In recent years, automated tests have gained ground in the software world, not only in initial levels such as unit tests and integration; today, they are trendy in end-user tests. Without a doubt, they have come with the strength to stay; they have been sometimes assumed, in my view, in a wrong way at this last level. Frequently, the first impulse of those who come across the benefits of automation for acceptance testing is to think that it solves everything. …

Surely all we have worked on QA, we have gone through that frustrating moment where after hours and hours of testing with extreme dedication and attention, when presenting the new functionality in the review, a few seconds are enough for someone without any effort to find a rookie mistake. And in front of everyone’s gaze, someone else throws the overwhelming question: “What did they test?” Or maybe: “How could they not have seen something so simple?

The most basic theory of tests says that having exhaustive tests is not an option, however, the rule is quite general and ambiguous, which makes it susceptible to multiple interpretations, such as: “ok, don’t test 100% test 70%!”.It doesn’t sound bad, but is it possible to test 70%? In most cases, it is not possible. Let’s see it with an example: a web application, not a very big one, a menu with ten options, and each one with an average of eight options. Half of these options display static content with five user interaction options. The other half contains forms…

Surely all we have worked on QA, we have gone through that frustrating moment where after hours and hours of testing with extreme dedication and attention, when presenting the new functionality in the review, a few seconds are enough for someone without any effort to find a rookie mistake. And in front of everyone’s gaze, someone else throws the overwhelming question: “What did they test?” Or maybe: “How could they not have seen something so simple?

The most basic theory of tests says that having exhaustive tests is not an option, however, the rule is quite general and ambiguous, which makes it susceptible to multiple interpretations, such as: “ok, don’t test 100% test 70%!”.It doesn’t sound bad, but is it possible to test 70%? In most cases, it is not possible. Let’s see it with an example: a web application, not a very big one, a menu with ten options, and each one with an average of eight options. Half of these options display static content with five user interaction options. The other half contains forms…