01 +0000 2012@DuncNisbet I think of tests with a potential to find a problem as "interesting" - ideas to pursue/try out Tue Jun 26 22:00:29 +0000 2012@rahul_verma me too - I thought your questions were good and relevant! Tue Jun 26 19:10:03 +0000 2012@rahul_verma look forward to it Tue Jun 26 19:02:17 +0000 2012@rahul_verma yes, that's what I aim for Tue Jun 26 19:01:48 +0000 2012@rahul_verma and how many other types and application of testing were they no good at? :) Tue Jun 26 19:00:47 +0000 2012@rahul_verma I think people mix up activities, skills, application and results - and makes our job (communicating about a product) harder :( Tue Jun 26 18:57:20 +0000 2012@rahul_verma applying a skill successfully in an activity (RT) does not make that a skill specific to that activity Tue Jun 26 18:56:15 +0000 2012RT @rvansteenbergen: Want tweets about the latest news from testblogs? Follow @testingref #softwaretesting Tue Jun 26 18:45:00 +0000 2012@jamesmarcusbach I'd love to hear about the other things Tue Jun 26 18:14:31 +0000 2012@rahul_verma agree! Tue Jun 26 18:11:32 +0000 2012@rahul_verma on a certain level performance testing and regression testing do have an overlap of ambition Tue Jun 26 18:10:10 +0000 2012@rahul_verma performance testers focus on an element of a product's characteristics - to measure, monitor, feedback, recommend and help tune Tue Jun 26 18:09:02 +0000 2012@rahul_verma exact knowledge is tricky to define.... Tue Jun 26 18:03:28 +0000 2012@rahul_verma I'd say the purpose distinguishes it as regression testing, but not necessarily the skills used Tue Jun 26 18:02:42 +0000 2012@rahul_verma Are you asserting that regression testing is done by regression testers? Tue Jun 26 13:05:39 +0000 2012@rahul_verma What skillset in "regression testing" is specific to regression testing? Tue Jun 26 13:03:13 +0000 2012@rahul_verma Think skillset rather than domain knowledge Tue Jun 26 13:02:15 +0000 2012@rahul_verma Is that separate from good testing? Tue Jun 26 12:55:56 +0000 2012@thetestmile And is that separate from good testing? Tue Jun 26 12:55:34 +0000 2012@rahul_verma Are you saying you can be a good tester and not a good regression tester, or vice versa? Tue Jun 26 12:55:08 +0000 2012Now where's the "black viper testing" skillset button.... Tue Jun 26 12:42:55 +0000 2012"Regression testing" is a "skill" in linkedin. Really? Discuss... #faketestingmeasurement Tue Jun 26 12:41:39 +0000 2012
Trap
The trap is to think of the skills associated with good regression testing as being separate from good testing.
In my context and experience, I consider good testing to take into account aspects of new features that interact with legacy (existing) features or systems. This implies that "regression testing" is included. But if I think of specific skills needed for regression testing then these are just as applicable to "non-regression testing":
  • Investigative skills about the domain (system)
  • Discussing development and the existing system with developers/BA's/stakeholders
  • Test Idea / Environment Analysis & Maintenance
  • Reporting and discussing the status of the system
  • etc, etc

So, is a "regression testing" skill separate or unique to regression testing? I assert not, and that it can be just as common in "non-regression testing" or any good testing.


To Be Continued
I have another trap about regression testing - coming soon.

References
[1] Satisfice: Transpective Dialogues for Learning http://www.satisfice.com/blog/archives/62

Other "traps":