get
decimal: generate five bytes by using HKDF. Take the first 13 bits and convert it to a decimal number (which will be a number between 0 and 8191 inclusive), and add 1000 (resulting in a number between 1000 and 9191 inclusive). Do the same with the second 13 bits, and the third 13 bits, giving three 4-digit numbers. In other words, if the five bytes are B0, B1, B2, B3, and B4, then the first number is (B0 << 5 | B1 3) + 1000, the second number is ((B1 & 0x7) << 10 | B2 << 2 | B3 6) + 1000, and the third number is ((B3 & 0x3f) << 7 | B4 1) + 1000. (This method of converting 13 bits at a time is used to avoid requiring 32-bit clients to do big-number arithmetic, and adding 1000 to the number avoids having clients to worry about properly zero-padding the number when displaying to the user.) The three 4-digit numbers are displayed to the user either with dashes (or another appropriate separator) separating the three numbers, or with the three numbers on separate lines.