1.6 PRINCIPLE NUMBER FOUR-ASK NOT FOR WHOM THE BELL TOLLS-ASK WHY?

---

1.6 PRINCIPLE NUMBER FOUR—ASK NOT FOR WHOM THE BELL TOLLS—ASK WHY?

If you attend conferences and seminars that address the topic of Software Metrics, the question often posed is: "what do I do with all this data now that I've collected it? "

One glib answer with a great deal of truth behind it is to give it to a statistician. There is a very good reason for this. Statisticians are trained to interpret data, which is an art in itself, but more importantly they realize that, for example, recognizing the presence of a trend is only the first step. The next thing to do is to find out why that trend is present. Why is something happening? Answer that question and you have useful information.

Of course, the answer to the question "why" may not be obvious from the data itself. Usually that data needs to be related back to the environment it came from. To illustrate this, imagine the trivial case where the number of defects found during testing suddenly rises from some norm. Let us also assume that the items being tested are essentially the same.

In Figure 1.1, each point could be an enhancement build on a generic software product where each build is of similar size and complexity.

Figure 1.1: Test Defects Report

What does our data tell us? Unless we relate it back to the environment, very little. Looking at that testing environment, we may find that at time t we changed our test strategy, perhaps introducing a new tool that automatically generates test cases, thus freeing up effort so that more testing could be carried out. An initial knee jerk reaction that quality (in terms of the development deliverables to testing) had gone down would have been false.

Learning to ask "why" is yet another important step towards making practical use of Software Metrics.

---