We discuss manual testing here. The actions are repetitive and mechanical so it seems that this is a strong candidate for automation. What is the value to manually testing?
Answer
When we’re testing a piece of software (anything really) we’re basing our analyses off of a specification, spec for short, of the expected behavior. Some initial manual testing based on the spec is helpful in focusing the types of automated tests we would like to perform and can highlight other aspects of the software that should be explored in automated tests which are not explicitly stated in the spec. This happens fairly often. Furthermore, once automated tests are in place, manual testing is helpful when exploring a failed automated test more carefully.
Nobody can reject that there are advantages of automation testing over manual testing click, specifically when it comes to saving time and enhance test accuracy. However, it is not just possible to have a robust QA process by relying entirely on automation. There are aspects of your QA that remain to be best managed by human engineers and manual testing.
One instance is that the automated tests are written by people and people make mistakes and can sometimes overlook or miss some things. Manual testing can serve as another layer of making sure that all bases are covered and if not they can go back to the drawing board and add more tests that may have been overlooked before.
Hi. Over 7 years of testing experience here. There are some things automation sucks at testing. e.g. Is that the right image? Is the text overlapping itself or images? It doesn’t do well at testing how humans will interact with the software, just how we expect or want them to. It can’t test audio. It misses a lot. Some memory leaks happen by staying on the site or clicking on items in a specific order, changing screen size. There are just too many variables. I have tested sites and software that had lots of automation, and still always found major defects. As a developer, unless you coded the entire site or software on your own you only really know the code you are touching and then move on and forget that stuff. Manual testers know how it functions in reality.
e.g. A prefill function field you touched may mess with prefill somewhere far from where you tested. Different prefills may end up in NAN. Databases you copy over may not copy over correctly or prefill obviously bad info. Also Dev machines and wifi are great but most client ones are much slower and may crash when yours won’t. Or a browser configuration or a company’s set up and software may mess yours up.
Or like when FLASH was used by developers all the time, it could flash red on videos and cause seizures. It may not be your code specifically but the client will still blame you for the awful video quality etc etc.
Automated Testing as stated in previous post doesn’t cover all scenarios. Also Depending on your Automation Framework. UI tests can be flaky. There are a lot of scenario based testing where there is a bunch of logic on the Backend that needs validating on the Front End over multiple Pages. Ya, you can automate that but the tests take a long time to run and imo it’s much better to have a human due the functional testing. There is just a lot of stuff a QA associate can find, especially having access to Pull Request if they understand code etc… Unit Tests I feel provide good value but maintaining and getting value out of a front end automation suite is difficult.
Automated testing and manual testing. We still need both of them.
Automated tests are better for testing quantitatively. Is it there? Are there enough data in array? Did it start from right index?
Manual tests are better for testing qualitativily. Is it right data, right place? Or right image? What about divs width? Should it be bigger? Is it right shade of gray?