Alternatives to “Manual Testing”: Experiential, Attended, Exploratory

August 24th, 2021

This is an extension on a long Twitter thread from a while back that made its way to LinkedIn, but not to my blog.

No one ever sits in front of a computer and accidentally compiles a working program, so people know — intuitively and correctly — that programming must be hard. But almost anyone can sit in front of a computer and stumble over bugs, so people believe — intuitively and incorrectly — that testing must be easy!

Testers who take testing seriously have a problem with getting people to understand testing work.

The problem is a special case of the insider/outsider problem that surrounds any aspect of human experience: most of the time, those on the outside of a social group—a community; a culture; a group of people with certain expertise; a country; a fan club—don’t understand the insider’s perspective. The insiders don’t understand the outsiders’ perspective either.

We don’t know what we don’t know. That should be obvious, of course, but when we don’t know something, we have no idea of how little we comprehend it, and our experience and our lack of experience can lead us astray. “Driving is easy! You just put the car in gear and off you go!” That probably works really well in whatever your current context happens to be. Now I invite you to get behind the wheel in Bangalore.

How does this relate to testing? Here’s how:

No one ever sits in front of a computer and accidentally compiles a working program, so people know—intuitively and correctly—that programming must be hard.

By contrast, almost anyone can sit in front of a computer and stumble over bugs, so people believe—intuitively and incorrectly—that testing must be easy!

In our world of software development, there is a kind of fantasy that if everyone is of good will, and if everyone tries really, really hard, then everything will turn out all right. If we believe that fantasy, we don’t need to look for deep, hidden, rare, subtle, intermittent, emergent problems; people’s virtue will magically make them impossible. That is, to put it mildly, a very optimistic approach to risk. It’s okay for products that don’t matter much. But if our products matter, it behooves us to look for problems. And to find deep problems intentionally, it helps a lot to have skilled testers.

Yet the role of the tester is not always welcome. The trouble is that to produce a novel, complex product, you need an enormous amount of optimism; a can-do attitude. But as my friend Fiona Charles once said to me—paraphrasing Tom DeMarco and Tim Lister—”in a can-do environment, risk management is criminalized.” I’d go further: in a can-do environment, even risk acknowledgement is criminalized.

In Waltzing With Bears, DeMarco and Lister say “The direct result of can-do is to put a damper on any kind of analysis that suggests ‘can’t-do’…When you put a structure of risk management in place, you authorize people to think negatively, at least part of the time. Companies that do this understand that negative thinking is the only way to avoid being blindsided by risk as the project proceeds.”

Risk denial plays out in a terrific documentary, General Magic, about a development shop of the same name. In the early 1990s(!!), General Magic was working on a device that — in terms of capability, design, and ambition — was virtually indistinguishable from the iPhone that was released about 15 years later.

The documentary is well worth watching. In one segment, Marc Porat, the project’s leader, talks in retrospect about why General Magic flamed out without ever getting anywhere near the launchpad. He says, “There was a fearlessness and a sense of correctness; no questioning of ‘Could I be wrong?’. None. … that’s what you need to break out of Earth’s gravity. You need an enormous amount of momentum … that comes from suppressing introspection about the possibility of failure.”

That line of thinking persists all over software development, to this day. As a craft, the software development business systematically resists thinking critically about problems and risk. Alas for testers, that’s the domain that we inhabit.

Developers have great skill, expertise, and tacit knowledge in linking the world of people and the world of machines. What they tend not to have—and almost everyone is like this, not just programmers—is an inclination to find problems. The developer is interested in making people’s troubles go away. Testers have the socially challenging job of finding and reporting on trouble wherever they look. Unlike anyone else on the project, testers focus on revealing problems that are unsolved, or problems introduced by our proposed solution. That’s a focus which the builders, by nature, tend ot resist.

Resistance to thinking about problems plays out in many unhelpful and false ideas. Some people believe that the only kind of bug is a coding error. Some think that the only thing that matters is meeting the builders’ intentions for the product. Some are sure that we can find all the important problems in a product by writing mechanistic checks of the build. Those ideas reflect the natural biases of the builder—the optimist. Those ideas make it possible to imagine that testing can be automated.

The false and unhelpful idea that testing can be automated prompts the division of testing into “manual testing” and “automated testing”.

Listen: no other aspect of software development (or indeed of any human social, cognitive, intellectual, critical, analytical, or investigative work) is divided that way. There are no “manual programmers”. There is no “automated research”. Managers don’t manage projects manually, and there is no “automated management”. Doctors may use very powerful and sophisticated tools, but there are no “automated doctors”, nor are there “manual doctors”, and no doctor would accept for one minute being categorized that way.

Testing cannot be automated. Period. Certain tasks within and around testing can benefit a lot from tools, but having machinery punch virtual keys and compare product output to specificed output is not more “automated testing” than spell-checking is “automated editing”. Enough of all that, please.

It’s unhelpful to lump all non-mechanistic tasks in testing together under “manual testing”. Doing so is like referring to craft, social, cultural, aesthetic, chemical, nutritional, or economic aspects of cooking as “manual” cooking. No one who provides food with care and concern for human beings—or even for animals—would suggest that all that matters in cooking is the food processors and the microwave ovens and the blenders. Please.

If you care about understanding the status of your product, you’ll probably care about testing it. You’ll want testing to find out if the product you’ve got is the product you want. If you care about that, you need to understand some important things about testing.

If you want to understand important things about testing, you’ll want to consider some things that commonly get swept a carpet with the words “manual testing” repeatedly printed on it. Considering those things might require naming some aspects of testing that you haven’t named before.

Think about experiential testing, in which the tester’s encounter with the product, and the actions that the tester performs, are indistinguishable from those of the contemplated user. After all, a product is not just its code, and not just virtual objects on a screen. A software product is the experience that we provide for people, as those people try to accomplish a task, fulfill a desire, enjoy a game, make money, converse with people, obtain a mortgage, learn new things, get out of prison…

Contrast experiential testing with instrumented testing. Instrumented testing is testing wherein some medium (some tool, technology, or mechanism) gets in between the tester and the naturalistic encounter with and experience of the product. Instrumentation alters, or accelerates, or reframes, or distorts; in some ways helpfully, in other ways less so. We must remain aware of the effects, both desirable and undesirable, that instrumention brings to our testing.

Are you saying “manual testing”? You might be referring to the attended or engaged aspects of testing, wherein the tester is directly and immediately observing and analyzing aspects of the product and its behaviour in the moment that the behaviour happens. And you might want to contrast that with the algorithmic, unattended things that machines do—things that some people label “automated testing”—except that testing cannot be automated. To make something a test requires the design before the automated behaviour, and the interpretation afterwards. Those parts of the test, which depend upon human social competence to make a judgement, cannot be automated.

Are you saying “manual”? You might be referring to testing activity that’s transformative, wherein something about performing the test changes the tester in some sense, inducing epiphanies or learning or design ideas. Contrast that with procedures that are transactional: rote, routine, box-checking. Transactional things can be done mechanically. Machines aren’t really affected by what happens, and they don’t learn in any meaningful sense. Humans do.

Did you say “manual”? You might be referring to exploratory work, which is interestingly distinct from experiential work as described above. Exploratory—in the Rapid Software Testing namespace at least—refers to agency; who or what is in charge of making choices about the testing, from moment to moment. There’s much more to read about that.

