DIGGING FOR CONTENTS IN METRICS VALUES Applicable processes for analysing metrics values to get worth insights into large software systems

Four of five software metrics programs fail [Pitt97]. On the other side many software projects fail because of budget or time overruns [Glas02]. From the technical point of view metrics can be applied to every system to get insights about the quality of large software systems. There exist many metrics, metric sets and metrics tools (cf. e.g. [Simo01]). Additionally there exist many studies on processes used for an efficient application of metrics. Nevertheless most of these processes are variations of the well known goalquestion-metric (GQM) paradigm that suggest a strong refinement of management goals among quality aspects to concrete metrics and its values ([BaWe84]). Often, however, the goals are not completely defined and it makes more sense to start from the metrics themselves. The metrics then deliver hints for questions and goals not yet considered. In the following we present three processes giving a wider area of how to successfully use metrics to get insights about large software systems. The processes itself are the result of customer projects done by the SQS-AG and focus on an efficient and effective quality assessment of large software systems. These processes are used to know the current state of quality, to detect concrete quality weaknesses and to suggest improvement actions.