If I would need new colleague

   If I would be QA team manager, and I would need to find some people to my team, what would I do? Yes, of course, I would need to go through CV and have interviews. CV part is not very interesting, but interview - that is totally other thing.

   Testing tester about what he (she, it or any other thing, I'm trying not to be sexist, but it is hard to use abstract word) knows and understands about testing. That is not just simple task with several question about Agile and how to register bug, it can be fun and challenging for interviewer and interviewee (had to Google these names). Over my experience had some of them. Some were interesting others were boring, and after couple minutes I knew I will not work here, but still spend my time for knowledge and experience. Without knowing which interview is bad, You can't know which was good, or how good.

   While thinking about previews interviews, I starting to remember some questions I got there which in my opinion I would use in my interviews. And got some I would not use, "why?" You ask, because they are stupid, sexist or impropriety for interviews. Good example would be "where You see yourself in five years?" well here if you pay me enough.

   But back to good ones. If QA does not have imagination, that person (this time sounds OK, but how often I can use person?) will not be the best for this role. Sometimes any person is better than nothing, but I am not talking about it now. For imagination:
We have medicine scale and eight coins. One of them are fake, lither then other seven. How least number of weighting we need to have to find fake coin?

   And if You thing about three, remember this is not about knowledge of math, imagination.

   Person needs to understand situations. It is not easy to understand that you fucked up, but who understands, it is really useful person.

    When incorrectly used word can postpone release? For example in Microsoft office packet?

Didn't expect that, didn't You? Now think, and yes everybody can guess about brand name, but think little bit deeper, I'm talking about user experience, not grammar Nazis.

   Then You have imagination and You are smart enough to know then time to stop. It is time for shit to get real. Understand testing is the biggest part of testers life. So:
A simple gaming system has been specified as a set of use cases. It has been tested by the supplier and is assessed as low risk and there is pressure to release the software into the market as soon as possible. Which test techniques would be most appropriate for this testing?

   This could make You feel that You seen this somewhere, and it is, I used it from external sources, thanks for that!

   And I would finish my interview with open question:
Which testing approach You prefer? And why?

For example me - I love exploratory testing, because searching for an issue or a bug just to find it is the reason I do this work.

   Expected to finish this couple weeks ago, but life is life, nana na nana.

Your friendly neighborhood Tester.


P.S.
Doesn't feel as decent post, but to lazy to rewrite, so let it be as it is.

These robots - taking over our jobs

   From last post so much happened, didn't even had time or energy to write about something. I don't even know is somebody reading this, but any way I want to keep writing, maybe some good idea will come up one day in this blog, and I will be able to use it.

   So now, after so much time without writing or thinking anything related with testing, I'll try to jump in, for me not very much explored, waters. Automatic testing or test automation or any other name You would like to call it. I will not create any test here, or write about frameworks or tools for testing, just ideas and other crap I can think of automation.

   Main task of test automation is same as in other areas where automation is used. Remove manual work and do work more efficient. And yes, this is achievable even in testing. But not always. And I will try to explain pros and cons of this, and it wouldn't be easy.

   Thinking about automation, people image, that robots will take away their jobs, and they are right, manual work often can be replaced with automated machine which will work day and night, will not have breaks, wouldn't complain, and if something happens - will not inform other about problem. Last part can be even updated to inform personal about issues, but again exhausting testing is impossible (what again, you never mentioned that before. Well I'm not, but I'm using this in real life almost weekly). And even human errors can be made, because still robots don't create robots with already programed stuff without human interception, and even more automated tests (as much as I know right know) are written by humans only.

   There are bunch of tools for test automation, companies are creating specific things for other companies for automation, companies by them self are creating things related to automation testing, and even random citizens creating things to automate testing for websites/software they created. But how often automation is needed in testing? Very often, but not always. There are ideologies that every test needs to be automated. But this is not true. Don't agree? Well, maybe You are write, but I think differently.

   There are bunch of cases where automation will take to long, or will not be so efficient or accurate as human. Automatic test can't decide that gray text on white background is hard to read - text is there, no errors in text, so it is correct, but user cannot see a shitty message You are showing to him. So automatic test passed, but manual test not. Other example can be testing website and searching elements by ID's - automatic test can find element with correct ID, it can make actions and report result, but user cannot find this element, because it is overlapped with table or something else. I can think of bunch similar test that will give on some level correct results while they are incorrect in human perception.

   But believe me - automatic tests are like a water to thirsty tester in a week of regressions with thousands of test cases, after months of repetitive work. Regressions period can be so boring and repetitive that at least part of test automated, even with bugs and required overview is better, then no automation.

   I think that You shouldn't trust person who says that automation is everything, and same way - who says manual testing is life. There should be equilibrium, balance in life. Eating only cake is not healthy, but not eating cake - huge part of joy of eating can be missed.

Your friendly neighborhood Tester