Wait… how are experiential and exploratory testing not the same?

You could be exploring—making unscripted choices—in a way entirely unlike the user’s normal encounter with the product. You could be generating mounds of data and interacting with the product to stress it out; or you could be exploring while attempting to starve the product of resources. You could be performing an action and then analyzing the data produced by the product to find problems, at each moment remaining in charge of your choices, without control by a formal, procedural script.

That is, you could be exploring while encountering the product to investigate it. That’s a great thing, but it’s encountering the product like a tester, rather than like a user. It might not be a great idea to be aware of the differences between those two encounters, and take advantage of them, and not mix those up.

You could be doing experiential testing in a highly scripted, much-less-exploratory kind of way; for instance, following a user-targeted tutorial and walking through each of its steps to observe inconsistencies between the tutorial and the product’s behaviour. To an outsider, your encounter would look pretty much like a user’s encounter; the outsider would see you interacting with the product in a naturalistic way, for the most part—except for the moments where you’re recording observations, bugs, issues, risks, and test ideas. But most observers outside of testing’s form of life won’t notice those those moments.

Of course, there’s overlap between those two kinds of encounters. A key difference is that the tester, upon encountering a problem, will investigate and report it. A user is much less likely to do so. (Notice this phenomenon, while trying to enter a link from LinkedIn’s Articles editor; the “apply” button isn’t visible, and hides off the right-hand side of the popup. I found this while interacting with Linked experientially. I’d like to hope that I would have find that problem when testing intentionally, in an exploratory way, too.)

There are other dimensions of “manual testing”. For a while, we considered “speculative testing” as something that people might mean when they spoke of “manual testing”; “what if?” We contrasted that with “demonstrative” testing—but then we reckoned that demonstration is not really a test at all. Not intended to be, at least. For an action to be testing, we would hold that it must be mostly speculative by nature.

And here’s the main thing: part of the bullshit that testers are being fed is that “automated” testing is somehow “better” than “manual” testing because the latter is “slow and error prone”—as though people don’t make mistakes when they apply automation to checks. They do, and the automation enables those errors at a much larger and faster scale.

Sure, automated checks run quickly; they have low execution cost. But they can have enormous development cost; enormous maintenance cost; very high interpretation cost (figuring out what went wrong can take a lot of work); high transfer cost (explaining them to non-authors).

There’s another cost, related to these others. It’s very well hidden and not reckoned: we might call it interpretation cost or analysis cost. A sufficiently large suite of automated checks is impenetrable; it can’t be comprehended without very costly review. Do those checks that are always running green even do anything? Who knows?

Checks that run red get frequent attention, but a lot of them are, you know, “flaky”; they should be running green when they’re actually running red. Of the thousands that are running green, how many should be actually running red? It’s cognitively costly to know that—so people routinely ignore it.

And all of these costs represent another hidden cost: opportunity cost; the cost of doing something such that it prevents us from doing other equally or more valuable things. That cost is immense, because it takes so much time and effort to to automate GUIs when we could be interacting with the damned product.

And something even weirder is going on: instead of teaching non-technical testers to code and get naturalistic experience with APIs, we put such testers in front of GUIish front-ends to APIs. So we have skilled coders trying to automate GUIs, and at the same time, we have non-programming testers, using Cypress to de-experientialize API use! The tester’s experience of an API through Cypress is enormously different from the programmer’s experience of trying use the API.

And none of these testers are encouraged to analyse the cost and value of the approaches they’re taking. Technochauvinism (great word; read Meredith Broussard’s book Artificial Unintelligence) enforces the illusion that testing software is a routine, factory-like, mechanistic task, just waiting to be programmed away. This is a falsehood. Testing can benefit from tools, but testing cannot be mechanized.

Testing must be seen as a social (and socially challenging), cognitive, risk-focused, critical (in several senses), analytical, investigative, skilled, technical, exploratory, experiential, experimental, scientific, revelatory, honourable craft. Not “manual” or “automated”. Let us urge that misleading distinction to take a long vacation on a deserted island until it dies of neglect.

Testing has to be focused on finding problems that hurt people or make them unhappy. Why? Because optimists who are building a product tend to be unaware of problems, and those problems can lurk in the product. When the builders are aware of those problems, the builders can address them. Whereby they make themselves look good, make money, and help people have better lives.

Rapid Software Testing (for American daytimes and European evenings) runs September 13-16, 2021. Register now!

Exact Instructions vs. Social Competence

July 5th, 2021

An amusing video from a few years back has been making the rounds lately. Dad challenges the kids to write exact instructions to make a peanut butter and jelly sandwich, and Dad follows those instructions. The kids find the experience difficult and frustrating, because Dad interprets the “exact” instructions exactly—but differently from the way the kids intended. I’ll be here when you get back. Go ahead and watch it.

Welcome back. When the video was posted in a recent thread on LinkedIn, comments tended to focus on the need for explicit documentation, or more specific instructions, or clear direction.

In Rapid Software Testing, we’d take a different interpretation. The issue here is not that instructions are unclear, or that the kids have expressed themselves poorly. Instead, we would emphasize that communicating clearly, describing intentions explictly, and performing actions appropriately all rely on tacit knowledge—knowledge that has not been made explicit. In that light, the kids did a perfectly reasonable job at describing the assignment.

Notice that the kids do not describe what peanut butter is; they do not have to not tell the father that one must twist the lid on the peanut butter jar to open it; nor do they have to explain that the markings on the paper are words representing their intentions. The father has sufficient tacit knowledge to be aware of those things. At a very young age, through socialization, observation, imitation, and practice, the dad acquired the tacit knowledge required to open peanut butter jars, to squeeze jelly dispensers without crushing them, to use butter knives to deliver peanut butter from jar to bread, to make reasonable inferences about what the “top” of the bread is, and so forth.

Even though he has sufficient tacit knowledge to interpret instructions for making a peanut butter and jelly sandwich, the dad pretends that he doesn’t. What makes the situation in the video funny for us and exasperating for the kids is our own tacit knowledge of things the father presumably should know as a normal American dad in a normal American kitchen. In particular, we’re aware that he should be able to interpret the instructions competently; to repair differences between the actions the kids intended him to take and the ones he chose to take.

In certain circles, there is an idea that “better requirements documents” or “clear communication” or “eliminating ambiguity” are royal roads to better software development and better testing. Certainly these things can help to some degree, but organizing teams and building products requires far more than explicit instructions. It requires the social context and tacit knowledge to interpret things appropriately. Dad misinterpreted on purpose. Development and testing groups can easily misintrepret by accident; unintentionally; obliviously.

Where do explicit instructions come from? Would they be any good if they weren’t rooted in knowledge about the customers’ form of life, and knowledge of the problems that customers face—the problems that the product could help to solve? Could they be expressed more concisely and more reliably when everyone involved had shared sets of feelings and mental models? And would exact instructions help if the person (or machine) receiving them didn’t have the social competence to interpret them appropriately?

In RST, we would hold that it’s essential for the tester to become immersed in the world of the product and in the customers’ forms of life to the greatest degree possible—a topic for posts to come.

Testers: Focus on Problems

June 16th, 2021

A tester writes:

“I’m testing an API. It accepts various settings and parameters. I don’t know how to get access to these settings from the API itself, so I’m stuck with modifying them from the front end. Moreover, some responses to specific requests are long and complicated, so given that, I have no idea how to test it! Online examples of API testing tend to focus on checking the response’s status code, or verification of the schema, or maybe for correctness of a single string. How can I make sure that the whole response is correct?”

