2) What similarities and differences in approach can be observed? The generation of shapes that conform to particular styles, using shape computation tools based on the mathematics of shape grammars [Stiny 1980], has been demonstrated in a number of domains [Prats et al 2006]. Researchers at the University of Leeds have built the world’s first and only 3D shape grammar implementation for curvilinear shapes [Chau 2004]. The basic elements of a shape grammar are shown in Figure 1. The box at the top of the figure shows an initial shape (that seeds the computation) and the two shape rules that are applied during the computation. The shapes at the bottom of the figure show a fragment of the network of shapes that can be computed from the initial shape through the application of the shape rules. The application of a shape rule involves two key steps. Firstly, the shape on the left-hand side of a rule must be identified in the shape from which a new shape is to be computed; this is referred to as “sub-shape detection”. Secondly, the rule is applied by replacing the sub-shape from the left-hand side of the rule with the shape on the right-hand side of the rule. Once a sub-shape has been detected, the Leeds system can automatically apply a rule. However, the sub-shapes have to be identified manually because the automatic detection of sub-shapes is an open research question within the shape grammar community.