Guideline: In More Details

Missing information at the time of setting of the test plan

For example, the allocation of test units and/or test techniques. Occasionally, information from the developers is lacking, so that it is not possible to arrive at a satisfactory distribution of test units. The test manager may also decide to delay the allocation of test techniques to test units until the testability review has been carried out (see Prepare (AST)).

Test plan

While in practice there has probably never been a project where the plan was carried out unchanged, this does not mean that the plan is somehow unimportant. On the contrary, the plan provides a common framework that makes correcting the process easier and more effective than when working without a plan.

Retesting

A specific part of the strategy is how to deal with retesting. Normally, a test delivers defects that are then reworked. A choice must then be made as regards retesting.

For example, limited retesting can be carried out, focusing only on the adjustment. Another possibility is to carry out retesting of the total function in which the adjustment was implemented, of the total function in conjunction with surrounding functions, or even of the total system. The change can also be retested with specific test cases and a regression test can be run on the (unchanged) rest of the system. The choice of the degree of retesting is made based on the risks. Sometimes guidelines are in place; sometimes the test manager determines the retesting level from case to case. In fact, the test manager takes a kind of mini test-strategy decision, with all the steps being gone through briefly.

The devil’s quadrangle

A familiar trend is symbolised in the ‘devil’s quadrangle’, with Time, Money, Functionality and Quality as the corner points. At the start of the project, there is a certain balance between the points. A predictable course of events is that all kinds of unforeseen events occur that introduce tension into the quadrangle (see diagram below). In particular, certain activities overrun (Time) and/or cost much more than was estimated (Money). The project manager corrects this by putting restrictions on the other corner points, i.e. Quality and Functionality.

Figure 1: The devil’s quadrangle

Although in itself this is not necessarily “wrong” behaviour on the part of the project manager, it is the test manager’s job to monitor Functionality and Quality. Bearing the quadrangle in mind, the test manager indicates the consequences of the project manager’s decisions and alerts the client, for example, if the choices repeatedly fall on restricting Quality and Functionality. Timely communication of this trend in particular is difficult, which emphases the importance of an independent test manager. Depending on their perceptions, the test manager is either the “conscience” or the “thorn in the side” of the project manager and/or client. This role requires a high degree of professionalism, for the test manager has to tread carefully regarding the politics of the various interests within and beyond the project.