Chapter 5: Applied Design Metrics

---

OVERVIEW

Key Points:

Using the poor relation of Software Metrics to radically improve quality

Managing complexity in software systems

McCabe Metrics

Information Flow Metrics

Now we come to an area that always strikes me as the poor relation in the area of metrics use. This is a great shame because the techniques are easy to apply, they cost very little in terms of additional effort, they are used by engineers or programmers and it is one form of Software Metric that can have a very real impact on the quality of the product delivered to the customer.

So what is this new wonder cure? Well, firstly they are not new, in fact the concepts are as old as systems theory itself and secondly, as with all Software Metrics, they are not a panacea. Put simply, I am referring to a set of techniques, firmly based upon measurement, that are applied to designs at different levels to assess the complexity of those designs. Hence the name that I choose to describe the collection of such techniques, "Applied Design Metrics." The basic premise behind such metrics is as follows; designs that exhibit higher levels of complexity will also exhibit poorer reliability and maintainability characteristics.

You can see by the wording of that last statement that I view the results of using these techniques as indicators of quality. I will not claim that a design with a high complexity rating will always produce a system that is unreliable and unmaintainable, but I do claim that such a system is potentially more likely to exhibit those characteristics. Applied Design Metrics provide indicators or pointers to likely trouble spots within systems and, at the highest level, to potentially troublesome systems. Having said that these metrics are indicators there is one example that I will discuss for which absolute figures are suggested as a boundary beyond which it is not safe to travel.

---