This is the second part of the series software testing is, based on the mind map I provided in the initial post, you can take a look here: http://mrslavchev.com/2016/11/04/software-testing-is-part-1/
I bet that big part of the readers of this blog will be puzzled by the presence of the words software and social science in the same sentence and that’s OK. Here’s how it goes.
The current state of thinking about software testing.
When I was a newbie, not that I consider myself experienced now, I am just an old newbie, so when I was a newbie and first started studying the software testing craft I was presented with the following picture of software testing.
Software testing is a technical activity that relies on understanding of the technical side of the product, such as:
- How the product is built – structure and frameworks
- How specific functionalities are implemented?
- What protocols are used for communication?
- How does an application works – databases, interfaces, memory, processor usage?
- How the internet operates – basic networking knowledge?
- How is the application dependent on its environment – OS, hardware, 3-rd party libraries?
- and many more
All of this is correct, but far from enough. It would be enough, if we only accept the definition that an application is just a series of instructions executed following a specific logic.
I like the example that Cem Kaner gives in the BBST foundation course, saying the above is like saying a house is just a bricks and materials ordered in a specific way, which is not true. A house is a place where someone can live and feel at comfortable, so there’s something more to a house than just the materials. Same with a software product – it’s not just the code, but it comes to partially replace a human interaction, but I will get into more depth in this a bit later.
The soft skills people.
Suddenly, when speaking of testing a person or two will show up saying that testing is not just technical, but incorporates some so-called “soft skills”. By “soft skills” they normally mean some part of negotiation skills, some part of team work abilities, some part of leadership, presentation skills, communication, etc. I say “some part”, because, so far, I didn’t hear or read any credible explanation what “are soft skills made of”, exactly, normally they are a small amount of this and that, etc.
I also agree, that this is important, too, but too general, in my opinion. It’s expected by a good professional in almost any area, interacting with clients/colleagues or any other human beings, to be a good communicator or a good negotiator and a good team player. So, I personally consider these – basic corporate survival skills, they are not something that’s strictly specific to software testing. So, what is it?
Software testing is rooted in social science.
I mentioned earlier that a software product is not just a bunch of computer instructions ordered following specific logic. Software applications came to existence in order to replace, although partially, a human interaction, therefore, software applications are not only there to do what they are supposed to do, but they also have to bring us comfort and satisfaction. To understand how an application comforts our clients and how we can measure this in a reliable way, we need to understand the human nature and this is where social sciences come to help us. Or, if I have to use the house analogy Kaner made, it will be more useful to know how to evaluate a comfortable house, rather than just knowing how it’s built.
So, many times, under the influence of Kaner, Bach, Bolton and the CDT community in general, I have claimed that software testing is a science, or may be not a science, but uses many scientific methods and approaches and it corresponds with other sciences, not only but we can apply knowledge from other sciences in testing, so our testing can evolve and be more diverse. The aim of this topic is to give some perspective how testing corresponds with social science and which branches of it, I believe were useful for me until now. It’s important to mention that I am not considering myself an expert in any of the following, so all of the conclusions are made by my personal, non-expert, observations.
In fact psychology is a science that can help any aspect of life and it has so many branches and sub-branches, that it will be practically impossible to cover all the possible applications it can have on software testing.
Psychology is science about the mind and behavior and that’s something that we rely on very heavily in testing, in two aspects:
In order to satisfy our clients’ needs, producing specific software product, we want to make them happy, we want to influence their emotions. And in no way, we want to influence their emotions in a negative way. That’s why, psychological knowledge can be useful in software testing, specifically in areas like usability evaluation. If you want more information on this I recommend you the book “The Design of Everyday Things” by Don Norman where you can learn how design of some utensils we use in our daily life has actually the ability to call for specific action.
The second aspect in which psychology can be very useful to software testing is sort of reflection oriented. We are professionals in assessing quality and performing experiments, but how can we say if we make the right experiments, if we have the correct judgement or in other words – how do we test a tester ?
Well, there’s no guideline actually, how to be a better tester or how to have the proper judgement, but we can be aware of what our own weak sides are or what known “defects and features” (because most of them are actually features) our minds have. And this is where cognitive psychology can help a lot. We need to know how our minds can trick us, how our perception might be distorted, absolutely naturally and how our judgement can become biased. Another interesting topic that cognitive psychology reviews is the heuristic method for solving complex problems. This is a topic that is often discussed in the CDT community and a popular approach of solving testing problems, even when people are unaware they are doing it and don’t call it a heuristic.
