Throughout the world numerous efforts to automate generalization are i
n progress. The results are yet to be satisfactory. Ample reasoning ca
n be given to justify the lack of success, the most important being th
at generalization is an ambiguous process, highly subjective which lac
ks definitive rules, guidelines or systematization. This paper deals w
ith the problem of generalization of vector data bases through the ana
lysis of recent developments and research in the field. These developm
ents tend to establish a promising framework which, with subsequent re
finements and the utilization of state-of-the-art computer technology,
may lead to successful results. What is needed is what lacks: Definit
ive rules in structuring the digital image of the world and developmen
t of expert systems which will intelligently manipulate this image.