We have spent months looking for the anatomy of space. I think each of us have really been searching for it for decades. But only recently have we decided to conjugate our efforts.

The arrangement of the features of a territory has been seen by many before us as something to be probed. The topographic endeavours of the explorers who have sliced the earth tried to make it finite. But even they know that they missed most of what was to be found.

Daratan has started their explorations with pen and paper. Their pulaus are three-dimensional projections of units extracted from a planet that doesn’t exist. And yet these representations are immediately naturalised, for they evoke geographic norms which describe the world to those who wouldn’t see it for what it is.

Anaglyphic took the opposite perspective, and set up abstraction generators that would enable others to create areas to be explored. The anaverse is a system which generates topographies according to simple rules which, called recursively, allow for unbounded diversity. Rather than observing space and trying to deduce what it is, they allow humans to build space and, in the process, prompt them to reflect on what they think it should be.

The temporary intersection of these individual journeys yielded pulau gen. In an attempt to demineralise both our media, we decided to populate the pulaus with guardian creatures. The whales, octopi and birds are as many reminders that a topographic representation is usually missing the one thing that matters the most about it: the life it can host.

======

This generator produces images in the style of Daratan, and anaverse native tokens in the style of Anaglyphic. Moons and creatures are hand drawn by Daratan, topographies and final compositions are written in javascript by Anaglyphic.

Push the S key to export a high-res (4000x4000) PNG file.

You can add a “size=” argument to specify another export size. Don’t do it if you’re not sure what you’re doing, this can be pretty intensive on your computer.

This can be made completely sizeless by vectorizing the assets. We found it to be a suboptimal solution for our purpose, but it can be done ex-post.