• Author
  • #12718

    First it was just test cases…now it’s the whole field of software testing.

    I dont think that testing is dead but Lee Hawkins begs to differ.

    He stresses that the rise of DevOps will change how testing is done and that the days of the manual tester may be limited. Two things that he highlights are acceptance criteria and automated testing that he suggests are limitations of DevOps

    What do you think? Has he a point? Are you worried as a manual tester?


    I think that in the past years software products and versions have been shipped at an alarming speed forcing users to keep up with this pace. Rapid user adaption to new software versions, layouts and functionalities has become the norm in technology. This is opposite to what was happening in the early days of computer software when companies were building them to help people do their work easier.
    A software tester’s role is not only to look for defects but to put him/herself in the user’s shoes and to signal if the software is not valuable for the user.
    Eliminating the testers out of the equation and working with DevOPS only will lead to the delivery of software that will bring a lot of value to businesses but less to users. On the long term people could end up using irrelevant software which would make them turn their backs on technology


    I do believe that the controversy beyond the words “testing” or “checking” is useful to move on and understand how Testing is changing. It is clear that the Testers role is in a big transformation, more and more since Agile and DevOps cultures have landed on IT departments.
    IMHO pointing towards CI/CD practices, a Tester have some choice:
    get closer to DevOps, and become an “AutomaTester” and be anyway a Tester even mixing exploration and coding, building a better and better Test Harness, helping the Team to assure deploying of a still-working product;
    get closer to ProductOwners and become an “ProducTester” and be anyway a Tester, knowing deeply every single detail about that product, leading releases, helping the Team to discover the unknown and be able to manage it.
    Of course that choice have to be empowered inside your company, any change can be useless if not accompanied by the common growth culture. Agile and DevOps cultures require a big change to everybody within an IT department:
    – Managers have to change their lead;
    – Programmers have to change their habits;
    – Tester have to change their skills.
    No one can do it alone.
    It could be not easy at all and it could take a very long time, you could fight with a long list of people and first you could lose every single of those battles.
    But it is an opportunity and as a Tester you should be curious and stubborn enough to want to know how it ends up… so I guess you may decide to take that opportunity!
    And that, to me, is what doesn’t let the Testing die!


    “The stories of my death is greatly overrated”

    There was a buzz around this in 2011:
    GTAC 2011: Opening Keynote Address – Test is Dead – YouTube


    Even I do not believe that testing is dead. Real testers (manual testing) will always find a lot more ways of breaking the system which automated testing cannot.


    Testing is not dead, testing hasn’t even grown up yet.
    Let us forget the silly “Is testing a science or an art?” discussion for now.
    Software nowadays is everywhere.It controls all our important systems, from our cars (hackable with great ease), our pacemakers (hackable with great ease), our lights (responsible for the biggest DDOS attack in years), our nuclear facilities (Stuxnet anybody?), our financial systems and voting computers (you do read the news don’t you?) and so we literary trust our lives to software.
    Problem is that we would not be allowed to live in a building designed and build along the lines we design and build software – let alone test it.
    For all fields of engineering there is a solid scientific base, the brakes in your car are more thoroughly tested and better designed than the software controlling them.
    In stead of making silly statements like these, and saying DevOps (or Agile, or ..) will change everything and testing is dead it would be wiser to concentrate on building a scientific foundation with reusable patterns to be used in certain situations (like any dome has the same basic design, a roof has support beams, and wings on planes follow aerodynamic principles) and methods of inspection to give qualified insight in the quality of the software.
    Maybe then we can be justified to trust our lives to software.


    Well, saying this that testing is dead is not the right statement, however nowadays need for manual testers has been continuously decreasing but not diminishes. Testing goes in parallel along with development and it will never come to an end so does the testing. Yeah, we can say automated testing is taking place of manual testers but again the process will take time.


    Lets start with the language we are using. The use of manual and automation testers is incorrect. Being able to code is a skill some testers have and some don’t we are all still testers. At the end of the day automation checks can only define the output of an algorithm. Skilled testers will always be needed as long as humans develop the code and use the applications.

    For me testing is only starting to get the credit and focus it deserves after years of been seen as a job anyone can do once there are “testing artifacts” created. (In Ireland anyway)


    The benefits of test automation are clear and for all to see. Regression tests run faster with more data. Test scenarios can be set up much faster. Test coverage improves. It is possible to improve auditing trails which makes regulators and stakeholders happy. Testers spend less time on repetitive tasks so can spend more time testing and thinking of test scenarios.

    Test automation also has costs. Some are clear but some are not.
    The costs associated with building and maintaining the test set are clear.
    Time spend by the tester in Functional Testing to make sure his/her test can be easily automated is noticeable but less clear to pinpoint.

    But the biggest cost can come from excluding exploratory or non scripted tests from the testing efforts.
    For example if we use a tool like Cucumber to document user stories, test scenarios and ultimately the automated tests the tester is clearly steered towards covering the described scenarios.
    This is fine but the “If – When – Then” form of describing testing scenarios is extremely limited and will only cover stories the users / stakeholders know and understand. Testing scenarios dictated by the code itself, strange behavior not thought of by either user or developer and non functional tests are very difficult or impossible to describe in these terms.
    Even when a human quickly glancing at a screen would immediately see something is wrong an automated script can and will miss it if the error situation is not described somewhere.
    It is important to be aware of the limitations of test automation and the dangers of grooming testers so that they start to adhere to the “If-When-Then” way of thinking.

    The gray matter between your ears is still the most important tool for a tester. Thinking is still the best means to come up with useful, efficient and effective tests. In the foreseeable future there is no amount of automation that is able to do it as effective and efficient as the human brain.

Viewing 9 posts - 1 through 9 (of 9 total)

You must be logged in to reply to this topic.