Quality — Metrics — Numbers — Consequences

Abstract

In this paper we present an overview over the measurement activities in which we are involved with several industry partners. We categorised these activities into four phases: The definition of quality and how to operationalise it by metrics; the selection, modification and creation of corresponding metrics; the handling of the large amount of calculated numbers; and the extraction of consequences that can be learned from them and that might improve the quality. Within every phase we tried to use some standard techniques. Our experiences show some shortcomings of these techniques within every phase, e.g. problems with using FCM, metrics suites, measurement theory, metrics filtering or metrics visualisation. We demonstrate this by examples, and we analyse and categorise the problems. Additionally we describe workarounds for these problems. Although we have to put into perspective the power of measurement, we show how a measurement program can be defined, how it has to be adjusted to different project environments, and how it can benefit the different people that are involved in a measurement process.