Similarly to how environmental information is incorporated into the production of the branching structures, geometrical aspects of modeling can be integrated as well [167]. Instead of a communication module, user-defined functions are inserted. To obtain natural effects, a small number of B-spline curves is sufficient.
Figure 5.13 shows the successive construction of a compound leaf. The user defines the leaves by first entering their length, then their size, the branching angle, the style form, and, finally, by adding a random offset angle (twisting). With this method, the disadvantage of the predominantly local modeling methodology can be partially eliminated using L-systems, since the position infor-
 |
 |
mation of the objects permits a function inquiry, and thus allows the user to integrate global aspects into the production. However, the formulation of the productions seems quite laborious.
In addition to modeling with Lindenmayer systems, Prusinkiewicz et al., extend the formalism using sequentially processed Chomsky grammars, which are applied in the simulation of structurally independent plant parts. Hereby, the L-systems regulate the exchange of materials and time-dependent processes.
In conjunction with the introduction of an interactive editor for adjustment of the functions, Prusinkiewicz et al. get closer to the efficient modeling of plants. This applies also for the Xfrog system described in the next chapter. However, beforehand we will introduce a few additional approaches that, although they work with rule-based systems, use a different modeling methodology.