A book that might be of great help to those curious of you, which makes a great analysis of all of these is: “Thinking, Fast and Slow” by Daniel Kahneman
Quantitative and qualitative research
Again, without any ambition to be an expert – this is something that is widely used in other fields like statistics and marketing and believe it or not a problem that we solve in testing every day.
In testing we often have that argument on whether or not our testing process should be structured around trying to provide some specific metrics or focus on experiment that is profound and provides valuable contextual information.
Seems that in science this is a problem that existed for ages in the face of quantitative and qualitative scientific research methods. And we deal with it in testing as well. How many times have you heard the argument on whether or not we should use metrics to direct our testing and if yes what kind of metrics, how do we translate these metrics in the language of quality and risks? How are they meaningful for us and for our peers?
On the other hand, there’s this group of people saying that metrics are easily manipulatable and they give us a good statistical results, but bad human or contextual results. They advice us to more “qualitative”, human and contextual approach in testing, which is in its nature experimentation.
Another question arises on how we choose one or the other, or if we have to choose at all? Do we need to combine them? If yes, how do we combine them, which is first, which is more important.
Seems to me we have a lot to learn from the field of qualitative and quantitative scientific research, in fact while preparing this article I reviewed some materials that made me think of a separate post on this topic only, so expect a follow-up on this in the near future.
This might seem a little bit odd to you may be, but expressing ourselves in testing is actually a vital part of our job. I often see people having arguments on the words that we use in testing or what a specific testing term means, what is the purpose of using this term over that term, I also often see these distinctions being qualified as “just semantics”, but I don’t think people realize how vital are these semantics for the quality of our craft. And believe me, this comes from me – a person that normally communicates his thoughts in testing in a language that’s not his native, so I am used to saying something stupid or incorrect and having to deal with the result.
Here’s how proper use of the language can be crucial for testing:
Bug reporting is not just a routine task that we do in our day-to-day activities, although it looks like one. Quality bug report is not “just a bug report”, it’s your evaluation of the quality of the product, sometimes it’s the only proof of your input to the product. So, preparing a high quality bug report is crucial for everyone in the team and the product itself.
The more you develop that skill, the more credibility you and your job gain in the eyes of development, analysts, management and so on. And believe me, no matter how much time and effort you’ve put into building it, you can ruin it within a day, by letting negligence crawl into your bug reporting language, so it is very important to be careful, professional and specific in what we report as a bug. Don’t forget there are people depending on our judgement.
Describe your strategy and tell it in a compelling way
I think I wrote about this for probably a thousand times and a thousand times more won’t be enough to put the accent on how important it is to be able to explain what you do in a compelling and professional way.
Turns out in my experience, that many testers, even experienced ones have no issues performing their job as testers, including all their routine tasks, all but one – being able to explain what do they do. In other words, in the words of Harry Collins in fact, they hold a lot of “tacit” knowledge and they have problems turning it into explicit, in fact sometimes we even fail to explain our explicit knowledge in a good way, without oversimplifications.
This is why linguistic knowledge and the rich vocabulary and the ability to be precise in what you do are so important. Yes, we do report bugs, but no one is looking for bug reporters, after all, they are looking for testing professionals. In order to be professional tester you have to know testing, to be able to do it well and to be able to explain and defend the approach that you’ve chosen and sound like a professional while you do it.
Philosophy has so many branches and so many great thinkers involved in it, that I can only wish I can make a profound analysis how it is involved in testing. Also, I am probably not going to dive too deep in it, as I plan to write a separate blog post on how software testing is related to epistemology.
Here’s just some quick thoughts on how philosophy is related to testing. What does philosophy means? From ancient Greek, the literal translation is – “love to wisdom”. Well, that’s interesting, I think I’ve seen this somewhere in IT as well, remember these guys that are always ready to learn more and more about the product, that claim software testing is perpetual learning ?
And there’s actually more, but hey, I don’t want to spoil all of it all at once.
And I think this topic became waaay much longer than planned. I hope with the above I gave some more ideas to think on next time you speak with someone on how “technical” testing is and I hope you consider the big part of social science knowledge that we use, in order to be valuable testers.
Of course, I would love to read your thoughts on the topic, that’s what comments are for. Like, share and retweets are always appreciated. Good luck. 🙂