One of the most publicized goals of modern computer society is to provide flexible access to information for everybody, anywhere, anytime. To make data management systems successful in satisfying user quest for information, an abstractolute requirement is the capability of tailoring the information to the peculiar needs of the targeted user. Unfortunately, this is a domain where current technology fails to support adequate functionality. Basically, existing approaches to such versatility rely on the idea that all possible views of a piece of information can be derived from a single ultimate representation. This may work in a centralized setting, but it is definitely unlikely to provide a sufficient framework in distributed, interoperable environments. This paper presents a research and development project that intends to explore the issue related to representation and management of multiple representations of the same reality. A specific focus of the project is spatio-temporal information and the support of multi-scale geographic databases, where objects are represented at many different resolution levels.