| 1/ | In September we (Frederik De Bleser and Tom De Smedt) started a design automation research project supervised by Lucas Nijs, at the St. Lucas school of arts in Antwerp. | Basically, our research deals with design automation: is it possible for a computer to make intelligent visual/communicative/marketing decisions based on a limited set of information; the same information any graphic designer would get during a briefing, brainstorm, conceptual planning phase. |
| 2/ | At the same time we're teaching graphic design students to work with DrawBot, a tool for scripting 2D compositions of text, images, vectors with Python, and saving them as a PDF. | A lot of their assignments deal with formal language. For example: what does love look like? Is it light or heavy? Sharp or soft? Red or green? And once you can discern "love"'s formal parameters, is it possible to write a Drawbot script that always produces something that resembles "love"? |
| 3/ | Right now, we're coding basic graphic intelligence in a hive: a network of ants that each do their job. They can work together. All the ants together make for an intelligent anthill (see Hofstadter). | For example, you would have a color expert system that is able to map lexical content to color, or a module that matches sharp forms to ice and fluffy forms to clouds. These two could be linked to produce colored shapes from a piece of text. The Aunt Hillary code package for DrawBot was some preliminary attempt at this. |
