nx_crockford32.nx

crockford32.nx -- Crockford Base32 encoding. Douglas Crockford 2002: a base32 variant designed for human transcription. Removes I, L, O, U (visually similar to 1, 1, 0, V respectively) from the alphabet. Accepts mixed case + hyphens on decode (so \"4CQ-5M4D\" and \"4cq5m4d\" both decode to the same bytes). Alphabet: 0123456789ABCDEFGHJKMNPQRSTVWXYZ 0 1 0123456789012345678901234567890 1 Used by: - ULID (Universal Unique Lexicographically-sortable Identifier) -- 26-char ID, Crockford-32 over 128 bits - License keys, short-URL tokens, password reset tokens - Voicemail PINs, serial numbers for hardware Not RFC-standardised (RFC 4648 base32 is similar but uses a different alphabet and doesn't allow hyphens/mixed case). We ship both: base32.nx (RFC 4648) and this module (Crockford). Invariants: C1 Encoder output is upper-case only, no padding. C2 Decoder accepts mixed case + '-' between groups. C3 Round-trip exact: decode(encode(x)) == x.