Question

give atleast 3 examples in which black box testing might give the impression that "everything's ok" while the white box tests might give an error

Answer 1

Answer:

In general, it can be easy to achieve, for example, branch coverage, with cases that do not test boundary values. Black box tests will expose these problems.

Some examples in block box testing:

Caught by Black-Box, but likely missed by White-Box

• Division by Zero – white box unlikely to hit the exact number, black box may find this as a boundary value.

• Possible off by one error – likely boundary case in black-box testing.

• All performance issues – white box is not concerned with this at all.

• Memory errors (very large input) – likely to be a block box testing boundary value, but not needed for white-box testing coverage.

• Rounding errors and overflow errors – again, likely block box testing boundary cases, but not needed to test for this to achieve any white-box testing coverage.

• Requirements not implemented (missing path problem) – block box testing would easily catch (missing function), white-box testing would not since we would not know the code is missing.

2 MS Calc : sqrt(4)=2, sqrt(2) =1.412........ again reverse the order i.e sqr(1.412....)=2,sqr(2)=4

Functionally this is right and even the display will show you four(4)

Now just subtract 4 from the result and you get a precision error. This would have been easily detected in unit(white box) testing.Think what a disaster it could be for a banking application.

3.when there is a never ending loop trigger by setting something in the application say. a vehicle reached Times Square=True and once true a while loop keeps initiating x classes. Functionally here nothing would be wrong but this could be a disaster in terms of memory leak