We discuss different 'evolution aims' or 'goal functions' obtained in diffe
rent fields of modeling 'real' nature, especially in ecology and physics. F
urthermore we argue that the infinite complexity of nature as a whole deman
ds also a model with infinite complexity. The only candidates with this pro
perty which are known up to now are the so called class IV cellular automat
a. In these models evolution is open, i.e. there exists no goal function wh
ich can determine the longtime behavior. It follows that it seems impossibl
e to ever find a universal goal function for the development in nature. (C)
2000 Elsevier Science B.V. All rights reserved.