Effort Estimation for Changing Requirements

Complex software systems have to cope with a steady stream of changing requirements. However, current software engineering techniques are ill equipped for assessing the impact of changes, either within a single system or across reused assets and the systems in which they are reused. This is a critical inhibitor for (amongst others) effort estimation in the object-oriented development process where reuse and incremental development play a crucial role. We have been confronted with this problem in deriving a commercial domain specific framework from a custom built application (in the area of broadcast planning). In this case the problem boils down to efficient effort estimation of customisations of the framework early in the development process. In the context of a research project we are investigating reuse contracts as a possible solution.