We’ve all heard this story before in one way or another, that manual testing is the dinosaur of the testing world. The story that machines and automation are going to remove the “fragile human” element from quality assurance. That the humble Manual tester is a weak link in the chain and that this link, must be removed from the Quality life-cycle to achieve a development utopia.
In the new era of self-healing tests and screenshots being attached to every step of these automated tests, there should be no reason for a Human to be the judge of quality right?
In a perfect world with perfect applications there would be very little argument for this approach, as with a perfect world the perfect applications would also be written by self-healing development…
But that simply doesn’t exist. No application is perfect.
If we are using AI and ML to heal our tests then surely, we should be using AI and ML to heal our Development processes to make them more efficient and to remove the human margin for error. I think most if not all Developers would call that absolute rubbish. So why do we accept the new overlords of testing as a replacement for people and knowledge?
Self-healing tests are an incredible step forward to a perfect development ecosystem but it does beg a question… At which point is the app broken first rather than the test?
In a self-healing ecosystem will the test be “fixed” when it encounters a bug, will it work around them to provide that sweet Green Pass that we all strive for? The answer to this is that a human has to at some stage define some sort of criteria to say whether this self-healing test is on the right track.
To use a comic book term “Who Watches the Watchmen? “
Your average SDET is quite far removed from the test scenario creation phase. This in a lot of scenarios comes from the Business first which is passed down to the test team then it finally reaches the SDET, so the SDET isn’t the watcher of the watchmen because their job is to create automation not define it.
The onus at the moment is to shift Manual testers and testing towards Automation, Manual testers can learn how to code and how to operate test automation tools with Youtube videos and tutorials because it’s really simple and easy and only costs £70. What a bargain!
But it’s not that easy is it?
Nothing ever is that easy a shift towards converting Manual testers into the masters of automation is missing the mark completely. Sure, you can watch a tutorial or go on a course but not everyone is wired in the same way. I’ve worked beside and met many Manual testers in my almost 10 years of working in testing and personally I wouldn’t swap a Manual tester for a self-healing automation tool.
Replacing manual testers skill-sets with automation specialist skill-sets doesn’t remove the issues with the Software Development life-cycle it just shifts an issue somewhere else. Adding a tool which self-heals their workflow certainly doesn’t solve this issue either. Manual testers and the processes which they undertake such as translating requirements into articulated test scenarios and observing application behaviour can’t be resolved by a tool which translates user journeys into code and then tries to fix it if the test doesn’t work anymore just doesn’t sound like it is a step forward.
There must always be a complimentary process beside Automation which generates business value and more importantly governs what Quality actually is.
In this non-perfect world of testing the wardens of testing are the Manual testers, the misunderstood, the ones with the unrivalled application knowledge, the ones who should all “pick up automation”, the protectors of quality.
From my perspective Manual testing shouldn’t go anywhere, as we get more and more far removed from the business requirements. There must always be a chain which holds the process together and from my experience this is Manual testers. Manual testing isn’t dead yet.
The humble Manual Tester is truly the Watcher of the Watchmen.