Developer-tester, Tester-developer

During this week I watched the following conversion between Robert C. Martin and Michael Bolton on Twitter:

Uncle Bob
@dwhelan: If you’ve enough testers you can afford to automate the functional tests. If you don’t have enough, you can’t afford not to.

Michael Bolton
Actually, it’s “if you have enough programmers you can afford to automate functional tests.” Why should /testers/ do THAT?

Uncle Bob
because testers want to be test writers, not test executers.

Michael Bolton
Testers don’t mind being test executors when it’s not boring, worthless work that machines should do. BUT testers get frustrated when they’re blocked because the some of the programmer’s critical thinking work was left undone.

Uncle Bob
if programmers did all the critical thinking, no testers would be required. testers should specifying at the front and exploring all through; not repeating execution at the end.

Michael Bolton
I’m not suggesting that programmers do all the critical thinking, since programmers don’t all of the project work. I am suggesting, however, that programmers could do more critical thinking about their own work (same for all of us). Testers can help with specification, but I think specification needs to come from a) those who want and b) those who build.

Over the weekend I thought through the reasoning. First of all I truly believe that both are right. Elisabeth Hendrickson stated this in the following way:

In any argument between two clueful people about The Right Way, I usually find that both are right.

Truly, both Uncle Bob and Michael Bob are two clueful people. In the conversation above they seem to be discussing about The Right Way, and I believe they are both right – to some degree, in some context, depending on the context. Basically this is how I got introduced to the context-driven school of testing.

The clueful thing I realized this morning was my personal struggle. Some months ago I struggled whether I am a tester-developer or a developer-tester. Again raising this point in my head with the conversation between a respected developer and a respected tester opened my eyes. Three years ago I started as a software tester at my current company. Fresh from university I was introduced into the testing department starting with developing automated tests based on shell-scripts. Finally I mastered this piece and got appointed to a leadership position about one and a half years later. Until that point I thought testing was mostly about stressing out the product using some automated scripts. Then I crossed “Lessons Learned in Software Testing” and got taught a complete new way to view software testing.

The discussion among these two experts in their field raised the point of my personal struggle. Testing is more than just writing automated scripts. Over and over executing the same tests again is a job a student can do. It’s not very thought-provoking and it gets boring. Honestly, I haven’t received a diploma in computer science (with a major in robotics) to stop thinking at my job. Therefore I got into development topics. Since our product doesn’t include good ways for exploratory testing, I came up with better test automation. Basically the tools I invented for the automated tests also aid in my quest to be good as an exploratory tester.

So what was I struggling with? Basically I realized that on my job the programmers get all the Kudos. That’s why I started investing time in becoming a better programmer. Meanwhile I found out that it’s also a fun thing to do. You can see the results whenever you run your programs. This is why I never gave up looking at code, dealing with it. Sure, it’s not what Michael Bolton or Jerry Weinberg mean with software testing. On the other hand it’s what I would like to do.

Basically I consider myself a developer for software test autmation. I have a background in software development and I have an understanding of software testing. (I leave it up to my clients and superiors whether I’m a good one or not.) In the way I understand my profession it’s necessary for me to know about both sides. This is also why I now realized that I am a tester-developer, not a developer-tester. As I just realized my personal struggle comes from the aspect that I am not sure, whether I really ever was a software tester or not. But for sure I have started to become a developer.

In order to conclude this posting with the initial discussion from the two experts in their fields, there is one thing left to say. As a software tester you may choose to become a developer of test automation. Robert Martin refers to these kinds of testers. On the other hand as a software tester you may choose to become a tester who applies critical thining. Michael Bolton refers to these kinds of testers. Whether or not to choose one path over the other may be up to you.

  • Print
  • Digg
  • StumbleUpon
  • del.icio.us
  • Facebook
  • Twitter
  • LinkedIn
  • Google Bookmarks

Leave a Reply

Your email address will not be published. Required fields are marked *