My reply:

You can’t.

That reply may sound disconcerting to many testers, but it’s true. The good news is that there’s a practical workaround when you focus less on demonstrating correctness and more on discovering problems.

Among other things, being unsure whether something is correct suggests problems on its own. For instance, your question reveals an immediate problem: you’re not familiar enough with the product or how to test it. Not yet. And, I hasten to point out: that’s okay.

Being unsure about how to test something is always the case to some degree. As you’re learning to test, and as you’re learning to test a product that’s new to you, some uncertainty and confusion is normal. Don’t worry about that too much. To test the product you must learn how to test it. You learn how to test the product by trying to test the product—and by reporting and discussing the problems you encounter. Learning is an exploratory process. Apply an exploratory approach to discover problems.

For instance: I can see from your question that you’ve already discovered a problem: you’ve learned that your testing might be harder or slower without some kind of tool support that allows you to set options and parameters quickly and conveniently. Report that problem. Solving it might require some kind of help from the designers and developers in supplying APIs for setup and configuration.

If those APIs don’t exist, that’s a problem: the intrinisic testability of your product is lower than it could be. When testing is harder or slower, given the limited time you have to test, it’s also shallower. The consequence of reduced testability is that you’re more likely to miss bugs. It would be a good idea to make management aware of testability problems now. Report those problems. If you do so, you’ll already have given at least part of an answer when management inevitability asks, perhaps much later, “Why didn’t you find that bug?”

Moreover, if those setup and configuration APIs don’t exist, there’s a good chance that it’s not only a problem for you; it will probably be a problem for people who want to maintain and develop the product, and for people who want to use it. Report that problem too.

If those APIs do exist, and they’re not described somehow or somewhere, that’s a problem for you right now, but sooner or later it will be a problem for others. Inside developers who maintain the product now and in the future need to understand the API, and outside developers who want to use the product through the API need to be able understand it quickly and easily too. Without description, the API is of severely limited usefulness. Report that problem. Mind you, missing documentation is a problem that you can help to address while testing.

If there is a description of the API, but the description is inaccurate, or unclear, or out of date, you’ll soon find that it doesn’t match the product’s behaviour. That’s a problem for several reasons: for you, the tester, it won’t be clear whether the product or the documentation is correct; inside developers won’t know whether or how to fix it; and outside developers will find that, from their perspective, the product doesn’t work, and they’ll be confused as to whether there’s an error in the documentation or a bug in the API. Report that problem.

If those APIs do exist and they’re documented but you don’t know how to design or perform tests, or how to analyse results efficiently (yet), that’s a problem too: subjective testability is low. The quality of your testing depends on your knowing the product, and the product domain, and the technology. That knowledge doesn’t come immediately. It takes working with something and with the people who built it to know how to test it properly, to learn how deeply it needs to be tested, to develop ideas about risk, and to recognize hidden, subtle, rare, intermittent, emergent problems.

To learn the product, you’ll probably need to be able to talk things over with the developers and your other testing clients, but that’s not all. To learn the product well, you’ll need experience with it. You’ll need to engage with it and interact with it. You’ll need to develop mental models of it. You’ll need to anticipate how people might use it, and how they might have trouble with it. You must play, feel your way around, puzzle things out, and be patient as you encounter and confront confusion. The confusion lifts as you immerse yourself in the world of the product.

However, management needs to know that that learning time is necessary. While you’re learning about where and how to find deep bugs, you won’t find them deliberately. At first, you’ll tend to stumble over bugs accidentally, and you might miss important bugs that are right in front of you. Again, that’s normal and natural until you’ve learned the product, have figured out how to test it, and have become comfortable with your oracles and your tool sets.

Hold up — what’s an oracle? An oracle is a means by which we recognize a problem when we encounter one in testing. And this is where we return to issues around correctness.

After making an API call as part of a test, you can definitely match elements of the response with a reference—an example, or a list, or a table, or a bit of JSON that someone has else provided, or that you’ve developed yourself. You could compare elements in the response individually, one by one, and confirm that each one seemed to be consistent with your reference. You could observe these things directly, with your own eyes, or you could write code to mediate your observation.

If you see some inconsistency, you can suspect that there’s a problem, and report that. If each element in the output matches the reference, and all of the elements match the reference, and there don’t seem to be any extra elements in the output, you can assert that the response appears to be correct, and from that you can infer that the response is correct. But even then, you can’t make sure that the response is correct.

One issue here is that correctness of output is always relative to something, and to someone’s notion of consistency with that something. You could assert that the response seems to be consistent with the developers’ intentions, to the degree that you’re aware of those intentions, and to the degree that your awareness is up to date. Of course, the developers’ intentions could be inconsistent with what the project manager wanted, and all of that could be inconsistent with what some customer would perceive as correct. Correctness is subject to the Relative Rule: that is, “correct” really means “correct to some person, at some time”.

If you don’t notice a problem, you can truthfully say so; you didn’t notice a problem. That doesn’t mean that product is correct. Correctness can refer to an overwhelming number of factors.

Is the output well-formed and syntactically correct? Is the output data accurate? Is it sufficiently precise? Is it overly precise? Are there any missing elements? Are there any extra elements? Has some function changed the original source data (say, from a database) while transforming it to an API response? Was the source data even correct to begin with?

Did the response happen in a sufficiently timely way? The output seemed to be correct, but is the system still in a stable state? The output appeared correct this time; will it be correct next time? Was the response logged properly? If there was an error, was an appropriate, helpful error message returned, such that it could be properly understood by an outside programmer? In terms of the questions we could ask, we’re not even getting started here.

Answering all such questions on every test is impractical, even with tool assistance. You could write code to check a gigantic number of conditions for an enormous number of outputs, based on a multitude of conditions for a huge set of inputs. That would result in an intractable amount of test output data to analyze, and would require an tremendous amount of code—and writing and debugging and maintaining all that code would be harder than writing and maintaining debugging code for the product.

The effort and expense wouldn’t be worthwhile, either. After something has been tested to some degree (for example, by developers’ low-level unit checks, or a smattering of integration checks, or by some quick direct interaction with the product), risk may be sufficiently low that we can turn our attention to higher-risk concerns. If that risk isn’t sufficiently low (perhaps because the developers haven’t been given time or resources to develop a reasonable understanding of the effects of changes they’re making), more testing on your part is unlikely to help. Report that problem.

So rather than focusing on correctness, I’d recommend that you focus on problems and risk instead; and that you report on anything that you believe could plausibly be a problem for some person who matters. There are (at least) three reasons for this; one social, one psychological, and one deeply practical.

The social reason is important, but sometimes a little awkward. The tester must be focused on finding problems because most of the time, no one else on the project is focused on risk and on finding problems. Everyone else is envisioning success.

Developers and designers and managers have the job of building useful products that make people’s problems go away. The tester’s unique role is to notice when the product is not solving those problems, and to notice when the product is introducing new problems.

This puts the tester at a different perspective from everyone else, and the tester’s point of view can seem disruptive sometimes. After all, not very many people love hearing about problems. The saving grace for the tester is that if there are problems about the product, it’s probably better to know about them before it’s too late, and before those problems are inflicted on customers.

