Formal variations in a generative system: syntagma as a hypothesis for semantic control

Abstract

To define the means of control for generative process in a computer system that aspires design automation depends on how we understand and organize the information encoded in the program. This must be considered in the light of contingent issues, such as purpose and the understanding of its user. As part of a research by design strategy, this article addresses a Rhinoceros / Grasshopper definition whose purpose is to aid a basic design course illustrating the syntactic relations of facade elements in infill architecture. The generative process produces composition variations through the selection and displacements of architectural elements. The development of the control module implies considering the main pedagogical purpose of this project: to promote an understanding of formal manipulation that emphasizes the link between syntactic operations and the semantic dimension of architectural form. A hypothesis will be presented to address this problem: a reorganization of the control module based on the improvement of a filter system previously developed. Following logical principles of Patterns Language, by Alexander et al. (1977), we tried to circumvent the problem of independent treatment of variables with the approximation of architectural precedents where elements act together to provide a coherent response to design problems with multiple and sometimes conflicting demands. The new modality is called “syntagmatic control” – a concept of linguistics that refers to a set of words ordered according to syntactic rules - and seeks to allow controlling variations in classes of elements using predefined combinatorial patterns based on the precedents studied.