“… we don’t need no thought control.” Today I heard this song during a sports course from CD, and was reminded about a piece I read last week, and my reaction on it. Before I start let me mention that I don’t intend to offend you if you have put work into the ISTQB (or any other test certification) syllabus, or if you are teaching it. I truly believe that your intentions were quite reasonable at the time when you started to work on it. That said, if your kids are watching over your should right now, send them to bed, or playing while you read the following.
The background on this blog entry is the eBook from the Software Testing Club that was published around Christmas and is called A tester is for life, not just for christmas. There are several answers from testers all around the globe to a set of questions. One of these questions deals with Testing Certification Schemes. Here is the answer I gave at the time:
Certification is the devil in disguise. Just like the medieval triumvirate of army, church, and government formed an evil axis, Certification-Institutes, certification course sellers, and HR-departments create machinery which blinds executives from proper action, meaningful leadership, and advancement in the craft.
I fully support this statement, even several months after answering the question.
Now, during the past week, Simon Morley put up another issue of the Carnival of Testers, where he referenced a blog series by Dorothy Graham on the origin of the ISTQB certification, where she cites parts of my answer, and motivates her work for the ISTQB syllabus. When I read her three part series, I noticed that my statements surely need some more explanations. And I would also like to add some pieces regarding her reaction.
The devil in disguise
The origin of this phrase for me was a collection of audio CDs from the German Heavy Metal band Running Wild. The albums “Pile of Skulls”, “Black Hand Inn”, and “Masquerade” deal with the medieval world where the church, kings and queens, and the military formed a triumvirate. This triumvirate caused lots of victims for the notion of power.
The three albums end in “The Masquerade” with the following dialogue between a “Redcoat” and three masqueraded persons:
Redcoat: Are you willing to raise our empire of hate and evil onto the whole world and to deliver man and every living creature onto our hands?
The Three: Yes, master!
Redcoat: Are you willing to tell man lies about world’s past, how he was made and to suppress the truth for all times?
The Three: Yes, master!
Redcoat: Are you willing to deliver your souls onto our hands to become powerful, famous, and rich?
The Three: Yes, master!
Redcoat: Ho Ho Ho! So it shall be done!
It might seem a bit unfair to say something about medieval popes, kings, and queens. Similarly it might appear to be unfair to say the same about the HR departments, certification institutes, and course sellers. On the other hand, that last question raises thoughts, maybe.
On the other hand, if you take a look onto the F.A.Q. of the ISTQB, it states that the ISTQB is a non-profit organization. They are just defining guidelines for training courses. But as someone outside the ISTQB I wonder where do they get funding from, and who takes care about the non-profit part of their organization. I can’t see where they spend money on, where money is going into that organization, and whether they don’t put something form the money of the training course aside. For me, having dealt more than 10 years in a non-profit sports club, this looks suspicious, but maybe their fundings are just open for members. Maybe.
This now leads to the points that Dorothy Graham raised as a historical motivation behind the course syllabus. From the ebook there were also arguments for certification. Dorothy states the following:
- certifications are a good starting point
- certifications provide knowledge of basic terminology
- certifications help get a first job
- certifications demonstrate that you are serious enough about testing to take an exam
Let’s have a look at each of these in turn.
Certifications are a good starting point
Certifications are built on a syllabus. They provide testers with a body of knowledge that is mandatory to all classes taught under this label. So, they may get you started. Sure.
But then I consider my course in software testing. Back, when I got out of university, I found my first job as a software tester. So far I had not heard about software testing, was rather oblivious to it. There was a course in university where this word was mentioned, specifically during the Software Engineering class when mentioning the V-Model, once, in one lesson. Even if there had been a course, thinking back to my university days, I wouldn’t have attended it for I found it to be boring.
Still, I started in April, at the end of April I had learned enough about the particular testing problem at the company to do some adaptations to the course how they tested software at that point. At the end of the second or third month, I remember that I sat in a status meeting, telling my colleagues that I finished my piece of work, when they gave me another topic to look into. This continued until I found myself a software testing group leader one and a half years later. Up until then I never attended any course on software testing, and still was productive – I think.
To date, I do not have any testing certification, and I am proud of this.
Sure, certifications can be a good starting point, but the ones that you would like to be working for you are surely not reliant on it.
Certifications provide knowledge of basic terminology
Certifications are built on a body of knowledge, so that everyone getting out of the class know what their colleagues are talking about, and vice versa. I love this scam the most from the ISTQB pros.
Recently I held a course at a client. The course was filled with about 13 of their testers. The testers came from two different departments. As part of my course I introduced the testing quadrants dimensions – business-facing vs. technology-facing and team support vs. product critique. Since they asked for a common language, I asked the participants to brainstorm together testing methods and techniques which they already did onto cards. Then I asked them to pin them to the board where I had drawn the quadrants on.
Both departments had sent their testers to training courses, and even all of them were certified under ISTQB. Still, they had cards pinned down to the quadrants named “Integrationtest” in two different places. When I asked about this ambiguity, they answered that the more technology-facing integration test probably is the integration test of some code modules, while the more business-facing is probably the one for several executables talking to each other. Hooray for the common knowledge.
Now, Integrationtest is not the only ambiguous term I found while doing classes. Others include “Regressiontest”, “Explorative Test”, and of course “functional test”. So, while the certification schemes put a lot of effort into creating a common term for things testers do, “Integration test” is still “integration test” to some person (at some time). Putting this in another way: The Oxford English Dictionary also collects common terms for English words. Still there are eight different definitions of the word “standard” as Michael Bolton once taught me. Having common used terms does not make them unambiguous in itself.