The psychological reason to focus on problems is that if you focus on correctness, that’s what you’ll tend to see. You will feel an urge to show that the product is working correctly, and that you can demonstrate success. Your mind will not be drawn to problems, and you’ll likely miss a bunch of them.

And here’s the practical reason to focus on problems: if you focus on risk and search diligently for problems and you don’t find problems, you’ll be able to make reasonable inferences about the correctness part for free!

Unlimited Charges

May 17th, 2021

I noticed something interesting while reviewing my credit card bills a couple of evenings ago: monthly charges for $9.99 from “Amazon Downloads”, going back several months.

I buy a lot of e-books. I looked for receipts from Amazon in email. I found a bunch, but none from Amazon for $9.99. I never delete email receipts; I put all of them into a separate folder so that I can collate them and have the relevant ones as records of business expenses at tax time. No receipts for $9.99 on or near the associated dates.

I did a Google search for “Amazon Downloads charge 9.99”; it autocompleted before I got to the 9.99 part. It seems that I’m not alone.

It turns out that Amazon Kindle Unlimited subscriptions get started in a very subtle and hard-to-notice way, and then automatically renew monthly, without receipts being issued. Other auto-renewing services do provide notification each business cycle; Amazon doesn’t, so it seems.

And I found it intriguing how quickly they processed a refund when I pointed this out.

Within moments of tweeting about this, I heard from another fellow who had the same experience. Indeed I’m not alone.

This really bugs me. On the other hand, by Amazon’s lights, this is not a bug; the system is doing what it is designed to do. I would like to believe that if I were a tester at Amazon, I would have noted the subtlety of the auto-renewal option and the absence of monthly receipts as Severity-1 problems, inconsistent with an image that the organization would like to project or defend. After this experience, and feeling like I’ve been duped, I hope I would be more likely to notice and report these problems. And I’d like to believe that a tester at Amazon did just that, but I don’t know; I can’t know.

I can be more certain that management has not so far seen this as a problem, and I have a problem with that. Meanwhile, lesson learned; keep reviewing those credit card purchases diligently every month.

“Manual Testing”: What’s the Problem?

April 27th, 2021

I used to speak at conferences. For the HUSTEF 2020 conference, I had intended to present a talk called “What’s Wrong with Manual Testing?” In the age of COVID, we’ve all had to turn into movie makers, so instead of delivering a speech, I delivered a video instead.

After I had proposed the talk, and it was accepted, I went through a lot of reflection on what the big deal really was. People have been talking about “manual testing” and “automated testing” for years. What’s the problem? What’s the point? I mulled this over, and video contains some explanations of why I think it’s an important issue. I got some people — a talented musician, an important sociologist, a perceptive journalist and systems thinker, a respected editor and poet, and some testers — to help me out.

In the video, I offer some positive alternatives to “manual testing” that are much less ambiguous, more precise, and more descriptive of what people might be talking about: experiential testing (which we could contrast with “instrumented testing”; exploratory testing (which we have already contrasted with “scripted testing”; attended testing (which we could contrast with “unattended testing”); and there are some others. More about all that in a future post.

I also propose how it came to be that important parts of testing — the rich, cognitive, intellectual social, process of evaluating a product by learning about it through experiencing, exploring and experimenting — came to be diminished and pushed aside by obsessive, compulsive fascination with automated checking.

But there’s a much bigger problem that I didn’t discuss in the video.

You see, a few days before I had to deliver the video, I was visiting an online testing forum. I read a question from a test manager who wanted to interview and qualify “manual testers”. I wanted provide a helpful reply, and as part of that, I asked him what he meant by “manual testing”. (As I do. A lot of people take this as being fussy.)

His reply was that he was wanting to identify candidates who don’t use “automated testing” as part of their tool set, but who were to be given the job of creating and executing manually scripted human-language tests and performing all the critical thinking skills that both approaches require.

(Never mind the fact that testing can’t be automated. Never mind that scripting a test is not what testing is all about. Never mind that no one even considers the idea of scripting programmers, or management. Never mind all that. Wait for what comes next.)

Then he said that “the position does not pay as much as the positions that primarily target automated test creation and execution, but it does require deeper engagement with product owners”. He went on to say that he didn’t want to get into the debate about “manual and automated testing”; he said that he didn’t like “holy wars”.

And there we have it, ladies and gentlemen; that’s the problem. Money talks. And here, the money—the fact that these testers are going to be paid less—is implicitly suggesting that talking to machines is more valuable, more important, than deeper engagement with people.

The money is further suggesting that skills stereotypically associated with men (who are over-represented in the ranks of programmers) are worth more than skills stereotypically associated with women (who are not only under-represented but also underpaid and also pushed out of the ranks of programmers by chauvinism and technochauvinism). (Notice, by the way, that I said “stereotypically” and not “justifiably”; there’s no justification available for this.)

Of course, money doesn’t really talk. It’s not the money that’s doing the talking.  It’s our society, and people within it, who are saying these things. As so often happens, people are using money to say things they dare not speak out loud.

This isn’t a “holy war” about some abstract, obscure point of religious dogma. This is a class struggle that affects very real people and their very real salaries. It’s a struggle about what we value. It’s a humanist struggle. And the test manager’s statement shows that the struggle is very, very real.

Suggestions for the (New) Testers

April 23rd, 2021

A friend that I’m just getting to know runs a training and skills development program for new testers. Today he said, “My students are now starting a project which includes test design, test techniques, and execution of testing. Do you have any input or advice for them?” Here’s my reply.

Test design, test techniques, and execution of testing are all good things. I’d prefer performing tests to “test execution”. In that preference, I’m trying to emphasize that a test is a performance, by an engaged person who adapts to what he or she is experiencing. “Test execution” sounds more like following a recipe, or a programmed set of instructions.

Of these things, my advice is to perform testing first. But that advice can be a little confusing to people who believe that testing is only operating some (nearly) finished product in a search for coding errors. In Rapid Software Testing, we take a much more expansive view: testing is the process of evaluating a product by learning about it through experiencing, exploring and experimenting, which includes to some degree questioning, studying, modeling, observation, inference, etc.

Testing includes analysis of the product, its domain, the people using it, and risk related to all of those. Testing includes critical thinking and scientific thinking. Testing includes performing experiments—that is, tests—all the way along. But I emphasized the learning part just back there, because testing starts with learning, ends with reporting what we’ve learned, feeds back into more learning, and is about learning every step of the way.

We learn more most powerfully from experiencing, exploring, and experimenting; performing experiments; performing tests. So, my advice to the new tester is to start with performing tests to study the product, without focusing too much on test design and test techniques, at first.

Side note: the “product” that you’ve been asked to test may not be a full, working, running piece of software. It may be a feature or component or function that is a part of a product. It may be a document, or a design drawing, a diagram, or even an idea for a product or feature that you’re being asked to review. In the latter cases, “performing a test” might mean the performance of a thought experiment. That’s not the same as the real-world experience of the running product, hence the quotes around “performing a test”. A thought experiment can be a great and useful thing to help nip bugs in the bud, before bugs in an idea turn into bugs in a product. But if we want to determine the real status of the real product, we’ll need to perform real testing on the real product.

So: learn the product (or feature, or design, or document, or idea), and identify how people might get value from it. Survey the product to identify its functions, features, and interfaces. Explore the product, and gain experience with it by engaging in a kind of purposeful play. Don’t look for bugs, particularly—not right away. Look for benefits. Look for how the product is intended to help people get their work done, to help them to communicate with other people, to help them to get something they want or need, to help them to have fun. Try doing things with the product—accomplishing a task, having a conversation, playing the game.

