08_Build_Your_Own — invent an encoding, then implement it

Level: 201 → 301 · a project, not a lesson

The fastest way to stop confusing code point with byte is to have designed both yourself. This chapter starts as one project — a deliberately silly text format with its own character set, its own variable-length encoding on 3-bit units, and its own container, specified precisely enough that you can implement it in Rust and know when you are done, because a Python reference implementation prints every expected result — and then steps back from it to ask what any format needs before it can leave your machine.

#	Page	What it is	Status
1	Tribit — the specification	The four layers, the rules, the test vectors, and a Rust API sketch	written
2	A record has to say what it is, how long it is, and whether it arrived	Record type, length and checksum, worked on Intel HEX — the three fields tribit does not have	written

It is also the answer to a question worth asking first: is re-inventing this the wrong approach? It is the classic right one — every serious treatment of UTF-8 ends with "now write an encoder" — provided two things stay separate that beginners merge: your code points and Unicode's, and the unit encoding and the container. The spec keeps all four apart on purpose.

The second page is the one to read after the implementation compiles. Tribit stops at the container; page 2 is about the fields you wrap around a container so a stranger's program can read it back, and it takes a shipped format — Intel HEX, from 1988 — as the worked example rather than inventing a second toy.