Thoughts on Manual QA
I've been doing a lot of automated testing lately, I just finished reading these excellent articles from Sauce Labs, Why is Manual QA Still So Prevalent? and Kickstart Your Automation Efforts which totally bash manual testing and prescribes that even the QA analysts run tests from that are automated scripts. I've distilled out the key points here to why a company should adopt automation over manual QA.
How do you know if you program works without testing it? Manual testing is how everyone first learns to program. It's an natural way of checking that it works, and when the app goes live and has users, that's what they are doing- manually using it. However, as your app grows you'll be spending more and more time "checking that you didn't break anything" than actually adding features.
This is the main reason why most teams who aren't doing automated testing aren't doing it- they simply don't know about it or don't know how to do it. Many companies have a "every programmer tests his or her own code" policy which is basically a cop-out answer because the programmers who were never familiar with testing will never do it, and then the other programmers see that no one cares if you don't have tests so nobody does it. Not only does it take a lot of individual learning to get comfortable with writing automated test scripts for your apps, but there needs to be an ongoing to from upper management to have a culture that actually values of automated testing. Integrating with sauce labs, running on CI servers, and generating reports all add more complexity to the topic, and many programmers simply get scared off. They "have real work to do" and go heads down with their gulp serve banging out features full of bugs. It's toughing teaching automated testing, and it's even tougher teaching the appreciation of automated testing.
In my AngularJS applications I've built some that use Protractor tests extensively and I've worked on those that have none at all. It's key to recognize that the Protractor tests are different from unit tests. While unit tests are important from checking property i/o of your public methods, Protractor testing is more of a selenium style automate-clicking-on-stuff type of testing. I think the author of the Sauce Labs posts was referring to basically Protractor testing (although he makes no mention of this, unit testing, or the distinction between the two). Often testing is an all or none type of thing; you have people that understand and incorporate both unit and e2e tests or someone that does no testing. Occasionally you'll find dev's who are all about unit testing but haven't yet unlocked the power of Protractor and only write a few e2e tests. On the flip side, I've also seen teams that do hardly any unit testing but have QA engineers writing extensive selenium tests. Then you have the uber-testing superstars who throw Cucumber.js into the mix for a third level of testing (see the Testing Triforce). The key here is that putting in the time to learn unit testing yourself and help you team members understand it will pay off may times over in the long run. Trust me! ;)
Your comment will be posted after it is approved.
Leave a Reply.
The posts on this site are written and maintained by Jim Lynch. About Jim...