Record your thoughts and ideas and feelings reasonably thoroughly. Pay attention to things that surprise you, or that trigger your interest, or that prompt curiosity. Note things that you find confusing, and notice when the confusion lifts. If you have been learning the product for a while, and that confusion hasn’t gone away, that’s significant; it means there’s some confusing going on. If you get ideas about potential problems (that is, risks), note those. If you get ideas for designing tests, or applying tools, note those too.

Capture what you’re learning in point form, or in mind maps, or in narratives of what you’re doing. Sketeches and diagrams can help too. Don’t make your notes too formal; formality tends to be expensive, and it’s premature at this stage. It might be a good idea to test with someone else, with one person focusing on interacting with the product, and the other minding the task of taking notes and observations. Or you might choose to narrate and record your survey of the product on video to review later on; or to use like the black boxes on airplanes to figure out what led to problems or crashes.

You’ll probably see some bugs right away. If you do, note them quickly, but don’t investigate them. If you spotted a bug this easily, this early, and you take a quick note about it, you’ll almost certainly be able to see the bug again later. Investigating shallow bugs is not the job at the moment. They job right now is to develop your mental model of the product, so that you become prepared to find bugs that are more subtle, more deeply hidden, and potentially much more important or damaging.

Identify the people who might use the product… and then consider other groups of people you might have forgotten. That would include novice users of the product; expert users of the product; experts in the product domain who are novice users of the product; impatient users; plodding users; users under pressure; disabled users… Consider the product in terms of things that people value: capability, reliability, usability, charisma, security, scalability, compability, performance, installability… (As a new tester, or a tester in training, you might know these as quality criteria.)

You might also want to survey the product from the perspective of people who are not users as such, but who are definitely affected by the product: customer support people; infrastructure and operations people; other testers (like testing toolsmiths, or accessibility specialists); future testers; current developers future developers… Think in terms of what they might value from the product: supportability, testability, maintainability, portability, localizability. (These are quality criteria too, but they’re focused on the internal organization more than on their direct benefit to the end user.)

Refine your notes. Create lists, mind maps, tables, sketches, diagrams, flowcharts, stories… whatever helps you to reflect on your experience.

Share your findings with other people in the test or development (or in this case, study) group. That’s very important. It’s a really good way both to share knowledge and to de-bias ourselves and to reveal things that we might have forgotten, ignored, or dismissed too quickly.

Have these questions in mind as you go: What is this that we’re building? Who are we building it for? How would they get value from it? As time goes by, you’ll start to raise other questions: What could go wrong? How would we know? How could people’s value might be threatened or compromised? How could we test this? How should we test this? Then you’ll be ready to make better choices about test design, and applying test techniques.

Of course, this isn’t just advice for the new tester. It applies to anyone who wants to do serious testing. Testing that starts by reading a document and leaps immediately to creating formal, procedurally scripted test cases will almost certainly be weak testing, uninformed by knowledge of the product and how people will engage with it. Testing that starts with being handed some API documentation and leaps to the creation of automated checks for correct results will miss lots of problems that programmers will encounter—problems that we could discover if we try to experience it the way programmers—especially outside programmers—will.

As we’re developing the product, we’re learning about it. As we’re learning the product, we’re developing ideas about what it is, what it does, how people might use it, and how they might get value from it, and that learning feeds back into more development. As we develop our understanding of the product more deeply, we can be much better prepared to consider how people might try to use it unsuccessfully, how they might misuse it, and how their value might be threatened. That’s why it’s important, I believe, to do test execution perform testing first—to prepare ourselves better for test design and for identifying and applying test techniques—so we can find better bugs.

This post has been greatly influenced by ideas on sympathetic testing that came to me—over a couple of decades—from Jon Bach, James Bach, and Cem Kaner.

Evaluating Test Cases, Checks, and Tools

April 11th, 2021

For testers who are being asked to focus on test cases and testing tools, remember this: a test case never finds a bug. The tester finds a bug, and the test case may play a role in finding the bug. (Credit to Pradeep Soundararajan for putting this so succinctly, all those years ago.)

Similarly, an automated check never finds a bug. The tester finds a bug, and the check may play a role in finding the bug.

A testing tool never finds a bug. The tester finds a bug, and the tool may play a role in finding the bug.

If you suspect that managers are putting too much emphasis on test cases, or automated checks, or testing tools—artifacts—, try this:

Start a list.

Whenever you find a bug, make a quick note about the bug and how you found it. Next to that, put a score on the value of the artifact. Write another quick note to describe and explain why you gave the the artifact a particular score.

Score 3 when you notice that an artifact was essential in finding the bug; there’s no way you could have found the bug without the artifact.

Score 2 if the artifact was significant in finding the bug; you could have found the bug, but the artifact was reasonably helpful.

Score 1 if the artifact helped, but not very much.

Score 0 if the artifact played no role either way.

Score -1 whenever you notice the artifact costing you some small amount of time, or distracting you somewhat.

Score -2 whenever the artifact when you notice the artifact costing you significant time or disruption from the task of finding problems that matter.

Score -3 whenever you notice that the artifact is actively preventing you from finding problems—when your attention has been completely diverted from the product, learning about it, and discovering possible problems in it, and has been directed towards the care and feeding of the artifact.

Notice that you don’t need to find a bug to offer a score. Pause your work periodically to evaluate your status and take a note. If you haven’t found a bug in the last little while, note that. In any case, every now and then, identify how long you’ve been on a particular thread of investigation using a test case, or a set of checks, or a tool. Evaluate your interaction with the artifact.

Periodically review the list with your manager and your team. The current total score might be interesting; if it’s high, that might suggest that your tools or test cases or other artifacts are helping you. If it’s low or negative, that might suggest that the tools or test cases or other artifacts are getting in your way.

Don’t take too long on the aggregate score; practically no time at all. It’s far more important to go through the list in detail. The more extreme numbers might be the most interesting. You might want to pay the greatest or earliest attention to the things that score the lowest and highest first, but maybe not. You might prefer to go through the list in order.

In any case, as soon as you begin your review of a particular item, throw away the score, because the score doesn’t really mean anything. It’s arbitrary. You could call it data, but it’s probably not valid data, and it’s almost certainly not reliable data. If people start using the data to control the decisions, eventually the data will be used to control you. Throw the score away.

What matters is your experience, and what you and the rest of the team can learn from it. Turn your attention to your notes and your experience. Then start having a real conversation with your manager and team about the bug, about the artifact or tool, and about your testing. If the artifact was helpful, identify how it helped, and how it might help next time, and how it could fool you if you became over-reliant on it. If the artifact wasn’t helpful, consider how it interfered with your testing, how you might improve or adjust it or whether you should put it to bed for a while or throw it away.

Learn from every discovery. Learn from every bug.

Related reading:

Assess Quality, Don’t Measure It

Flaky Testing

February 22nd, 2021

The expression “flaky tests” is evidence of flaky testing. No scientist refers to “flaky experimental results”. Scientists who observe inconsistency don’t dismiss it. They pay close attention to it, and probe it. They redesign their experiments or put better controls on them.