While reading up Dorothy’s write-up it occurred to me that this was one of the historical reasons to create this certification at all. Looking back on the course from last week, they didn’t seem to have accomplished it. Keeping in mind the Relative Rule (X is X to some person), it occurs to me that the goal of creating a common language is the wrong goal in first place. What we should be seeking is a common understanding. Language may help with this, but language alone is not enough. For something as complex as software testing we have to talk and discuss about what we do, in order to reach that common understanding. Words alone are not sufficient at all.
Certifications help get a first job
I strongly offend this position. Certfication may appear to have helped with it, but it could be that there was a different self-esteem after taking the certification course. Maybe the tester had more self-esteem during the interview than before, that got him onto his or her first job. So it was not certification in itself that helped with the self-esteem, but the reaction to passing the certification test that did.
Certifications demonstrate that you are serious enough about testing to take an exam
Certifications demonstrate that you have put your body into some class. Sure. Maybe there is a questionnaire which have to answer after the course to the cert. But was it the class you are serious about, or rather your seriousness about testing that did? Who paid the rather expensive costs for the class? Was it the person herself? Or was it the company she worked for at that time? Are these questions asked during a job interview?
Back in 2009, I submitted a talk on Agile practices in a traditional environment to the Agile Testing Days. While I was preparing my talk, I noticed that I could attend one of the tutorials. I asked my employer whether he would pay for it, but he refused to do so. So I paid my own expenses for the ATDD course from Elisabeth Hendrickson.
Consider a job interviewer who was sent to a certification program and passed the test, and one who took his own expenses for an uncertified course. Which one of the two showed more dedication to testing?
Against testing as a second-class activity
In part 2 of Dorothy’s blog series the main argument for the creation of the course syllabus was to overcome the second-class citizenship of testers in organizations. Back in the 1990s she writes testers were a necessary evil in most organizations, if testing was done at all.
Reflecting on maybe twenty years of testing certification, it appears to me that this didn’t change. I have seen companies in the past five years which treated testers as second-class citizens. I have seen companies working on software used in hospital that did not have a single tester in their company at all. And I have seen testing being done by programmers and business experts.
Based upon what I have seen, I acknowledge that the initial reasons to create a testing certification program were honorable, but it appears to me that the certification program did not deliver anything at all about.
So, I’m asking, but what does deliver a first-class citizenship to testers? I have no data on this, but I sincerely believe that integrity, support for customers, programmers, and project managers helps in large extents to deliver this. A while ago, Matt Heusser had an awesome blog entry on this: How to be a first-class citizen as tester. To cite from that piece:
if you want to be a first-class citizen, one way to do it is by focusing on demonstrating value and competence.
Pointing to my cert does not demonstrate any value or competence, but rather that you are guilty of killing a tree.
Certification does not assess testers
Now, let me get to the final piece of my argument. Certifications do not assess testers for their skills. Answering a multiple-choice test does not make them any competent in applying all the theory they have learned. But my point goes way beyond this single point.
With their body of knowledge the syllabus creates the ground for the third obstacle to innovation that Jerry Weinberg mentions in Becoming a technical leader. The first and the second great obstacles to innovation are according to Weinberg self-blindness – I can’t see that I’m wrong – and no-problem syndrome – I don’t admit that I’m wrong. The third obstacle is the belief in a single possible solution, and that science has this answer handy for us.
Yes, the body of knowledge creates a thought-limiting notion. Testers working after the syllabus are working in a mind-set that anything outside the syllabus is impossible. I had just one of them in my class last week. I met another at a local user group meeting in Germany, both pray to the same single solution belief. The belief was that anything not in the requirements document is not a requirement. This is just plain wrong. I could educate the tester last week, but failed to do so at the user group meeting.
A software product is developed in order to solve a problem for some particular person. When you build a software to track your working hours per clients in a consultancy then you are probably solving the problem of getting all the hours from all your colleagues together in one place. Or you are solving the problem of billing all of your clients. Or… Now, if any of these problems is not solved, then the person that has the problem – your boss, your clients, your company – is unlikely to state that his problem is solved whether the problem is stated in the requirements document or not.
It appears to me that certifications currently create ground for the single-solution fallacy in software testing. Especially the notion that everything that is a requirement is put up in a requirements document. Now, go back to your documents, read it line by line and count the times that you could add “… and the software shall not make the house explode” to any of the sentences.
Let me recap my points. Certifications may appear to help you get a first job, but ask yourself if self-esteem was the real enabler in this, or rather the paper. Certifications are not a good starting point, since the syllabus may turn out a thought-limiting – or though control. Rather consider what could help you, and seek a course adapted to your needs. Seriousness for testing and knowledge of testing is better shown by showing skill and self-education on the job. Second-class citizenship is not solved by a piece of dead wood, but rather by showing your capabilities, and helping your team understand your challenges. Finally, certification does not show any skills in the tester, or whether she may apply it.
Last, the German DLRG also has a certification. The German DLRG is the institution for rescue swimmers in Germany. They provide baywatch services as well as services within indoor and outdoor swimming pools.
If you achieve to get their gold certification, you are allowed to take on work on your own. After two years this certification is downgraded to a silver certification which lasts all your life long. Any rescue swimmer employing organization in Germany understands the need to assess the skill of their rescue swimmers. Do you have a life-long valid testing certification? Do you need to show your skills?