This study is an empirical investigation of the relations between model measurements and model quality. The main objective of this study is to dene a set of model metrics which correlate strongly with failure reports and experts opinion. Then, to use these metrics as indicators on model quality and predictors of model failure. This study targets a group of `Simulink' models used in the development of embedded software within the automotive industry. Model Based Development has been a fundamental method in the development of embedded software in the automotive industry for the last two decades.
Alongside empirical values, this study distinguish between all factors and variables
which may efect the quality of a model as an inuence network. By identifying the
inuential factors it will be possible to prevent negative development.