Recently, our team have been finished phrase I test, in the phrase, we are responsibility for test plan analyze, test case design, test case develop, test case implementment and perform test case.
Overall these procedure, we have been experienced one-loop test, at this procedure, we also met some problems, for example, we haven't traced develop status very well, when system have been done, we haven't produced some test design document, also, one reason is developers haven't updated test cases in time, the other I haven't considered completely, maybe It's the my first test task, so I can't control it by some method, or I don't know how to deal with some situation.
So, accroding to this experience, I try to summary:
Firstly, should trace schedule in time, distinguish some of time point, and everybody task, know when to begin, when to end, what environment shall we have?
Secondly,should design test case step by step, don't wait for all of environment and dev doc have finished, because our project carry out repeatedly, so when we haven't detail info, we should try to do it, and when info update, we can update our test cases, too.
Thirdly, when we have been designed our test cases, it doesn't mean that test cases design have been finished, when we performming our test cases, we can find some problem, i.e. expected result haven't matched our test cases, maybe it's a bug, but sometime maybe we haven't understand system requirment, so at this situation, we should update our test cases in time, also, there are some other situation, for example, our test cases exist some ambiguity term or enironment etc, according to different situation, we should clear ambiguity section.
Fourthly, we should trace bug info in time, via bug analysis, we can find some new bugs and some of bugs can instruct next phrase test, and also analyze those bugs we can find where are more bugs happen? and so on.
Finally, we should often verify our test cases, maybe request developer members to take part in, via verification, we can find a lot of good idea, and it's a good manner to improve our test cases.
Overall these procedure, we have been experienced one-loop test, at this procedure, we also met some problems, for example, we haven't traced develop status very well, when system have been done, we haven't produced some test design document, also, one reason is developers haven't updated test cases in time, the other I haven't considered completely, maybe It's the my first test task, so I can't control it by some method, or I don't know how to deal with some situation.
So, accroding to this experience, I try to summary:
Firstly, should trace schedule in time, distinguish some of time point, and everybody task, know when to begin, when to end, what environment shall we have?
Secondly,should design test case step by step, don't wait for all of environment and dev doc have finished, because our project carry out repeatedly, so when we haven't detail info, we should try to do it, and when info update, we can update our test cases, too.
Thirdly, when we have been designed our test cases, it doesn't mean that test cases design have been finished, when we performming our test cases, we can find some problem, i.e. expected result haven't matched our test cases, maybe it's a bug, but sometime maybe we haven't understand system requirment, so at this situation, we should update our test cases in time, also, there are some other situation, for example, our test cases exist some ambiguity term or enironment etc, according to different situation, we should clear ambiguity section.
Fourthly, we should trace bug info in time, via bug analysis, we can find some new bugs and some of bugs can instruct next phrase test, and also analyze those bugs we can find where are more bugs happen? and so on.
Finally, we should often verify our test cases, maybe request developer members to take part in, via verification, we can find a lot of good idea, and it's a good manner to improve our test cases.
扫一扫
5019
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
