A quite new series of postings on the Software Testing Club´s Blog is asking for the communities answers on frequently asked questions among the testing world:
Our community holds the wisdom to frequently asked questions we constantly being asked again and again.
Help out some testers in need!
Answer one as many as you dare, you choose!
Selected responses will be published in the next issue of The Testing Planet.
Find my answers below:
I hear rumours that our customers are using a diverse set of browsers. I only test on IE. How can I improve my coverage?
The testing effort is -at first sight- proportional to the number of browsers you have to test on. Why? Because you have to execute the same tests in every browser to ensure everything is working fine in every browser.
The no-brain-solution would be, to hire more people who then execute the tests. Probably the most expensive solution.
The step which brings imho the quickest win would be, to try to identify areas, where you can go sure, that the AUT (application under tests) behaves the same in a subset of browsers, this might be the case for e.g. browsers using the same engine (for example the Webkit is used by Chrome & Safari), or for parts of the application that happens on server-side, e.g. a search.
But what you definitely should go for is test automation. Using a tool like Selenium or Watir (which are for free btw) brings you the high benefit of having your tests automated, and with those tool´s ability to execute the same scripts on a couple of browsers with nearly no extra efforts, testing the same AUT in multiple browsers, the nightmare of having to do this manually really looses its scare.
I’m recruiting at the moment and am thinking of going down the „certification required“ recruitment route. (i.e. all applicants must have many certifications). Is this a good strategy or should I be looking for others skills instead or as well as?
It depends.
First of all, you should have clearly defined what profiles you are looking for.Then you can check if at least parts of this profile you´re looking for can be covered by available certifications.
If so: How much do you trust this certifications? Is it really hard to get them? Is it essential to have a lot of experience in the area of expertise to get the certification? Or is the exam a non-brainer and everybody can get it, just by attending a 1-day-seminar? If the certification is kind of ‚worthless‘, forget it…What in my opinion is more important then having a certain certification is the willingness to learn and a general understanding on the area of expertise. This is what you should talk about with a potential candidate during in interview, e.g. for testing: Has the candidate a sense for quality? Is he interested in delivering also unit tests in addition to his code? Is he willing to learn more about that?
My manager has said we are doing either automated testing or manual testing. This sounds wrong to me. How can I convince my manager that a mix of both might be good?
To quote a former manager of mine: „A goal of 100 % test automation is not a strategy, it´s a fantasy.“.
It is just not possible to automate 100% of all tests, be it for technical reasons, be it, that automated testing would be too expensive, or whatever…
Meaningful testing always requires to be done by a human, it needs real thinking, instead of just stupid repeating of the same scripts over and over again…Furthermore also automated testing is not possible without manual efforts. E.g. if an automated test fails, a human needs to investigate for what reason: Was it real bug? Who enters this into the bug tracker? Was it a false negative? For what reason? Needs the test script to be updated?
Update: All of my answers have been chosen and are part of Issue No. 5 of The Testing Planet.