When someone refers to an automated check (or a suite of them) as a “flaky test”, the suggestion is that it represents an unreliable experiment. That assumption is misplaced. In fact, the experiment reliably shows that someone’s models of the product, check code, test environment, outcomes, theory, and the relationships between them are misaligned.

That’s not a “flaky experiment”. It’s an excellent experiment. The experiment is telling you something crucial: there’s something you don’t know. In science, a surprising, perplexing, or inconsistent result prompts scientists to begin an investigation. By contrast, in software, an inconsistent result prompts some people to shrug and ignore what the experiment is trying to tell them. Then they do weird stuff like calculating a “flakiness score”.

Of course, it’s very tempting psychologically to dismiss results that you can’t explain as “noise”, annoying pieces of red junk on your otherwise lovely all-green lawn. But a green lawn is not the goal. Understanding what the junk is, where it is, and how it gets there is the goal. It might be litter—it it might be a leaking container of toxic waste.

It’s not a great idea to perform a test that you don’t understand, unless your goal is to understand it and its relationship to the product. But it’s an even worse idea to dismiss carelessly a test outcome that you don’t understand. For a tester, that’s the epitome of “flaky”.

Now, on top of all that, there’s something even worse. Suppose you and your team have a suite of 100,000 automated checks that you proudly run on every build. Suppose that, of these, 100 run red. So you troubleshoot. It turns out that your product has problems indicated by 90 of the checks, but ten of the red results represent errors in the check code. No problem. You can fix those, now that you’re aware of the problems in them.

Thanks to the scrutiny that red checks receive, you have become aware that 10% of the outcomes you’re examining are falsely signalling failure when they are in reality successes. That’s only 10 “flaky” checks out of 100,000. Hurrah! But remember: there are 99,900 checks that you haven’t scrutinized. And you probably haven’t looked at them for a while.

Suppose you’re on a team of 10 people, responsible for 100,000 checks. To review those annually requires each person working solo to review 10,000 checks a year. That’s 50 per person (or 100 per pair) every working day of the year. Does your working day include that?

Here’s a question worth asking, then: if 10% of 100 red checks are misleadingly signalling a problem, what percentage of 99,900 green checks are misleadingly signalling “no problem”? They’re running green, so no one looks at them. They’re probably okay. But even if your unreviewed green checks are ten times more reliable than the red checks that got your attention (because they’re red), that’s 1%. That’s 999 misleadingly green checks.

Real testing requires intention and attention. It’s okay for a suite of checks to run unattended most of the time. But to be worth anything, they require periodic attention and review—or else they’re like smoke detectors, scattered throughout enormous buildings, whose batteries and states of repair are uncertain. And as Jerry Weinberg said, “most of the time, a nonfunctioning smoke alarm is behaviorally indistinguishable from one that works. Sadly, the most common reminder to replace the batteries is a fire.”

And after all this, it’s important to remember that most checks, as typically conceived, are about confirming the programmers’ intentions. In general, they represent an attempt to detect coding problems and thereby reduce programmers committing (pun intended) easily avoidable errors. This is a fine and good thing—mostly when the effort is targeted towards lower-level, machine-friendly interfaces.

Typical GUI checks, instrumented with machinery, are touted as “simulating the user”. They don’t really do any such thing. They simulate behaviours, physical keypresses and mouse clicks, which are only the visible aspects of using the product—and of testing. GUI checks do not represent users’ actions, which in the parlance of Harry Collins and Martin Kusch are behaviours plus intentions. Significantly, no one reduces programming or management to scripted and unmotivated keystrokes, yet people call automated GUI checks “simulating the user” or “automated testing”.

Such automated checks tell us almost nothing about how people will experience the product directly. They won’t tell us how the product supports the user’s goals and tasks—or where people might have problems getting what they want from the product. Automated checks will not tell us about people’s confusion or frustration or irritation with the product. And automated checks will not question themselves to raise concern about deeper, hidden risk.

More worrisome still: people who are sufficiently overfocused, fixated, on writing and troubleshooting and maintaining automated checks won’t raise those concerns either. That’s because programming automated GUI checks is hard, like all programming is hard. But programming a machine to simulate human behaviours via complex, ever-changing interfaces designed for humans instead of machines is especially hard. The effort easily displaces risk analysis, studying the business domain, learning about users’ problems, and critical thinking about all of that.

Testers: how much time and effort are you spending on care and feeding of scripts that represents distraction from interacting with the product and searching for problems that matter? How much more valuable would your coding be if it helped you examine, explore, and experiment with the product and its data? If you’re a manager, how much “testing” time is actually coding and fixing time, in which your testers are being asked to fuss with making the checks run green, and adapting them to ongoing changes in the product?

So the issue is not flaky tests, but flaky testing talk, and flaky test strategy. It’s amplified by referring to “flaky understanding” and “flaky explanation” and “flaky investigation” as “flaky tests”.

Some will object. “But that’s what people say! We can’t just change the language!” I agree. But if we don’t change the way we speak —and the way we think along with it—we won’t address the real flakiness, which the flakiness in our systems, and the flakiness in our understanding and explanations of those systems. With determination and skill and perseverance, we can change this. We can help our clients to understand the systems they’ve got, so that they can decide whether those are the systems they want.

Learn about how to focused on fast, inexpensive, powerful testing strategies to find problems that matter. Register for classes here.

Necessary Confusion and the Bootstrap Heuristic

February 11th, 2021

I’m testing a test tool at the moment. I’m investigating it for a talk. The producers of the tool claim to have hundreds of thousands of users. A few positive remarks appear in a scrolling widget on the product’s web site from people purported to be users.

Me, I can’t make head or tail of the product. It doesn’t seem to do what it’s supposed to do. It looks like a chaotic mess. It’s baffling; it’s exasperating. I don’t know where to start in analysing it and preparing a report. I’m confused. But I’m okay with that.

Any worthwhile testing starts with some degree of necessary confusion.

Why? Because worthwhile testing is primarily about learning something about a product and learning about how to test it in a complex and uncertain space. That’s by nature confusing, and that’s normal.

If the test space is neither complex nor uncertain, and if there’s little risk, you may not need to test at all, and a simple demonstration might do the trick. Knowing that the product can work might be enough, for the moment. That’s why, for developers, performing checks and automating them at the unit level can make a lot of sense. Those checks tend to address specific, atomic conditions; they’re simple to develop and perform and encode; and they provide quick feedback without slowing down development.

A product gets built from small, discrete components. Through small, gradual changes, it turns into something much bigger and more complex, with interacting components and emergent behaviours that are non-trivial.

An encounter with anything non-trivial that you’re not familiar with tends to be messy and confusing at first. At the same time, as a working tester, you’re probably under pressure to “get things right the first time” or “get everything sorted from the beginning”. But having everything sorted really means that we’re at the end of something that was unsorted, and we’re at the beginning of the next unsorted thing!

In Rapid Software Testing, we refer to the Bootstrap Conjecture:

Any process we care about that is done both well and efficiently began by being done poorly and inefficiently.

Therefore, having “done something right the first time” probably means that it wasn’t really right, or it wasn’t really the first time, or that it was trivial, or that you got lucky.

In learning about something complex and in learning how to test it, there are frequent periods of confusion. In fact, if we’re dealing with something complex and we feel we’re sure about how to test it, that should prompt us to pause and reflect: why are we so sure?

