--- name: "letterize" section: "1" platform: "linux" package: "letterize" version: "1.5-1" release: "Debian 13 (Trixie)" platforms: ["linux"] url: "https://new.carta.tech/letterize" --- # letterize(1) phone-number to letter-mnemonic generator\ ## NAME letterize - phone-number to letter-mnemonic generator\ ## SYNOPSIS **letterize** \[word...\] ## DESCRIPTION This program tries to help you find a letter mnemonic matching a given phone number. It emits to standard output each possible pronounceable mnemonic, one per line, using the American standard mapping of dial letters to numbers (2 goes to ABC, 3 to DEF, 4 to GHI, 5 to JKL, 6 to MNO, 7 to PRS, 8 to TUV, 9 to XYZ). The program uses a table of pronounceable letter-triples derived from a dictionary scan. Each potential mnemonic must be such that all of its letter-triples are in the table to be emitted. About 30% of possible triples are considered pronounceable. A typical 7-digit phone number has 19,683 possible mnemonics, but this test usually cuts the list down to a few hundred or so, a reasonable number to eyeball-check. For some numbers, the list will, sadly, be empty. It’s best to leave out punctuation such as dashes and parens. ## BUGS The filtering method doesn’t know what plausible medial triples are not reasonable at the beginnings and ends of words. I’m not sure what table position 0 (which is what 0 and 1 are mapped to) means. If you figure it out, you tell me. I really should have generated my own table, but that would have been more work than this seemed worth — if your number contains either, you probably need to generate your mnemonic in disjoint pieces around the digits anyway. ## AUTHOR Eric S. Raymond \. It’s based on a table of plausible letter-triples that had no name attached to it. Surf to \ for updates and related resources.