Automatic generalization of the spatial data has been a goal of many international researches carried out in scientific centers worldwide. Unfortunately, neither current state of art nor the existing technology makes it possible to achieve it. One of the main factors making the automated generalization of spatial data difficult is a complexity of the process. The answer to this problem is trying to test systematizing of activities being a part of the generalization tasks considered as a model. The author characterizes existing model types of the generalization process as well as she describes some experiments concerning generalization of selected elements of the General Geographic Database (GGD) in an interactive mode (performed in the DynaGEN environment). Furthermore, she proposes an idea of a constraint-based modeling in the Clarity software.

13-26