Bitspeak

Converts hex strings to pronounceable words. This may be a good format to print small identifiers or data that needs to be pronounced. For example, if we display the 0x44042f as Base64, it may become something like hS_$, which is pronounced as "height uppercase ess underline dollar sign". With Bitspeak, the same hex becomes "kupakare", which is much easier to say. That's all...

Usage

const bs = require("bitspeak");

console.log(bs.hexToBitspeak6("476ffe")); // "kunizezi"
console.log(bs.bitspeak6ToHex("kunizezi")); // "476ffe"

Examples

Bitspeak has two versions, 6-bit and 8-bit per syllable. The 6-bit version is more accurate since syllables don't get mixed up, but the 8-bit one is more round since it matches 1 byte per syllable.

Pronounce

On the 6-bit version, consonants are picked to maintain an unique map between the pronounce and the writing. Consonants like x should be understood as sh from shower, not x from dex, and consonants like g should be read as in game, even when they're followed by an i or e. Here is a table with the proper sounds:

Similarly, each syllable must be pronounced in isolation: kupakare is coo - pa - ka - re, not coo - pa - care.

Specification

Bitspeak has two flavors, 6-bit and 8-bit per syllable. They use the following tables:

6-Bit

// Consonants:

0000 -> p
0001 -> b
0010 -> t
0011 -> d
0100 -> k
0101 -> g
0110 -> x
0111 -> j
1000 -> f
1001 -> v
1010 -> l
1011 -> r
1100 -> m
1101 -> n
1110 -> s
1111 -> z

// Vowels:

00 -> a
01 -> u
10 -> i
11 -> e

8-bit

// Consonants:

0000 -> p
0001 -> b
0010 -> t
0011 -> d
0100 -> k
0101 -> g
0110 -> sh
0111 -> j
1000 -> f
1001 -> v
1010 -> l
1011 -> r
1100 -> m
1101 -> y
1110 -> s
1111 -> z

// Vowels:

0000 -> a
0001 -> e
0010 -> i
0011 -> o
0100 -> u
0101 -> an
0110 -> en
0111 -> in
1000 -> un
1001 -> on
1010 -> ai
1011 -> ei
1100 -> oi
1101 -> ui
1110 -> aw
1111 -> ow

Why?

For when you have to type, vocalize or memorize bytes for whatever reason. For example, typing a shortened URL in a phone, manually accessing an uuid in a database, or memorizing a crypto private key. Sure, that should not happen, but if it does, Bitspeak is supposed to make the process slightly less painful, as it is arguably easier to write "kupakare" correctly than "hS_$" or "44042f".

1. Typing special characters and uppercase letters in a phone is painful.

2. The pronounce is shorter (kupakare vs heightuppercaseessunderlinedollar vs fourfourzerofourtwoeff).

3. It is gentler to human menory as it has less traps ("was it uppercase or lowercase?").

We're considering using something like that for displaying the unique identifiers of Formality files, a new programming and proof language.

Relevant XKCD