Necessary confusion is confusion for which we do not have an algorithmic resolution. To resolve necessary confusion, we must explore a complex solution space using heuristics (that is, means of solving problems that could work but that might fail) and bounded rationality (that is, reasoning in a space where there are limits on what we know and what we can know). To overcome confusion, we have to play, puzzle, make conjectures, perform experiments, miss stuff, ask questions, make mistakes, and be patient. Necessary confusion always occurs during deep learning and innovation.

We’re often trained in our cultures, in our social groups, and in our schooling to deny that we’re confused. That gets ramped up as soon as we get into the software business: appearing not to know something is socially awkward; almost seen as a sin in some circles of knowledge work. Confusion can make us uncomfortable.

As a tester, you could just write (or worse, run) a bunch of automated scripts that check a new product or feature for specific, anticipated errors. If you do that without exploring the product and preparing your mind your testing will be blind to important bugs that could be there.

No set of instructions can teach you everything you need to learn about a product, and about the ways in which diverse people will try to use it. No formal procedure can anticipate how you or other people will experience the product. No testing framework will handle surprising behaviour without you learning how to deal with that framework. No tool, no “AI”, can determine whether the product is operating correctly, or whether a product manager will regard a red bar as something that amounts to an important bug. Complete and correct knowledge about those things isn’t available in advance.

You can learn how to test in advance. That will avoid some unnecessary confusion during testing. You can learn about the technology and domain of your product in advance, and that will avoid more unnecessary confusion during testing. You can learn to use particular tools in advance, and that might spare you some unnecessary confusion during testing too.

But you can’t deeply learn a new product or feature before encountering and interacting with it. The confusion you experience when learning a product is necessary, temporary, and healthy.

The key is to accept the confusion; to recognize that it’s okay to be confused. As we interact with the product and the people around it; as we gain experience; as we practice new skills and apply new tools, some of the confusion lifts.

Start with a survey of the product. Take a tour of the interfaces — the GUI, the command line, the API. Play with it. List out its key features. Create an outline of what is there to be tested. Consider who might use it, and for what. Build on your ideas of how they might value it, and how their value might be threatened. Think about data that gets taken in, processed, stored, retrieved, rendered, displayed, and deleted. How could any of that get messed up? How could the data be mishandled, misrepresented, excessively constrained, insufficiently constrained, or Just Plain Wrong?

And then iterate. Go through the same process with each function and feature, getting progressively deeper as you go. Maybe write little snippets of code to generate some data, or to analyze the output. (Have you been working with a product for a long time? This cycle is fractal; it applies to new functions or features, or to repairs in a product you know well.)

As we learn about the product domain; as we go about the business of sensemaking; as we develop our mental models; as we talk about the product and the problems we observe… more of the confusion dissipates. This can all happen remarkably quickly if we allow ourselves just a little time for experiencing, exploring, and experimenting with the product. Ironically, we must deliberately require and allow ourselves room for spontaneity. We need to be brave and open enough to help our managers understand how necessary that kind of work is — and how powerful it can be.

When we embrace the confusion and lean in, things begin to get clearer, our code and maps and lists get tidier, our notions of risk get sharper, and we’re better prepared to search for problems. And then we’re more likely to find the deep, dangerous problems that matter—the ones that everyone has missed so far. At the beginning, though, that process starts as we pull ourselves up by our own bootstraps.

The Bootstrap Heuristic is: begin in confusion; end in precision.

Oh… and that test tool that I’m testing? There’s a reason that I’m confused: I’ve got a confusing product in front of me. The product is inconsistent with claims that its producers make about it. The product’s behaviour is inconsistent with its purpose. It seems incapable of keeping track of its state. It provides misleading results. For outsiders, it seems designed to provide the impression that testing is happening, without any real testing going on. From the inside perspective of a tester, it’s baffling, and that’s largely because it doesn’t work.

So there’s another heuristic: persistent confusion about a product—confusion that doesn’t go away—is often a pointer to serious problems in it. If you, as a tester, can’t make sense of a product, how will the product’s customers make sense of it?

After working with this product for a little more than an hour, much of the confusion I referred to above has evaporated, and I can prepare a report with confidence.

I’m only left with one thing that I find confusing:

How can anybody be fooled by a tool like this?

First Aid for the Mission Statement

January 23rd, 2021

A while back, a tester brought a patient in for treatment. It wasn’t a human patient; it was a sentence about building and testing in an organization. The tester asked me for help.

“Could you provide me with a first aid kit for this statement that came from my management?”

“We have to move on to DevOps to be able to release code more often but we also have to increase testautomation in any way we can and minimize manual time consuming testing.”

This is the sort of statement that needs more than first aid; it needs emergency room treatment. We’ll start with handling some critical problems right away to get the patient stabilized. Then we need to prepare the way for longer-term recovery, so that the patient can be restored to good health and become an asset to society.

I’ll suggest a number of quick treatments. As I do that, I’ll identify why I believe the patient needs them.

  • Replace “have to” with “choose to” in each case.

Unless someone is about to run afoul laws of nature, of government, or of ethics, no one has to do anything. People and organizations choose to do things. It’s important to preserve your agency. When you have to do things, you don’t have control over them. When you choose to do things, you remain in charge.

  • Replace “move on to DevOps” with “apply DevOps principles and practices”.

DevOps is not something to do; it’s a set of ideas and approaches for getting important things done. The central principle of DevOps—that development and operations people work together to support the needs of the business—is what matters most. If that principle is absent in the organization, there are practically infinite ways in which things will go wrong.

There will be more to say about DevOps-related practices and principles as we proceed.

  • Replace “to be able to” with “to be better able to”.

Not being a thing in and of itself, DevOps is not necessary to be able to do anything in particular. There were plenty of organizations building and releasing software successfully before DevOps, and there will be plenty of successful organizations long after DevOps is forgotten. Nonetheless, many of the ideas currently associated with DevOps could be very helpful.

DevOps doesn’t guarantee success, but applying its core principle might improve the odds.

  • Replace with “release code” with “release valuable products”.

The product is not the code, and the code is not the product. The product is the sum of code, software platforms, machinery, data, documentation, and customer support. The product is all of that together, delivering whatever experience, value, and problems that the customer encounters, good and bad. The code is part of the product, and the code enables the product. The code controls the mechanical parts of the product. That’s not to diminish the importance of the code. The code makes the product possible. If it’s a software product and there’s no code, there’s no product. If there is code, and it’s bad code, it’s less likely that we have a good product.

It’s a good thing to release valuable products. Therefore it’s a good thing to understand the code that we’re releasing—but not just the code, and not just its behaviour. That’s not trivial, but it’s the easy part of testing. The harder part of testing is understanding the relationships between the code, its behaviour and the people who will be using the product or otherwise interacting with it—including developers, operations people, and testers.

  • Replace “more often” with “at an efficient and sustainable pace”.

Delivering working software frequently is one of the principles behind the Manifesto for Agile Software Development. DevOps principles and practices are intended to support agility. Building efficiently and frequently—certainly a DevOps practice, but not a practice exclusive to DevOps—affords the opportunity to discover whether there are problems that threaten value before we inflict them on customers.”More often” isn’t the point, though, because “more often” might be good, bad, or irrelevant.

Consider the extremes.

When we build a product very rarely, problems get buried under layers of increasing complexity in the product and our inexperience with it. Testers become overwhelmed with the volume of learning and investigation (and, very probably, bug reporting) to be done.

