Encode Your Data With Emoji

Originally published at: https://boingboing.net/2020/06/10/encode-your-data-with-emoji.html


Every generation gets the ROT13 they deserve.


Messages full of emojis already seem encoded to me. :older_woman: Kids today won’t be happy until all logophiles are trapped in a hieroglyphic hellscape! Fortunately, I can encode my data using cursive text. Workers at the post office have started complaining that they can’t understand the script written on envelopes. Mwah hah hah! :nerd_face:


Ecoji is like Base64 encoding, but it’s Base1024 and uses emojis instead of the alphabet.

I was excited at first, but then I realized that I was flashing back to the Uuencoding adventures of my youth (i.e. porn).

Text is dead. Time to move on.

Time for machine learning image analysis and synthesis encoding.

Here’s how it works.

  • Write out a string of emojis expressing a thought, preferably a complex thought.
  • Write a second version, another similar string of emojis, expressing an equivalent or similar thought
  • Iterate as many versions as you can stand, building up a corpus of different ways to express some idea or set of ideas.
  • Use some Machine Learning mumbo-scripto to analyze the corpus
  • View the Output and promptly experience an appropriate emotive response ⁠— personally I live for amused bewilderment, and often the mumbo-scripto delivers the goods

Then there is the translation game from Philip K. Dick’s Galactic Pot-Healer, using networked computers to translate idiomatic phrases from one language into another and back again, seeking peculiarly unhelpful yet poetic results.

Dick’s poor schlubs played the Game with words, but we could use our mumbo-scripto to play the translation game with emoji.

If you have not read Dick, you might start with GPH. It’s short, for one thing. And it’s cohesive (some of his work is ragged, needs editing). And while some people say they can’t stand GPH, I find it beautiful and charming and true to the author’s heart every time I read it, which is about yearly.


Some rather inefficient message density there(!) This encoder appears to produce 40 emoji for some 49 boring old fashioned ascii-ish characters (BOFAC). Whereas if we examine some vital emoji encoding oh…say like “Woman in Suit Levitating:business_suit_levitating:t2:‍♀ (the critical applications of which are abundantly obvious to all, no?) then we learn that is encoded with something like ‘U+FE0F’ which is 16 bits, so at least we should be getting twice the BOFAC for each Tanabata tree :tanabata_tree: emoji, and/or a compression rate of something like 2, no?



Ixnay on the ecretsay oomerbay odecay!


This topic was automatically closed after 5 days. New replies are no longer allowed.