1. Beauty and evolution. One may continue to speculate as follows: A society with a distribution of faces corresponding to an algorithmically simple prototype face may have an evolutionary advantage. This is because face recognition (based on the given hardware, the brain) may be more successful or efficient in such a society. Evolutionary pressure may favor beautiful prototypes, where beauty is defined by the nature of the computations our brain handles well. On the other hand, the nature of these computations is influenced by typical face recognition tasks to be solved. It is difficult to analyze such mutual dependencies.
2. Learning coding schemes. Most certainly, the fractal circle scheme I discuss is different from the typical human coding scheme. Therefore, the most beautiful cartoons relative to the circle scheme will be different from the most beautiful cartoons relative to the coding scheme of most humans. Unfortunately, I cannot extract the latter (although many artists implicitly try to guess it, I believe). However, humans can learn new coding schemes. In particular, it is not hard to learn the circle scheme. Therefore I hope that some of the cartoons in this paper will be easily accessible to some readers.
3. Something ``beautiful'' needs not be ``interesting''. Interest has to do with the unexpected. But not everything that is unexpected is interesting--just think of white noise. One reason for the interestingness (for some observers) of some of the pictures shown here may be that they exhibit unexpected structure. Certain aspects of these pictures are not only unexpected (for a typical observer), but unexpected in a regular, non-random way. The formalization of ``interestingness" requires an extension of the formalism above. This, however, is beyond the scope of this discussion.