Why Models and Code Should be Treated as Friends

Abstract

Various approaches have been proposed to face the difficulties related to constructing and maintaining modern software systems. Often, they incorporate models in some part of the development or evolution process. Even the use of models at runtime seems to receive more and more attention as a way to enable the quick, systematic and automated application of change operations on software as it executes. Assuming that existing systems have been largely developed in code and that novel target architectures depend on – or even embed – models to some extent, the possible roles of models and code as well as their interaction and interchangeability need to be thoroughly examined. In this position paper, we attempt to initiate a discussion on why models and code should become closer friends.