Context-sensitivity and visual engagement in generative systems

Once recorded in a sketch or model, the formal composition of a building influences its subsequent development. Each change or addition must fit together visually with the rest. This contextual influence is rarely supported by generative systems, which favor a top-down derivation of solutions from goals. In contrast, this research investigates the potential of generative systems to support the bottom-up process of exploring a formal situation and following its visual cues. The paper defines human–computer interaction guidelines for such a system and then examines a prototype. Results suggest that generative systems can provide context-sensitive support and that tacit methods of human–computer interaction could make generative systems more palatable to architectural designers.