I encode all programming languages from English syntax to Kanji and Hanzi and back to English when Decoding. This is for a smaller total amount of token characters being used back and forth to AI models with understanding. So not to hit AI character token limits as easily.