The data portion of a test case includes simulated input data, live input data, and predicted output data. Simulated input data are prepared beforehand to exercise the system during a given test. For example, it your system is a payroll processing system, these inputs may include employee data, hours worked, tax information, and deductions.

   Live inputs are those that cannot be prepared ahead of time, such as telemetry data from satellite. This kind of input provides advantages difficult to get from simulated input data: (a) randomness, and (b) normal "garbage" that your system must be handled properly.

   Predicted outputs are written forecasts of the exact data that should result from a given test case, where such forecasting is possible. If you determine results in advance, it is only necessary to compare actual with predicted test outputs in order to know the effectiveness of the test. Predicting outputs in advance offers advantages other than simplifying posttest analysis. It saves calendar time because the prediction can be done in parallel with other activities, before the actual tests are run. Also, using predicted outputs can be more accurate than on-the-spot posttest analysis because the latter often gives rise to hasty decisions.

Go to...

Script | Checklist