I’d like to share a work in progress. I’ve been tossing around ideas of what to do with our 3D printer, what might be interesting and engaging. I eventually arrived at this concept.

Each country is represented in their own language, the type creating the shape of the country, each done by hand. I hope that the hand-craft aesthetic of the type is an interesting juxtaposition with the process and material. Each country will be extruded outward a different distance, to correspond to some data set, such as population. Perhaps a series of these could account for multiple data sets.

This post is just to show the proof of concept. I’m working on a piece of software to generate the files to send to the 3D printer. Read on for how I’m doing this.

The basic workflow I’m trying to achieve is providing a series of SVG vector files, in addition to a data set, and having the application output an STL file ready for 3D print consumption. There are a few obstacles in this process. First off I have to parse the SVG files. Since they are XML based, this is not too hard. Using TinyXML and borrowing bits and pieces of the Candy code base, reading in the shapes is successful.

I chose to use the Mercator map projection for these files. It allows for relatively easy conversion back to longitude/latitude coordinates and the distortion in size is relatively uniform. Meaning square things show up as squares on the globe rather than squashed rectangles. The tradeoff is no form on the poles.

Next challenge is generating the STL files. Luckily I’m not the first to tackle this in the processing community. Marius Watz provides a wonderful library which includes an STL renderer. Totally fantastic. All that I have to do is provide geometry in triangles. This means chopping apart the SVG bezier curves, and representing the faces of the extruded shapes, which are spherical. This was a huge pain.

This triangle mesh is concave, complex, and often has holes. Headache extraordinaire. I’m not familiar with much computational geometry, and I’m sure there are a thousand better faster more robust ways of doing just this. My approach was to first represent a grid of triangles inside the shape, and then do a walking path inside the shape to connect together the last few triangles.

All in all, I’m pretty excited by the results of the past day or two. I’m looking forward to a lot of work tweaking hand typography and a test print in a few weeks.