Existing software process models such as Waterfall and XP are characterised by unstated assumptions, a consequence of which is that we can not easily compare models or transfer data from one model to another. This means that software planners have no mechanism for selecting process activities that are best suited to individual projects. In this paper, we propose a framework for modelling software processes that supports representation and comparison of different kinds of software process. Our framework is based on a lift in focus from 'choosing activities' to 'identifying project objectives and selecting activities to meet those objectives'. We overview some evidence to support the claims of representation and comparison and discuss benefits and limitations of the approach.