In requirements engineering, abstraction permits the isolation of relevant properties in a problem domain from the rest, making the problem more tractable for engineers.
Finding the right level of abstraction for a model is essential, because a model at the wrong level of abstraction looses much of its value when it is used to answer questions about an original under study. Too abstract, a model will lack some important details and its readers may draw imprecise or incorrect conclusions from it. Conversely, a model containing irrelevant details will be larger and more complex than necessary.
Today, deciding how much and which details to mention in a model is based on the modeler's skill and experience. We think that building models at the right level of abstraction is too important to depend solely on the modeler's skills. Thus, our research project aims to define objective measurement of a model's abstractness and systematic guidance to attain the right level of abstraction.
A major step into this direction is determining which parts of a model are actually are needed for the model to fulfil its purpose. In more formal terms, we define the purpose of a model as a set of operations that can be performed on the model. We have developed the notion of footprints of model operations, i.e. the part of a model actually used by a set of operations. Measuring or estimating footprints is a first step towards measuring a model's abstractness.
C. Jeanneret (2009). Finding the Right Level of Abstraction. Doctoral Symposium of the 17th IEEE International Requirements Engineering Conference (RE'09). Atlanta, GA, USA. [pdf]
C. Jeanneret, M. Glinz and B. Baudry (2011). Estimating Footprints of Model Operations. Proceedings of the 33rd International Conference on Software Engineering (ICSE 2011). Honolulu, HI, USA. [pdf]
C. Jeanneret, M. Glinz and T. Baar (2012). Modeling the Purposes of Models. Modellierung 2012. Bamberg, Germany. [pdf]