The centre for environment computing and architecture continues to experiment with new ways to form, and this paper presents three recent projects from the MSc programme. The three projects all share underlying assumptions about the use of generative algorithms to constructform, using fractal decomposition, lindenmayer systems and the marching cubes algorithm respectively to construct three dimensional “architecturali objects. The data needed to drive the morphology however ranges from formal proportional systems and Genetic L systems programming through swarming systems to perceptive self organising neural nets. In all cases, the projects pose the question what is architectural form. While after Stanford Anderson (Anderson 66) we know it is simplistic to say that it is an automatic outcome of a proper definition of the brief, it is also difficult to accept that the form of a building is an entirely abstract geometrical object existing without recourse to social or contextual justification. In anattempt to resolve these issues we have turned to the study of systems and general system theory as a way of understanding the mechanics of emergence and morphogenesis generally, and the