#primify

Primify takes an image and embeds it into a prime number, so that, when viewed from a distance, it resembles the original image. It was created by Levi Borodenko, a masters student in mathematics and statistics at the University of Oxford. While it works particularly well for logos and simple line art, was actually first developed for portraits.

The idea came while I was a counselor at PROMYS -- a summer camp providing an immersion into rigorous mathematics. We had a secret Santa and me being a poor student, I tried to come up with a gift idea that would light a maths majors heart without being pricey. So I decided to find a prime that looked like the face of the person that I had to gift. I coded up a quick and dirty script (back then I was using python to find the number and then mathematica to find the next prime). After doing this for a few years I decided that more people should be able to enjoy generating these primes. This is when I coded up the open-source tool.

In primify, the underlying data is not only used in a graphical way, but also holds information when read as text; in this case, a single prime number. This is multicoding, where content holds multiple meanings when read in different systems. It's a strategy familiar not only from esolangs like legit, which multicodes between git commands and code, and polyglots, where one program executes successfully in multiple programming languages. The strategy here, where the text functions as a picture, is more akin to how multicoding functions in obfuscated code that builds on concrete poetics, where the opacity of the code is offset with a visual clue as to its function. We can see this in the example cited here by Adrian Cables, or the cauldron example in this post about the IOCCC. The term multicoding comes from the seminal paper on esolangs and code art, "A Box Darkly" by Matteas and Montfort (as "multiple coding").