The new medium of computer-aided design requires changes to the creative problem-solving methodologies typically employed in the development of new visual designs. Most theoretical models of creative problem-solving suggest a linear progression from preparation and incubation to some type of evaluative study of the "inspiration." These models give a communicable structure to infinitely variable creative experiences, but that perspective may need to be altered in the integration of computer applications into design education. In its infancy, computer-aided design merely saved engineering students the tedium of computation. Later on, computers were used to assist in drafting. Currently the computer can help with many aspects of visual design, including allowing for three-dimensional study models and providing access to helpful newsgroups and remote resources through the Internet. As long as students have the advantages of some previous knowledge of the programs and of appropriate hardware, computer applications can help them represent their ideas graphically. The computer-aided design environment is characterized by several qualities that require a move away from the linear problem-solving paradigm: (1) interactivity with programs, or even with the Internet, provides a cycle of immediate feedback which does not lend itself to assembly-line design or learning; (2) the visually mediated form of thinking is more holistic than linear; (3) the open-ended, discovery-oriented dynamic seems to operate without strict rules of causation; and (4) its ability to empower individual designers to make decisions conflicts with older views of the designer as a detached observer. The new model for creative problem solving is a feedback loop, an ongoing cyclical process of discovery and evaluation. Three separate studies of computer-aided design studios from fall 1993 to summer 1995 found these qualities at work with varying degrees of success; in some cases, interactivity was slow to develop, while in others, lack of personal empowerment or too much concentration on product over process was a problem. In some situations, moreover, the technology was not being used to its full potential. Future research is recommended. (Contains 13 figures and 19 references.) (BEW)