Yeah, I was joking. UTF-32 is a waste of 11 perfectly good bits[1], so you could just as well write each code point as 21 bits (0o0000000–0o7777777), and call i

Yeah, I was joking. UTF-32 is a waste of 11 perfectly good bits[1], so you could just as well write each code point as 21 bits (0o0000000–0o7777777), and call it UTF-21. 24 bits (3 × 0x00–0x7F) would be fine too, but 21 is the smallest possible fixed-width Unicode encoding. [1]: https://github.com/evincarofautumn/protodata https://github.com/evincarofautumn/protodata