Even when testing one variable at a time, it is impossible to test everything that can be thrown at that variable, and testing variables together can result in unexpected errors.

• Testing variables together is known as combination testing or configuration testing.

• The number of tests required for combination testing increases exponentially as more variables are added.

• Combination errors can be surprising and may seem unreasonable.

00:01

Combinations of data can lead to overflow or rounding errors, which can cause problems in the field.

• Overflow example given with a complex travel plan on an airline ticket.

• Combination testing techniques can optimize sampling to find combination-related errors.

• All pairs sampling technique would have yielded 80 tests out of 1600 for printers with video cards with memory configurations.

• Errors that occur in time and vocabulary related to program paths explained.

• Flowcharts and flow graphs used to represent program paths.

02:28

The lecturer discusses the importance of testing for consequences and provides examples of testing sequences.

• Testing for consequences is important when the program does something more complicated with a variable.

• Testing for consequences is one of the things that distinguishes skilled testers from juniors.

04:59

The  impossibility of testing all possible paths in a program.

• There are 25 possible paths in which the program exits the second time it hits X.

• The number of paths increases exponentially as the loop is executed multiple times.

• Testing all possible paths is not feasible, especially for complex programs.

• Testing a limited number of paths may not uncover certain bugs, such as a memory leak.

07:27