When we build a product frenetically, we also get infrequent experience with it, because each new build comes along before we can gain experience with last one. In this case, testers get overwhelmed by the pace of the builds, and shallow testing, at best, is all that’s possible.

There is probably business value in being able to deploy software promptly, but that’s not the same as deploying software constantly. There might be contexts in which frequent deployment might provide real business value. In other contexts, a firehose of deployment can disrupt your customers, such that they don’t care about the deployment, but they do care about the disruption. The key is to recognize what context you’re in, and to minimize costly disruption.

One reasonable compromise in most cases is to set up systems to build the product easily and therefore frequently. In the build and in the processes leading up to it, include a smattering of automated checks to provide a quick alert to problems close to the surface. From the stream of builds, choose one periodically, and spend some time testing each one deeply to find rare, hidden, or subtle problems that can escape even a disciplined development process.

  • Replace “but we also” with “Thus”.

The first clause in the statement sets up the second clause. The first clause doesn’t undermine the second clause; the first clause puts legs under the second.

In our emergency treatment, let’s use “thus” to reattach the legs to the body. Also, we’ll add a sentence break, giving the patient a little more room to breathe.

  • Replace “we have to increase” with “we choose to apply”.

We’ve already covered the “we have to” part of this replacement.

Increasing something is not necessarily a good thing. In engineering work, everything is a tradeoff between desirable factors. Every activity that we might consider valuable comes with some degree of opportunity cost, reducing our capacity to do other things that we might also consider valuable. When we choose to apply something, we can choose to apply it more, or less, or just as we’re currently doing, to obtain the greatest overall benefit.

  • Replace “testautomation” with “powerful tools”.

I’m sincerely hoping that testautomation is a typo, rather than a new term. We’ll put a space in there.

There are many wonderful ways to apply tools in testing. Automated checking is one of them; only one of them.

Tools can help us to build the product, to prepare for testing, and to reconfigure our systems efficiently for better coverage. Tools can help us to probe the internals of the system to see things that would otherwise be invisible. Tools can help us to collect and represent data for analysis; to see patterns in output. Tools can help us to record and review our work.

Replacing “test automation” with “powerful tools” could help reduce the risk that “test automation” will be interpreted only as “output checking”.

  • Replace “in any way we can” with “in ways that help us”.

There are some things that we can do that we probably don’t want to do. We can serve steak with turpentine sauce, but no one should eat it and it’s a waste of good turpentine.

Tools can definitely help us with building the product quickly, and with identifying specific functional problems that might threaten its value. Tools can also be overapplied, reducing our engagement and human interaction with the product.

Fixation on tooling to exercise functions in the product can be a real problem if we forget that people use software. Even if our product is a software service, its API is used by people directly—programmers putting the API to work—and indirectly—end users who interact with the product through an interface designed for non-programmers. Functional correctness is important; so are parafunctional elements of the product: usability, performance, supportability, testability…

Tools can help us to focus our attention on important observations. Tools can also dazzle or distract us, diverting our focus from other important observations. We choose to apply tools not in any we can, but in ways that help—and our choices can include changing or dropping tools when they’re not helping.

Consequently, it might be a good idea to remember what we’re using the tools for. So, in addition to the replacement, let’s add “to build the product, to understand it, and to identify problems efficiently”.

  • Replace “and minimize manual time consuming testing” with… something else.

This particular wound has become infected, and there’s a lot of debris in it. It requires a fair amount of cleanup, emergency surgery, and some stitches.

One principle of DevOps is the idea that teams use “practices to automate processes that historically have been manual and slow”. That’s a good idea for tasks that can be mechanized, and that benefit from being mechanized. It’s not such a great idea to forget that many tasks—and parts of tasks—are non-routine, rely on expertise and tacit knowledge, and can’t be made explicit or mechanized.

Programming involves strategizing, interpreting, designing, speculating, reflecting, analysing… Testing involves all of those things too, and more. Organizations reasonably want programmers work quickly, but no one suggests “minimizing manual time-consuming programming”. This is because no one considers programming a manual process. Programming is an intellectual process; a cognitive process; a social process. So is testing.

Programmers type; that’s the manual part of programming. Just as for programming, the central work of testing is not the typing. No one refers to the typing part as “manual programming”; and when a programmer sets a build process in motion or takes advantage of plugins in an integrated development environment, no one refers to “automated programming”.

Just as there is no manual programming, and no automated programming, there is no manual testing, and there is no automated testing. There is testing.

“Manual” and “Automated” Testing

The End of Manual Testing

Anything worth doing requires some time and effort, and we usually want to apply our limited time and effort to things that are worth doing. It does make sense to minimize or eliminate the amount of time that we spend on unimportant things. It also makes sense to apply appropriate effort to things that are worth doing; to maintain their value; and to increase that value where possible.

Just as some tasks in programming can be carried out by time-saving tools like compilers, some of the tasks in testing can be carried out by time-saving tools like automated checks.

Compiling, though, is not the central task of programming. The central task of programming is modeling and expressing things in the human world in a way that machinery can deal with them. We can then use that programmed machinery to extend, enhance, accelerate, intensify, or enable human capabilities. All that requires a significant degree of preparation, technical savvy and social judgement—and time for cycles of design, experimentation, learning, and refinement. No one should complain about this taking the time it needs.

Checking is not the central task of testing. The central task of testing is the search for problems that matter—ways in which the software fails to meet the needs or desires of its users, or introduces new problems of its own. That requires not only checks for problems that we can anticipate, but a search for problems that we didn’t anticipate.  Like development work, testing work also requires a signficant degree of preparation, technical savvy and social judgement—and time for experiencing, exploring, discovery, and investigation. No one should complain about this taking the time it needs.

At least one class of testing tasks is even faster than running automated checks: the tasks that we choose not to do, because cost, value, and risk don’t warrant them. It’s worth the investment to pause every now and then to assess the relative value of unattended automated checks; instrumented tool-supported testing; and direct, unmediated experience with the product.

The trick here is to set ourselves up to do the fastest, least expensive testing that fulfills the mission of finding problems that matter before it’s too late. One way to get there is to apply fast, easy, non-interruptive checks that don’t slow down development. Then, periodically, do deep testing to find rare, subtle, hidden, intermittent, emergent bugs that might elude even highly capable and well-disciplined programmers.

So, replace “and minimize manual time consuming testing” with with “We want to minimize distracting, unhelpful, or unnecessary work. We want to and maximize our ability to evaluate and learn about the product both efficiently AND sufficiently deeply.”

With all those replacements, the text might be longer, but it’s more accurate and more precise; bigger and stronger.


We choose to move to DevOps to be better able to release valuable products at an efficient and sustainable pace; thus we choose to apply powerful tools in ways that help us to build the product, to understand it, and to identify problems efficiently. We want to minimize distracting, unhelpful or unnecessary work. We want to maximize our ability to evaluate and learn about the product both efficiently AND sufficiently deeply.

Every recovering patient can use motivation and support, so we’ll set our patient up with a motivating and supporting statement, and send them on their way together.

As developers, testers, and operations people working togther, our goal is to enable and support the business by delivering, testing, building, and deploying valuable, problem-free products. As testers, our special focus is to help people to become aware of any important problems that would threaten the value of the product to people that matter; to help our clients determine whether the product that they’ve got is the product they want.

Rapid Software Testing Explored Online for North American time zones runs March 1-4, 2021. Learn more about the class, and then register!