This paper investigates the use of typological knowledge in
the visual modality through a computer framework that combines
multidisciplinary technologies from computer science, that is,
artificial intelligence, software engineering, database system,
and programming language, to help provide solutions and services
to building designers. The solving of design problems frequently
involves visual thinking, which has to do with the intensive
use of visual knowledge like pictures, images, and other types
of visual displays. The recognized power of typological knowledge
in design problem solving is applied to support the exploration
of a diversity of possible design solutions represented in a
pictorial mode. The innovative use of computer science technologies
enables a smooth link of visual typological knowledge with the
design goals. Within the framework, a core technology was designed
to respond to a designer's specific needs through dynamic
user viewpoint generation, so that design solutions are associated
with relevant (retrieved) visual typologies from the knowledge
base. This has been achieved in a two-way process, in which
the designer establishes an interactive dialogue with an
experimental computerized framework.