Timeline for Should a developer also act as a tester?
Current License: CC BY-SA 3.0
9 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 21, 2011 at 23:15 | comment | added | quickly_now | @Ben Aston: A developer should still be doing unit tests, integration tests, etc. Just not exclusively. The blind spots won't go away just because you want them to. | |
| Aug 21, 2011 at 17:10 | comment | added | deadalnix | Trying do push that where I work. This is great. However, I do think that you have to test your own code as much as possible before leting another dev testing it. | |
| Aug 21, 2011 at 13:33 | comment | added | 52d6c6af | "But testing your own code is not a good move" I disagree. Consider unit tests, integration tests and acceptance tests - all can be written by the implementing developer, in collaboration with other stakeholders if need be. Indeed, I perceive an industry move towards more automation of testing and a commensurate reduction in the size of the QA function, with a greater emphasis on the source of the problem - code quality and requirements communication (a process/organisational issue). | |
| Aug 21, 2011 at 1:34 | comment | added | quickly_now | @StasM - agreed, with one small qualification: I have found that coming back to my own code months later, I can see the faults and can do a better job of testing it objectively. But test your own after writing is very hard indeed. | |
| Aug 21, 2011 at 0:02 | comment | added | StasM | +1 it is impossible to properly test your own code. It is amazing which tricks the mind can play on you - you'll be 100% sure you coded and tested some function and it will take somebody else to show to you it's actually doesn't work except in very narrow case and it'd be obvious for you once shown - but you would never see it yourself. The mind uses cognitive shortcuts, and in testing those make impossible for the person who designed and developed the code to properly test it. | |
| Aug 20, 2011 at 10:47 | comment | added | Falcon | I generally agree with cross-testing but on some teams that'll introduce conflicts. Some people enjoy blaming others ("my stuff works, yours not, lol, I'm so much better than you") and that is unacceptable. I've witnessed that numerous times. Crosstesting should only be done between colleagues who respect each other. On my team I've introduced the nameless developer who is blamed for every bug to avoid that anyone loses his/her face. Bugs are nameless, they happen. | |
| Aug 20, 2011 at 8:23 | comment | added | quickly_now | Oh yes, sure. Completely agree. It's just that when you can't get 100% of what you want, you might have to settle for less. You know that less is not so good but it is better than nothing. | |
| Aug 20, 2011 at 8:19 | comment | added | Saeed Neamati | +1 because of the idea of cross-testing. Yeah, it might work. However, still developers tend to work with other software (for example Gmail) more professionally. | |
| Aug 20, 2011 at 8:12 | history | answered | quickly_now | CC BY-SA 3.0 |