I had heard a great deal about this little book. Some who had read it appreciated its premise, some were in fairly fundamental disagreement. If a text generates polar opposites of agreement, then that immediately piqued my interest! So lets begin with that premise:
“A Test Axiom is something we believe to be self evident and we cannot imagine an exception to it”
I feel this is indeed a risky premise for an approach to testing, could be easily misinterpreted as a set of iron laws to be followed, which will magically output an appropriate approach to testing. With this in mind I set about the enjoyable challenge of the dissection of these axioms. Lets take for example:
“Testing requires a known, controlled environment”
There are absolutely benefits to this statement, but also damaging flipsides to your test approach. A known, controlled environment is limited in variance, therefore only able to expose bugs of a certain nature. In addition, tests run in an environment such as this can be false signals, as the variance and scale of the real world changes outcomes.
On the reverse of this, I found a number of ‘axioms’ more challenging:
“Testing needs stakeholders”
I can imagine a great deal of variance here in terms of who the stakeholder is, their agenda and beliefs but testing without an audience? Can I imagine where this is not axiomatic? Stakeholders may see testing as a ‘hoop to jump through’ rather than an a skilful way of providing them the information they need, and feel they don’t need testing, but testing needs stakeholders to provide context for scope and priority decisions.
The ‘Axioms’ form part of the ‘First Equation of Testing’:
“Axioms + Context + Values + Thinking = Approach”
I found this to be another challenging assertion, as the application of axioms in an equation could be interpreted as a formula for success, whereas the real challenge of testing exists with the spaces between the constituent parts of the formula and how they interact. I see danger in creating formula’s and checklists for testing, as it perpetuates the linear, tickbox impression of testing as a craft. In fairness to the author, the overall tone of the piece encourages the application of the axioms and formula as part of a wider toolkit.
Although I found myself disagreeing with (what I saw as) the overall premise of the text, the author strongly advocates a focus on stakeholders and, by extension, providing the right information at the right time to those who make decisions. These sections are well worth reading and paying attention to, I certainly have applied some of those ideas to my recent work and provided an excellent framework for thought and approach. The author builds from a fairly formal approach to testing to give due attention to the spectrum of formality and the value of a varied approach. Initially I felt the text suffered from a lack of acknowledgement of the complexity of contemporary systems but this grew as the text progressed, which helped to provide a more rounded view of testing.
I found the authors real world experience shone through towards the end of the text, the reality of delivery is evident, although I think the author leans too far towards testing being concerned with the cost of failure rather than the benefit of the system meeting its mission. Both are important but I prefer the positive side of this coin and I believe testing enjoys a better profile when approached from this standpoint.
Thoroughly recommended read for testers of all experience and seniority, I will use some of the ‘axioms’ as part of my testing toolkit, although with an eye on their fallibility. I’ll end with my favourite ‘axiom’, which is certainly one I believe in:
“Testing never finishes. It stops.”