Surrogate escaping / UTF-8b is kinda genius. Unicode has two reserved ranges of 256 values each ("surrogates") reserved for the UTF-16 encoding, which are used

Surrogate escaping / UTF-8b is kinda genius. Unicode has two reserved ranges of 256 values each ("surrogates") reserved for the UTF-16 encoding, which are used there for encoding stuff outside the BMP. Surrogate escaping realizes that 1) These never appear in valid UTF-8, because UTF-8 is a variable-length integer, and doesn't need surrogates 2) So you can use them to pass-through binary data in UTF-8.