Traditionally, complex data models are developed to completely describe constructed facilities. These data models could be significantly simplified if the amount of information to be represented is reduced to basic or primitive data, from which more complex data can be deduced and extracted as required. This paper describes a building design framework developed in a solid modeling platform that is capable of automatically generating complex building information to support various aspects of building design. The topological relationships between building components are used as illustrative examples to demonstrate the practicality of the proposed system. A number of algorithms to deduce the topological information are also described.