dilton_dalton

 Profile Answers by dilton_dalton 

• Dec 31st, 2007

 

First, unless there are other verification records available, it is not possible to determine in advance if the software being tested is buggy or not.  There are methods available to determine this but they depend on unit level verification methods such as inspection and unit test.  At the end I have put a URL to a presentation that demonstrates one such method.

I am assuming that the testing that you refer to is system level testing at the end of the project. 

Defects are like mushrooms growing in the woods:  the more there are the easier they are to find.  Or put another way, the more there are, the more a given amount of effort will find.  Testing buggy software yields a lot of defects for the effort expended.  Perfect software would yield no defects no matter how much testing was done.  So one way to judge the level of defects is to measure the amount of actual test effort spent to find a defect.  The slope of the plot of cumulative defects vs cumulative test effort will give you that information.  If you use both manual and automated testing, the effort for one will have to be adjusted.  The closer the slope is to zero, the better the software, all else being equal. 

Check out this presentation. http:// www . thedreaming . org/~tomw/Testing-%20Anti_Quality%20process%202_0 . ppt