DECODING PARAMETRIC DESIGN DATA Towards a Heterogeneous Design Search Space Remix IMMANUEL KOH 1 , PAUL KEEL 2 and JEFFREY HUANG 3 1,3 Singapore University of Technology and Design, Singapore 1,3 {immanuel_koh|jeffrey_huang}@sutd.edu.sg 2 Massachusetts Institute of Technology, United States of America 2 keel@mit.edu Abstract. Designers or Non-Designers are not able to effectively ac- cess, view, search, discover, collect, reuse, remix and share parametric design data (PDD) for either professional or educational purposes. PDD here refers to the meta-data of 3D models generated by visual dataflow modelling software packages used in CAD/CAM industry. This inef- fectiveness is a direct consequence of the deliberately proprietary na- ture of most PDD file formats and the restricted use within their respec- tive desktop-based software environments. This paper presents an ini- tial software prototype capable of automating the process of decoding a commonly used PDD file format and then re-encoding it with new set of metrics to facilitate multiple PDD searchability, comparability and interoperability, via an integrated web interface querying a design data repository. All PDDs are conceptualized as genealogies of numerical or geometric transformations and explicitly encoded with a graph-based data structure. The goal is to eventually learn from its own big data and begin to artificially generate novel PDDs heterogeneously. Keywords. Design Decoder; Design Space Exploration; Parametric Design; Visual Analytics; Design Data. 1. Introduction The ability to seamlessly access, search, compare, reuse, recombine and share mul- tiple different parametric design data(PDD) could significantly facilitate the pro- cess of design search space exploration. PDD here refers to the meta-data of 3D models generated by visual dataflow modelling software packages used in CAD/- CAM industry. However, the diverse and deliberately proprietary set of paramet- ric design software environments and file formats make these tasks difficult to achieve. This paper presents an initial prototype capable of automating the decod- ing and re-encoding process of multiple PDDs (figure 1) belonging to a commonly P. Janssen, P. Loh, A. Raonic, M. A. Schnabel (eds.), Protocols, Flows and Glitches, Proceedings of the 22 nd International Conference of the Association for Computer-Aided Architectural Design Research in Asia (CAADRIA) 2017, 117-127. © 2017, The Association for Computer-Aided Architectural Design Research in Asia (CAADRIA), Hong Kong.